Getting started with Kandinsky
- Introduction
- Setup
- 1) Loading the toy dataset
- 2) Initializing Kandinsky data
- 3) Visualizing the tissue
- 4) Creating tissue region masks
- 5) Normalizing gene expression
- 6) Neighbourhood composition
- 7) Neighbourhood-based clustering
- 8) Co-localisation / dispersion analysis
- 9) Hot and cold expression areas
- 10) Changing the neighbourhood definition
- 11) Downsizing dense neighbourhoods
- Next steps
Introduction
Kandinsky is an R package for neighbourhood analysis of spatial omics data. Starting from cell or spot coordinates and their associated gene/protein expression, Kandinsky groups cells or spots into neighbourhoods (c/s-NBs) and uses this spatial context to:
- describe and cluster cells/spots according to the composition of their neighbourhood
- test for spatial co-localisation or dispersion between cell/spot types
- detect hot and cold spatial expression areas for any gene, protein, or signature score
This vignette walks through the core Kandinsky
workflow using a small simulated dataset bundled with the
package, so that it can be run quickly on any machine without
downloading external data. The toy dataset was generated with SRTsim and mimics a
spatial transcriptomics section of the human dorsolateral prefrontal
cortex (DLPFC), with spots annotated into 6 cortical layers
(Layer1-Layer6) plus white matter
(WM).
For more advanced, technology-specific tutorials (CosMx, Xenium, IMC, Visium) that reproduce the analyses of the Kandinsky paper, see the other vignettes.
1) Loading the toy dataset
Kandinsky ships a small simulated dataset, already stored as a
Seurat object, that can be retrieved with
system.file(). If Kandinsky needs to call any external
Python tool (for instance, some of the optional benchmarking/clustering
functions described at the end of this vignette), the path to a
compatible Python environment can be set once at the top of the
script:
We then load the toy dataset:
data = system.file('extdata','SRTsim_simulated_data_Seurat.rds', package = 'Kandinsky')
data = readRDS(data)
data
#> An object of class Seurat
#> 80 features across 3611 samples within 1 assay
#> Active assay: RNA (80 features, 0 variable features)
#> 1 layer present: counts
#> 1 image present: imageThe object contains 3611 spots and 80 genes, with a single
RNA assay. Spot-level metadata stores the spatial
x/y coordinates together with a
label column reporting the simulated cortical layer of
origin:
head(data@meta.data)
#> orig.ident nCount_RNA nFeature_RNA x y
#> AAACAAGTATCTCCCA.1 SeuratProject 1204 67 440.6391 381.0981
#> AAACAATCTACTAGCA.1 SeuratProject 206 53 259.6310 126.3276
#> AAACACCAATAACTGC.1 SeuratProject 630 61 183.0783 427.7678
#> AAACAGAGCGACTCCT.1 SeuratProject 927 69 417.2367 186.8137
#> AAACAGCTTTCAGAAG.1 SeuratProject 635 70 152.7003 341.2691
#> AAACAGGGTCTATATT.1 SeuratProject 605 70 164.9415 362.9163
#> label Cell
#> AAACAAGTATCTCCCA.1 Layer3 AAACAAGTATCTCCCA-1
#> AAACAATCTACTAGCA.1 Layer1 AAACAATCTACTAGCA-1
#> AAACACCAATAACTGC.1 WM AAACACCAATAACTGC-1
#> AAACAGAGCGACTCCT.1 Layer3 AAACAGAGCGACTCCT-1
#> AAACAGCTTTCAGAAG.1 Layer5 AAACAGCTTTCAGAAG-1
#> AAACAGGGTCTATATT.1 Layer6 AAACAGGGTCTATATT-1
table(data$label)
#>
#> Layer1 Layer2 Layer3 Layer4 Layer5 Layer6 WM
#> 273 253 989 218 673 692 5132) Initializing Kandinsky data
Kandinsky uses spatial coordinates to build cell/spot polygons and to
identify neighbourhoods, relying on the kandinsky_init()
function. Because our toy dataset was not generated by one of the
natively-supported platforms (Visium, VisiumHD, CosMx, Xenium,
MERSCOPE), we set tech = 'other' and manually point
Kandinsky to the columns containing spot centroid coordinates using
xcoord_other and ycoord_other.
We define spot neighbourhoods (s-NBs) using the Delaunay
neighbour method (nb.method = 'D'), considering
only the sphere of influence (SOI) subgraph (soi = TRUE) to
limit spurious neighbour links:
data = kandinsky_init(data, tech = 'other', nb.method = 'D', soi = T,
xcoord_other = 'x', ycoord_other = 'y')
#> Building Kandinsky slot "sf"...
#> Building Kandinsky slot "nb"...
#> Building Kandinsky slot "tx"...
#> Building Kandinsky slot "fov_mask"...
#> Kandinsky data loaded into Seurat object!This adds a new kandata object to the Seurat
tools slot, accessible through the KanData()
accessor:
slotNames(KanData(data))
#> [1] "platform" "nb" "sf" "nb.type"
#> [5] "spot_distance" "texture" "nnMat" "tx"
#> [9] "fov_mask" "img" "mask"
#Visualize lines connecting each spot to its neighbours
nbs=as(spdep::nb2lines(nb=KanData(data,'nb')$neighbours, coords = KanData(data,'sf')$geometry), 'sf')
ggplot()+theme_classic()+geom_sf(data=nbs,color='red')
KanData() returns:
sf: ansfobject storing spot geometries (and coordinates)nb: the sparse neighbour adjacency matrix in a weighted list (listw) formattx/img/fov_mask: additional slots populated depending on the technology (empty here, since we usedtech = 'other')
We can inspect the neighbourhood network we just built with
nb_summary():
nb_summary(data)
#> Neighbour method: Delaunay triangulation (D), soi = TRUE
#> Total cells/spots: 3611
#> Total neighbour links: 21228
#> Minimum number of neighbours per cell/spot: 1
#> Mean number of neighbours per cell/spot: 5.88
#> Median number of neighbours per cell/spot: 6
#> Max number of neighbours per cell/spot: 7
#> Number of cells/spots with no neighbours: 03) Visualizing the tissue
Kandinsky provides the KanPlot() function to map
cells/spots into space and colour them according to any metadata column
or gene/feature of interest. Let’s visualize the simulated cortical
layers:
We can also inspect the distribution of s-NB sizes across spots with
nbSizePlot(), which is useful when working with
distance-based neighbour definitions
(nb.method = 'C'/'M') that can produce a
variable number of neighbours per spot/cell. Since we used a Delaunay
definition here, given the regular disposition of Visium spots, most of
the spots have 6 neighbours:
4) Creating tissue region masks
Kandinsky can draw a spatial mask polygon around a
group of cells/spots sharing the same annotation, using
create_cell_masks(). Internally, dbscan is
used to find spatially-dense clusters of the cells/spots of interest,
and concaveman is then used to trace a concave hull around
each cluster.
This is useful, for instance, to outline anatomical regions
(e.g. white matter, tumour core, follicles) directly from spot/cell
annotations, without needing a separate segmentation mask image. Let’s
outline the white matter (WM) region of our simulated
tissue:
wm_mask = create_cell_masks(data, label = 'label', label_class = 'WM',
eps = 30, minPts = 6, concavity = 2,buffer=1)
wm_mask = sf::st_union(wm_mask) #Only needed when multiple masks are generated in different regions
wm_mask
#> Geometry set for 1 feature
#> Geometry type: POLYGON
#> Dimension: XY
#> Bounding box: xmin: 138.3339 ymin: 340.9892 xmax: 286.9136 ymax: 504.8254
#> CRS: NA
#> POLYGON ((138.3789 470.5587, 138.3789 470.5629,...
#Label spots falling within white matter mask using st_contains()
data$within_wm_mask = 'N'
data@meta.data[sf::st_contains(wm_mask,KanData(data,'sf')$geometry) %>% unlist(),]$within_wm_mask = 'Y'
knitr::kable(table(data$within_wm_mask,data$label))| Layer1 | Layer2 | Layer3 | Layer4 | Layer5 | Layer6 | WM | |
|---|---|---|---|---|---|---|---|
| N | 273 | 253 | 989 | 218 | 673 | 692 | 0 |
| Y | 0 | 0 | 0 | 0 | 0 | 0 | 513 |
The output is an sf polygon data frame (one row per
detected spatial cluster) that can be overlaid on top of a
KanPlot():
KanPlot(data, feature = 'label',pt.size=2) +
geom_sf(data = wm_mask, inherit.aes = FALSE, fill = NA, color = 'white', linewidth = 1) +
labs(fill = 'Layer')
Key parameters to tune are eps/minPts
(density-clustering sensitivity, passed to
dbscan::dbscan()) and
concavity/lengthThreshold (shape detail,
passed to concaveman::concaveman()); a buffer
distance can also be added to expand/shrink the resulting mask. When
multiple samples are stored in the same object, sample_key
can be set so that masks are computed independently for each sample.
5) Normalizing gene expression
Before any downstream expression-based analysis, we normalize raw
counts using the standard Seurat NormalizeData()
function:
data = NormalizeData(data, verbose = FALSE)
data = UpdateSeuratObject(data)
#> Validating object structure
#> Updating object slots
#> Ensuring keys are in the proper structure
#> Ensuring keys are in the proper structure
#> Ensuring feature names don't have underscores or pipes
#> Updating slots in RNA
#> Updating slots in image
#> Setting assay used for NormalizeData.RNA to RNA
#> Validating object structure for Assay5 'RNA'
#> Validating object structure for SlideSeq 'image'
#> Object representation is consistent with the most current Seurat versionWe can now visualize the spatial expression of any gene with
KanPlot(). Gene ENSG00000168314
(MOBP, a myelin/white-matter marker) is a good example, since
it should be enriched in the WM region of our simulated
tissue:
KanPlot(data, feature = 'ENSG00000168314', palette_cont = 'viridis',pt.size=2) +
labs(fill = 'MOBP')
6) Neighbourhood composition
The nnMat() function summarizes, for every spot, the
composition of its s-NB in terms of any categorical annotation - here,
the simulated layer of origin (label). The resulting matrix
is stored in the nnMat Kandinsky slot:
data = nnMat(data, label = 'label', return.seurat = TRUE)
#> Creating neighbour matrix using pre-defined neighbour network
knitr::kable(head(KanData(data,'nnMat')$nnMat))| Layer1 | Layer2 | Layer3 | Layer4 | Layer5 | Layer6 | WM | tot_nn | label | |
|---|---|---|---|---|---|---|---|---|---|
| AAACAAGTATCTCCCA.1 | 0 | 0 | 6 | 0 | 0 | 0 | 0 | 6 | Layer3 |
| AAACAATCTACTAGCA.1 | 5 | 0 | 0 | 0 | 0 | 0 | 0 | 5 | Layer1 |
| AAACACCAATAACTGC.1 | 0 | 0 | 0 | 0 | 0 | 0 | 6 | 6 | WM |
| AAACAGAGCGACTCCT.1 | 0 | 0 | 6 | 0 | 0 | 0 | 0 | 6 | Layer3 |
| AAACAGCTTTCAGAAG.1 | 0 | 0 | 0 | 0 | 5 | 0 | 0 | 5 | Layer5 |
| AAACAGGGTCTATATT.1 | 0 | 0 | 0 | 0 | 0 | 6 | 0 | 6 | Layer6 |
Each row reports, for a given spot, how many of its neighbours belong
to each layer, plus the total number of neighbours (tot_nn)
and the spot’s own annotation.
This neighbourhood composition matrix can be directly queried using
nn_query(), which internally relies on
dplyr::filter()-compatible syntax. For instance, we can
flag Layer6 spots that have at least one WM
spot in their neighbourhood - i.e. spots close to the
layer6/white-matter boundary:
data = nn_query(data,
query = "WM >= 1 & label == 'Layer6'",
anno_name = 'near_WM')
table(data$near_WM)
#>
#> N Y
#> 3556 55data$boundary_plot = ifelse(data$near_WM == 'Y', 'Layer6 (near WM)',
ifelse(data$label == 'WM', 'WM', 'Other'))
KanPlot(data, feature = 'boundary_plot',pt.size=2,
palette_discrete = c('Other' = 'grey85', 'WM' = '#7B1B38', 'Layer6 (near WM)' = '#1A6AA0')) +
labs(fill = '')
nn_query() can be also used to directly perform
differential expression analysis between two group of spots/cells with
different neighbourhood composition. To do this, nn_query()
internally call Seurat::FindMarkers() function.
label and which arguments define the cell/spot
type annotation used to generate the neighbourhood composition matrix
and the specific cell/spot type to consider for differential expression
analysis, respectively. Additional parameters for
Seurat::FindMarkers() can also be provided through
nn_query().
For instance, we can compare the gene expression profile
Layer6 spots that have at least one WM spot in
their neighbourhood versus the rest of Layer6 spots,
looking for genes with a |log2 fold-change| higher than log2(1.5):
library(presto) # Speed up differential expression analysis
#> Loading required package: Rcpp
#> Loading required package: data.table
#> data.table 1.17.8 using 4 threads (see ?getDTthreads). Latest news: r-datatable.com
#>
#> Attaching package: 'data.table'
#>
#> The following objects are masked from 'package:dplyr':
#>
#> between, first, last
de = nn_query(data,
query = "WM >= 1",
anno_name = 'near_WM',
diffexpr = T,label='label',which = 'Layer6',logfc.threshold = log2(1.5))
#> Differential expression between 55 near_WM-Y and 637 near_WM-N cells
knitr::kable(head(de))| p_val | avg_log2FC | pct.1 | pct.2 | p_val_adj | |
|---|---|---|---|---|---|
| ENSG00000197971 | 0 | 1.607163 | 1.000 | 0.981 | 0 |
| ENSG00000131095 | 0 | 1.651362 | 1.000 | 0.826 | 0 |
| ENSG00000173786 | 0 | 1.658929 | 0.982 | 0.739 | 0 |
| ENSG00000123560 | 0 | 1.361294 | 1.000 | 0.926 | 0 |
| ENSG00000109846 | 0 | 1.422594 | 1.000 | 0.799 | 0 |
| ENSG00000107317 | 0 | 1.062969 | 1.000 | 0.943 | 0 |
The same neighbourhood composition matrix underlies Kandinsky’s unsupervised neighbourhood-based clustering, described next.
7) Neighbourhood-based clustering
Once a neighbourhood composition matrix has been built with
nnMat(), Kandinsky can group spots/cells into
c/s-NBs-based clusters according to the similarity of
their neighbourhood composition. The default method,
nbCluster(), relies on model-based clustering (R package
mclust); the argument n_clust accepts a vector
of candidate cluster numbers, and the optimal one is selected via the
Bayesian information criterion (BIC):
data = nbCluster(data, n_clust = 3:9)
#> Performing model-based clustering using mclust...
#> fitting ...
#> | | | 0% | |= | 1% | |= | 2% | |== | 3% | |=== | 4% | |==== | 5% | |==== | 6% | |===== | 7% | |====== | 8% | |====== | 9% | |======= | 10% | |======== | 11% | |======== | 12% | |========= | 13% | |========== | 14% | |=========== | 15% | |=========== | 16% | |============ | 17% | |============= | 18% | |============= | 19% | |============== | 20% | |=============== | 21% | |================ | 22% | |================ | 23% | |================= | 24% | |================== | 25% | |================== | 26% | |=================== | 27% | |==================== | 28% | |===================== | 29% | |===================== | 30% | |====================== | 31% | |======================= | 32% | |======================= | 33% | |======================== | 34% | |========================= | 35% | |========================= | 36% | |========================== | 37% | |=========================== | 38% | |============================ | 39% | |============================ | 40% | |============================= | 41% | |============================== | 42% | |============================== | 43% | |=============================== | 44% | |================================ | 45% | |================================= | 46% | |================================= | 47% | |================================== | 48% | |=================================== | 49% | |=================================== | 51% | |==================================== | 52% | |===================================== | 53% | |===================================== | 54% | |====================================== | 55% | |======================================= | 56% | |======================================== | 57% | |======================================== | 58% | |========================================= | 59% | |========================================== | 60% | |========================================== | 61% | |=========================================== | 62% | |============================================ | 63% | |============================================= | 64% | |============================================= | 65% | |============================================== | 66% | |=============================================== | 67% | |=============================================== | 68% | |================================================ | 69% | |================================================= | 70% | |================================================= | 71% | |================================================== | 72% | |=================================================== | 73% | |==================================================== | 74% | |==================================================== | 75% | |===================================================== | 76% | |====================================================== | 77% | |====================================================== | 78% | |======================================================= | 79% | |======================================================== | 80% | |========================================================= | 81% | |========================================================= | 82% | |========================================================== | 83% | |=========================================================== | 84% | |=========================================================== | 85% | |============================================================ | 86% | |============================================================= | 87% | |============================================================== | 88% | |============================================================== | 89% | |=============================================================== | 90% | |================================================================ | 91% | |================================================================ | 92% | |================================================================= | 93% | |================================================================== | 94% | |================================================================== | 95% | |=================================================================== | 96% | |==================================================================== | 97% | |===================================================================== | 98% | |===================================================================== | 99% | |======================================================================| 100%
table(data$nbClust.D_soiTRUE)
#>
#> 1 2 3 4 5 6 7 8 9
#> 602 585 202 280 253 144 493 887 165
Alternative clustering methods
Kandinsky also implements simpler/alternative neighbourhood-based
clustering strategies. kmeans_clust(), inspired by the ISCHIA package, applies
k-means on the (row-normalised) neighbourhood composition matrix instead
of model-based clustering, again testing a range of candidate cluster
numbers and selecting the one minimizing the within-cluster sum of
squares:
data = kmeans_clust(data, n_clust = 3:9, label = 'label')
#> Creating neighbour matrix using pre-defined neighbour network
#> Optimal number of clusters: 7
table(data$kmeans_clusters)
#>
#> 1 2 3 4 5 6 7
#> 846 147 608 1457 182 232 139data$kmeans_clusters = as.character(data$kmeans_clusters)
KanPlot(data, feature = 'kmeans_clusters',pt.size=2) + labs(fill = 'k-means cluster')
We can quantify the agreement between the two clustering strategies
using the adjusted Rand index (ARI, via
mclust::adjustedRandIndex()):
data$nbClust.D_soiTRUE = as.character(data$nbClust.D_soiTRUE)
mclust::adjustedRandIndex(data$nbClust.D_soiTRUE, data$kmeans_clusters)
#> [1] 0.6143368Other alternative neighbourhood-clustering wrappers are available for
users who want to benchmark Kandinsky against external tools, including
spatopic_clust() (SpaTopic), hoodscanR_clust()
(hoodscanR) and giotto_clust() (Giotto) - all
pure-R/Bioconductor based - as well as cellcharter_clust(),
spatialleiden_clust() and mesa_clust(), which
call Python-based tools (CellCharter, SpatialLeiden, MESA respectively)
through reticulate. The latter require a Python environment
with the relevant dependencies installed, pointed to via their
python_path argument (this is what the pypath
variable defined at the top of this vignette is for):
# Requires a python environment with CellCharter/squidpy/scanpy/scvi installed (see pypath above)
data = cellcharter_clust(data,n_clust = 3:9,mode = 'rna', python_path = pypath)
#> Initializing CellCharter with reticulate
#> Loading CellCharter and its dependencies...
#> Converting Seurat object to anndata...
#> Running CellCharter...
#> Training: 0%| | 0/400 [00:00<?, ?it/s]Epoch 1/400: 0%| | 0/400 [00:00<?, ?it/s]Epoch 1/400: 0%|▎ | 1/400 [00:01<07:08, 1.07s/it]Epoch 1/400: 0%|▏ | 1/400 [00:01<07:08, 1.07s/it, v_num=1, train_loss=252]Epoch 2/400: 0%|▏ | 1/400 [00:01<07:08, 1.07s/it, v_num=1, train_loss=252]Epoch 2/400: 0%|▍ | 2/400 [00:01<05:43, 1.16it/s, v_num=1, train_loss=252]Epoch 2/400: 0%|▍ | 2/400 [00:01<05:43, 1.16it/s, v_num=1, train_loss=217]Epoch 3/400: 0%|▍ | 2/400 [00:01<05:43, 1.16it/s, v_num=1, train_loss=217]Epoch 3/400: 1%|▌ | 3/400 [00:02<04:51, 1.36it/s, v_num=1, train_loss=217]Epoch 3/400: 1%|▌ | 3/400 [00:02<04:51, 1.36it/s, v_num=1, train_loss=207]Epoch 4/400: 1%|▌ | 3/400 [00:02<04:51, 1.36it/s, v_num=1, train_loss=207]Epoch 4/400: 1%|▊ | 4/400 [00:02<04:07, 1.60it/s, v_num=1, train_loss=207]Epoch 4/400: 1%|▊ | 4/400 [00:02<04:07, 1.60it/s, v_num=1, train_loss=202]Epoch 5/400: 1%|▊ | 4/400 [00:02<04:07, 1.60it/s, v_num=1, train_loss=202]Epoch 5/400: 1%|█ | 5/400 [00:03<03:42, 1.77it/s, v_num=1, train_loss=202]Epoch 5/400: 1%|█ | 5/400 [00:03<03:42, 1.77it/s, v_num=1, train_loss=200]Epoch 6/400: 1%|█ | 5/400 [00:03<03:42, 1.77it/s, v_num=1, train_loss=200]Epoch 6/400: 2%|█▏ | 6/400 [00:03<03:36, 1.82it/s, v_num=1, train_loss=200]Epoch 6/400: 2%|█▏ | 6/400 [00:03<03:36, 1.82it/s, v_num=1, train_loss=198]Epoch 7/400: 2%|█▏ | 6/400 [00:03<03:36, 1.82it/s, v_num=1, train_loss=198]Epoch 7/400: 2%|█▍ | 7/400 [00:04<03:34, 1.83it/s, v_num=1, train_loss=198]Epoch 7/400: 2%|█▍ | 7/400 [00:04<03:34, 1.83it/s, v_num=1, train_loss=197]Epoch 8/400: 2%|█▍ | 7/400 [00:04<03:34, 1.83it/s, v_num=1, train_loss=197]Epoch 8/400: 2%|█▌ | 8/400 [00:04<03:27, 1.88it/s, v_num=1, train_loss=197]Epoch 8/400: 2%|█▌ | 8/400 [00:04<03:27, 1.88it/s, v_num=1, train_loss=196]Epoch 9/400: 2%|█▌ | 8/400 [00:04<03:27, 1.88it/s, v_num=1, train_loss=196]Epoch 9/400: 2%|█▊ | 9/400 [00:05<03:19, 1.96it/s, v_num=1, train_loss=196]Epoch 9/400: 2%|█▊ | 9/400 [00:05<03:19, 1.96it/s, v_num=1, train_loss=195]Epoch 10/400: 2%|█▊ | 9/400 [00:05<03:19, 1.96it/s, v_num=1, train_loss=195]Epoch 10/400: 2%|█▉ | 10/400 [00:05<03:09, 2.06it/s, v_num=1, train_loss=195]Epoch 10/400: 2%|█▉ | 10/400 [00:05<03:09, 2.06it/s, v_num=1, train_loss=194]Epoch 11/400: 2%|█▉ | 10/400 [00:05<03:09, 2.06it/s, v_num=1, train_loss=194]Epoch 11/400: 3%|██▏ | 11/400 [00:06<03:11, 2.03it/s, v_num=1, train_loss=194]Epoch 11/400: 3%|██▏ | 11/400 [00:06<03:11, 2.03it/s, v_num=1, train_loss=193]Epoch 12/400: 3%|██▏ | 11/400 [00:06<03:11, 2.03it/s, v_num=1, train_loss=193]Epoch 12/400: 3%|██▎ | 12/400 [00:06<03:17, 1.97it/s, v_num=1, train_loss=193]Epoch 12/400: 3%|██▎ | 12/400 [00:06<03:17, 1.97it/s, v_num=1, train_loss=192]Epoch 13/400: 3%|██▎ | 12/400 [00:06<03:17, 1.97it/s, v_num=1, train_loss=192]Epoch 13/400: 3%|██▌ | 13/400 [00:07<03:15, 1.98it/s, v_num=1, train_loss=192]Epoch 13/400: 3%|██▌ | 13/400 [00:07<03:15, 1.98it/s, v_num=1, train_loss=192]Epoch 14/400: 3%|██▌ | 13/400 [00:07<03:15, 1.98it/s, v_num=1, train_loss=192]Epoch 14/400: 4%|██▊ | 14/400 [00:07<03:22, 1.91it/s, v_num=1, train_loss=192]Epoch 14/400: 4%|██▊ | 14/400 [00:07<03:22, 1.91it/s, v_num=1, train_loss=191]Epoch 15/400: 4%|██▊ | 14/400 [00:07<03:22, 1.91it/s, v_num=1, train_loss=191]Epoch 15/400: 4%|██▉ | 15/400 [00:08<03:28, 1.85it/s, v_num=1, train_loss=191]Epoch 15/400: 4%|██▉ | 15/400 [00:08<03:28, 1.85it/s, v_num=1, train_loss=190]Epoch 16/400: 4%|██▉ | 15/400 [00:08<03:28, 1.85it/s, v_num=1, train_loss=190]Epoch 16/400: 4%|███▏ | 16/400 [00:08<03:17, 1.95it/s, v_num=1, train_loss=190]Epoch 16/400: 4%|███▏ | 16/400 [00:08<03:17, 1.95it/s, v_num=1, train_loss=190]Epoch 17/400: 4%|███▏ | 16/400 [00:08<03:17, 1.95it/s, v_num=1, train_loss=190]Epoch 17/400: 4%|███▎ | 17/400 [00:09<03:11, 2.00it/s, v_num=1, train_loss=190]Epoch 17/400: 4%|███▎ | 17/400 [00:09<03:11, 2.00it/s, v_num=1, train_loss=189]Epoch 18/400: 4%|███▎ | 17/400 [00:09<03:11, 2.00it/s, v_num=1, train_loss=189]Epoch 18/400: 4%|███▌ | 18/400 [00:09<03:16, 1.94it/s, v_num=1, train_loss=189]Epoch 18/400: 4%|███▌ | 18/400 [00:09<03:16, 1.94it/s, v_num=1, train_loss=189]Epoch 19/400: 4%|███▌ | 18/400 [00:09<03:16, 1.94it/s, v_num=1, train_loss=189]Epoch 19/400: 5%|███▊ | 19/400 [00:10<03:22, 1.89it/s, v_num=1, train_loss=189]Epoch 19/400: 5%|███▊ | 19/400 [00:10<03:22, 1.89it/s, v_num=1, train_loss=188]Epoch 20/400: 5%|███▊ | 19/400 [00:10<03:22, 1.89it/s, v_num=1, train_loss=188]Epoch 20/400: 5%|███▉ | 20/400 [00:10<03:12, 1.97it/s, v_num=1, train_loss=188]Epoch 20/400: 5%|███▉ | 20/400 [00:10<03:12, 1.97it/s, v_num=1, train_loss=188]Epoch 21/400: 5%|███▉ | 20/400 [00:10<03:12, 1.97it/s, v_num=1, train_loss=188]Epoch 21/400: 5%|████▏ | 21/400 [00:11<03:08, 2.01it/s, v_num=1, train_loss=188]Epoch 21/400: 5%|████▏ | 21/400 [00:11<03:08, 2.01it/s, v_num=1, train_loss=187]Epoch 22/400: 5%|████▏ | 21/400 [00:11<03:08, 2.01it/s, v_num=1, train_loss=187]Epoch 22/400: 6%|████▎ | 22/400 [00:11<03:16, 1.93it/s, v_num=1, train_loss=187]Epoch 22/400: 6%|████▎ | 22/400 [00:11<03:16, 1.93it/s, v_num=1, train_loss=187]Epoch 23/400: 6%|████▎ | 22/400 [00:11<03:16, 1.93it/s, v_num=1, train_loss=187]Epoch 23/400: 6%|████▌ | 23/400 [00:12<03:22, 1.86it/s, v_num=1, train_loss=187]Epoch 23/400: 6%|████▌ | 23/400 [00:12<03:22, 1.86it/s, v_num=1, train_loss=186]Epoch 24/400: 6%|████▌ | 23/400 [00:12<03:22, 1.86it/s, v_num=1, train_loss=186]Epoch 24/400: 6%|████▋ | 24/400 [00:13<03:18, 1.90it/s, v_num=1, train_loss=186]Epoch 24/400: 6%|████▋ | 24/400 [00:13<03:18, 1.90it/s, v_num=1, train_loss=186]Epoch 25/400: 6%|████▋ | 24/400 [00:13<03:18, 1.90it/s, v_num=1, train_loss=186]Epoch 25/400: 6%|████▉ | 25/400 [00:13<03:20, 1.87it/s, v_num=1, train_loss=186]Epoch 25/400: 6%|████▉ | 25/400 [00:13<03:20, 1.87it/s, v_num=1, train_loss=186]Epoch 26/400: 6%|████▉ | 25/400 [00:13<03:20, 1.87it/s, v_num=1, train_loss=186]Epoch 26/400: 6%|█████▏ | 26/400 [00:14<03:08, 1.98it/s, v_num=1, train_loss=186]Epoch 26/400: 6%|█████▏ | 26/400 [00:14<03:08, 1.98it/s, v_num=1, train_loss=185]Epoch 27/400: 6%|█████▏ | 26/400 [00:14<03:08, 1.98it/s, v_num=1, train_loss=185]Epoch 27/400: 7%|█████▎ | 27/400 [00:14<03:01, 2.06it/s, v_num=1, train_loss=185]Epoch 27/400: 7%|█████▎ | 27/400 [00:14<03:01, 2.06it/s, v_num=1, train_loss=185]Epoch 28/400: 7%|█████▎ | 27/400 [00:14<03:01, 2.06it/s, v_num=1, train_loss=185]Epoch 28/400: 7%|█████▌ | 28/400 [00:14<03:00, 2.06it/s, v_num=1, train_loss=185]Epoch 28/400: 7%|█████▌ | 28/400 [00:14<03:00, 2.06it/s, v_num=1, train_loss=184]Epoch 29/400: 7%|█████▌ | 28/400 [00:14<03:00, 2.06it/s, v_num=1, train_loss=184]Epoch 29/400: 7%|█████▋ | 29/400 [00:15<02:59, 2.07it/s, v_num=1, train_loss=184]Epoch 29/400: 7%|█████▋ | 29/400 [00:15<02:59, 2.07it/s, v_num=1, train_loss=184]Epoch 30/400: 7%|█████▋ | 29/400 [00:15<02:59, 2.07it/s, v_num=1, train_loss=184]Epoch 30/400: 8%|█████▉ | 30/400 [00:15<02:57, 2.09it/s, v_num=1, train_loss=184]Epoch 30/400: 8%|█████▉ | 30/400 [00:15<02:57, 2.09it/s, v_num=1, train_loss=184]Epoch 31/400: 8%|█████▉ | 30/400 [00:15<02:57, 2.09it/s, v_num=1, train_loss=184]Epoch 31/400: 8%|██████ | 31/400 [00:16<02:57, 2.08it/s, v_num=1, train_loss=184]Epoch 31/400: 8%|██████ | 31/400 [00:16<02:57, 2.08it/s, v_num=1, train_loss=183]Epoch 32/400: 8%|██████ | 31/400 [00:16<02:57, 2.08it/s, v_num=1, train_loss=183]Epoch 32/400: 8%|██████▎ | 32/400 [00:16<02:55, 2.10it/s, v_num=1, train_loss=183]Epoch 32/400: 8%|██████▎ | 32/400 [00:16<02:55, 2.10it/s, v_num=1, train_loss=183]Epoch 33/400: 8%|██████▎ | 32/400 [00:16<02:55, 2.10it/s, v_num=1, train_loss=183]Epoch 33/400: 8%|██████▌ | 33/400 [00:17<02:51, 2.14it/s, v_num=1, train_loss=183]Epoch 33/400: 8%|██████▌ | 33/400 [00:17<02:51, 2.14it/s, v_num=1, train_loss=183]Epoch 34/400: 8%|██████▌ | 33/400 [00:17<02:51, 2.14it/s, v_num=1, train_loss=183]Epoch 34/400: 8%|██████▋ | 34/400 [00:17<02:47, 2.18it/s, v_num=1, train_loss=183]Epoch 34/400: 8%|██████▋ | 34/400 [00:17<02:47, 2.18it/s, v_num=1, train_loss=182]Epoch 35/400: 8%|██████▋ | 34/400 [00:17<02:47, 2.18it/s, v_num=1, train_loss=182]Epoch 35/400: 9%|██████▉ | 35/400 [00:18<03:01, 2.01it/s, v_num=1, train_loss=182]Epoch 35/400: 9%|██████▉ | 35/400 [00:18<03:01, 2.01it/s, v_num=1, train_loss=182]Epoch 36/400: 9%|██████▉ | 35/400 [00:18<03:01, 2.01it/s, v_num=1, train_loss=182]Epoch 36/400: 9%|███████ | 36/400 [00:18<02:58, 2.04it/s, v_num=1, train_loss=182]Epoch 36/400: 9%|███████ | 36/400 [00:18<02:58, 2.04it/s, v_num=1, train_loss=181]Epoch 37/400: 9%|███████ | 36/400 [00:18<02:58, 2.04it/s, v_num=1, train_loss=181]Epoch 37/400: 9%|███████▎ | 37/400 [00:19<03:00, 2.01it/s, v_num=1, train_loss=181]Epoch 37/400: 9%|███████▎ | 37/400 [00:19<03:00, 2.01it/s, v_num=1, train_loss=181]Epoch 38/400: 9%|███████▎ | 37/400 [00:19<03:00, 2.01it/s, v_num=1, train_loss=181]Epoch 38/400: 10%|███████▌ | 38/400 [00:19<03:04, 1.96it/s, v_num=1, train_loss=181]Epoch 38/400: 10%|███████▌ | 38/400 [00:19<03:04, 1.96it/s, v_num=1, train_loss=181]Epoch 39/400: 10%|███████▌ | 38/400 [00:19<03:04, 1.96it/s, v_num=1, train_loss=181]Epoch 39/400: 10%|███████▋ | 39/400 [00:20<02:59, 2.01it/s, v_num=1, train_loss=181]Epoch 39/400: 10%|███████▋ | 39/400 [00:20<02:59, 2.01it/s, v_num=1, train_loss=181]Epoch 40/400: 10%|███████▋ | 39/400 [00:20<02:59, 2.01it/s, v_num=1, train_loss=181]Epoch 40/400: 10%|███████▉ | 40/400 [00:20<02:54, 2.06it/s, v_num=1, train_loss=181]Epoch 40/400: 10%|███████▉ | 40/400 [00:20<02:54, 2.06it/s, v_num=1, train_loss=180]Epoch 41/400: 10%|███████▉ | 40/400 [00:20<02:54, 2.06it/s, v_num=1, train_loss=180]Epoch 41/400: 10%|████████ | 41/400 [00:21<02:50, 2.11it/s, v_num=1, train_loss=180]Epoch 41/400: 10%|████████ | 41/400 [00:21<02:50, 2.11it/s, v_num=1, train_loss=180]Epoch 42/400: 10%|████████ | 41/400 [00:21<02:50, 2.11it/s, v_num=1, train_loss=180]Epoch 42/400: 10%|████████▎ | 42/400 [00:21<02:47, 2.13it/s, v_num=1, train_loss=180]Epoch 42/400: 10%|████████▎ | 42/400 [00:21<02:47, 2.13it/s, v_num=1, train_loss=180]Epoch 43/400: 10%|████████▎ | 42/400 [00:21<02:47, 2.13it/s, v_num=1, train_loss=180]Epoch 43/400: 11%|████████▍ | 43/400 [00:22<02:44, 2.16it/s, v_num=1, train_loss=180]Epoch 43/400: 11%|████████▍ | 43/400 [00:22<02:44, 2.16it/s, v_num=1, train_loss=179]Epoch 44/400: 11%|████████▍ | 43/400 [00:22<02:44, 2.16it/s, v_num=1, train_loss=179]Epoch 44/400: 11%|████████▋ | 44/400 [00:22<02:43, 2.17it/s, v_num=1, train_loss=179]Epoch 44/400: 11%|████████▋ | 44/400 [00:22<02:43, 2.17it/s, v_num=1, train_loss=179]Epoch 45/400: 11%|████████▋ | 44/400 [00:22<02:43, 2.17it/s, v_num=1, train_loss=179]Epoch 45/400: 11%|████████▉ | 45/400 [00:23<02:48, 2.11it/s, v_num=1, train_loss=179]Epoch 45/400: 11%|████████▉ | 45/400 [00:23<02:48, 2.11it/s, v_num=1, train_loss=179]Epoch 46/400: 11%|████████▉ | 45/400 [00:23<02:48, 2.11it/s, v_num=1, train_loss=179]Epoch 46/400: 12%|█████████ | 46/400 [00:23<02:51, 2.07it/s, v_num=1, train_loss=179]Epoch 46/400: 12%|█████████ | 46/400 [00:23<02:51, 2.07it/s, v_num=1, train_loss=179]Epoch 47/400: 12%|█████████ | 46/400 [00:23<02:51, 2.07it/s, v_num=1, train_loss=179]Epoch 47/400: 12%|█████████▎ | 47/400 [00:24<02:55, 2.01it/s, v_num=1, train_loss=179]Epoch 47/400: 12%|█████████▎ | 47/400 [00:24<02:55, 2.01it/s, v_num=1, train_loss=178]Epoch 48/400: 12%|█████████▎ | 47/400 [00:24<02:55, 2.01it/s, v_num=1, train_loss=178]Epoch 48/400: 12%|█████████▍ | 48/400 [00:24<02:57, 1.99it/s, v_num=1, train_loss=178]Epoch 48/400: 12%|█████████▍ | 48/400 [00:24<02:57, 1.99it/s, v_num=1, train_loss=178]Epoch 49/400: 12%|█████████▍ | 48/400 [00:24<02:57, 1.99it/s, v_num=1, train_loss=178]Epoch 49/400: 12%|█████████▋ | 49/400 [00:25<02:49, 2.07it/s, v_num=1, train_loss=178]Epoch 49/400: 12%|█████████▋ | 49/400 [00:25<02:49, 2.07it/s, v_num=1, train_loss=178]Epoch 50/400: 12%|█████████▋ | 49/400 [00:25<02:49, 2.07it/s, v_num=1, train_loss=178]Epoch 50/400: 12%|█████████▉ | 50/400 [00:25<02:47, 2.10it/s, v_num=1, train_loss=178]Epoch 50/400: 12%|█████████▉ | 50/400 [00:25<02:47, 2.10it/s, v_num=1, train_loss=178]Epoch 51/400: 12%|█████████▉ | 50/400 [00:25<02:47, 2.10it/s, v_num=1, train_loss=178]Epoch 51/400: 13%|██████████ | 51/400 [00:26<02:45, 2.11it/s, v_num=1, train_loss=178]Epoch 51/400: 13%|██████████ | 51/400 [00:26<02:45, 2.11it/s, v_num=1, train_loss=177]Epoch 52/400: 13%|██████████ | 51/400 [00:26<02:45, 2.11it/s, v_num=1, train_loss=177]Epoch 52/400: 13%|██████████▎ | 52/400 [00:26<02:42, 2.14it/s, v_num=1, train_loss=177]Epoch 52/400: 13%|██████████▎ | 52/400 [00:26<02:42, 2.14it/s, v_num=1, train_loss=177]Epoch 53/400: 13%|██████████▎ | 52/400 [00:26<02:42, 2.14it/s, v_num=1, train_loss=177]Epoch 53/400: 13%|██████████▍ | 53/400 [00:26<02:42, 2.14it/s, v_num=1, train_loss=177]Epoch 53/400: 13%|██████████▍ | 53/400 [00:26<02:42, 2.14it/s, v_num=1, train_loss=177]Epoch 54/400: 13%|██████████▍ | 53/400 [00:26<02:42, 2.14it/s, v_num=1, train_loss=177]Epoch 54/400: 14%|██████████▋ | 54/400 [00:27<02:37, 2.20it/s, v_num=1, train_loss=177]Epoch 54/400: 14%|██████████▋ | 54/400 [00:27<02:37, 2.20it/s, v_num=1, train_loss=177]Epoch 55/400: 14%|██████████▋ | 54/400 [00:27<02:37, 2.20it/s, v_num=1, train_loss=177]Epoch 55/400: 14%|██████████▊ | 55/400 [00:27<02:36, 2.21it/s, v_num=1, train_loss=177]Epoch 55/400: 14%|██████████▊ | 55/400 [00:27<02:36, 2.21it/s, v_num=1, train_loss=176]Epoch 56/400: 14%|██████████▊ | 55/400 [00:27<02:36, 2.21it/s, v_num=1, train_loss=176]Epoch 56/400: 14%|███████████ | 56/400 [00:28<02:37, 2.19it/s, v_num=1, train_loss=176]Epoch 56/400: 14%|███████████ | 56/400 [00:28<02:37, 2.19it/s, v_num=1, train_loss=176]Epoch 57/400: 14%|███████████ | 56/400 [00:28<02:37, 2.19it/s, v_num=1, train_loss=176]Epoch 57/400: 14%|███████████▎ | 57/400 [00:28<02:36, 2.19it/s, v_num=1, train_loss=176]Epoch 57/400: 14%|███████████▎ | 57/400 [00:28<02:36, 2.19it/s, v_num=1, train_loss=176]Epoch 58/400: 14%|███████████▎ | 57/400 [00:28<02:36, 2.19it/s, v_num=1, train_loss=176]Epoch 58/400: 14%|███████████▍ | 58/400 [00:29<02:34, 2.21it/s, v_num=1, train_loss=176]Epoch 58/400: 14%|███████████▍ | 58/400 [00:29<02:34, 2.21it/s, v_num=1, train_loss=176]Epoch 59/400: 14%|███████████▍ | 58/400 [00:29<02:34, 2.21it/s, v_num=1, train_loss=176]Epoch 59/400: 15%|███████████▋ | 59/400 [00:29<02:34, 2.21it/s, v_num=1, train_loss=176]Epoch 59/400: 15%|███████████▋ | 59/400 [00:29<02:34, 2.21it/s, v_num=1, train_loss=176]Epoch 60/400: 15%|███████████▋ | 59/400 [00:29<02:34, 2.21it/s, v_num=1, train_loss=176]Epoch 60/400: 15%|███████████▊ | 60/400 [00:30<02:32, 2.22it/s, v_num=1, train_loss=176]Epoch 60/400: 15%|███████████▊ | 60/400 [00:30<02:32, 2.22it/s, v_num=1, train_loss=175]Epoch 61/400: 15%|███████████▊ | 60/400 [00:30<02:32, 2.22it/s, v_num=1, train_loss=175]Epoch 61/400: 15%|████████████ | 61/400 [00:30<02:33, 2.21it/s, v_num=1, train_loss=175]Epoch 61/400: 15%|████████████ | 61/400 [00:30<02:33, 2.21it/s, v_num=1, train_loss=175]Epoch 62/400: 15%|████████████ | 61/400 [00:30<02:33, 2.21it/s, v_num=1, train_loss=175]Epoch 62/400: 16%|████████████▏ | 62/400 [00:31<02:35, 2.17it/s, v_num=1, train_loss=175]Epoch 62/400: 16%|████████████▏ | 62/400 [00:31<02:35, 2.17it/s, v_num=1, train_loss=175]Epoch 63/400: 16%|████████████▏ | 62/400 [00:31<02:35, 2.17it/s, v_num=1, train_loss=175]Epoch 63/400: 16%|████████████▍ | 63/400 [00:31<02:33, 2.20it/s, v_num=1, train_loss=175]Epoch 63/400: 16%|████████████▍ | 63/400 [00:31<02:33, 2.20it/s, v_num=1, train_loss=175]Epoch 64/400: 16%|████████████▍ | 63/400 [00:31<02:33, 2.20it/s, v_num=1, train_loss=175]Epoch 64/400: 16%|████████████▋ | 64/400 [00:31<02:34, 2.17it/s, v_num=1, train_loss=175]Epoch 64/400: 16%|████████████▋ | 64/400 [00:31<02:34, 2.17it/s, v_num=1, train_loss=175]Epoch 65/400: 16%|████████████▋ | 64/400 [00:31<02:34, 2.17it/s, v_num=1, train_loss=175]Epoch 65/400: 16%|████████████▊ | 65/400 [00:32<02:36, 2.14it/s, v_num=1, train_loss=175]Epoch 65/400: 16%|████████████▊ | 65/400 [00:32<02:36, 2.14it/s, v_num=1, train_loss=174]Epoch 66/400: 16%|████████████▊ | 65/400 [00:32<02:36, 2.14it/s, v_num=1, train_loss=174]Epoch 66/400: 16%|█████████████ | 66/400 [00:32<02:41, 2.07it/s, v_num=1, train_loss=174]Epoch 66/400: 16%|█████████████ | 66/400 [00:32<02:41, 2.07it/s, v_num=1, train_loss=174]Epoch 67/400: 16%|█████████████ | 66/400 [00:32<02:41, 2.07it/s, v_num=1, train_loss=174]Epoch 67/400: 17%|█████████████▏ | 67/400 [00:33<02:39, 2.08it/s, v_num=1, train_loss=174]Epoch 67/400: 17%|█████████████▏ | 67/400 [00:33<02:39, 2.08it/s, v_num=1, train_loss=174]Epoch 68/400: 17%|█████████████▏ | 67/400 [00:33<02:39, 2.08it/s, v_num=1, train_loss=174]Epoch 68/400: 17%|█████████████▍ | 68/400 [00:33<02:45, 2.00it/s, v_num=1, train_loss=174]Epoch 68/400: 17%|█████████████▍ | 68/400 [00:33<02:45, 2.00it/s, v_num=1, train_loss=174]Epoch 69/400: 17%|█████████████▍ | 68/400 [00:33<02:45, 2.00it/s, v_num=1, train_loss=174]Epoch 69/400: 17%|█████████████▋ | 69/400 [00:34<02:42, 2.04it/s, v_num=1, train_loss=174]Epoch 69/400: 17%|█████████████▋ | 69/400 [00:34<02:42, 2.04it/s, v_num=1, train_loss=174]Epoch 70/400: 17%|█████████████▋ | 69/400 [00:34<02:42, 2.04it/s, v_num=1, train_loss=174]Epoch 70/400: 18%|█████████████▊ | 70/400 [00:34<02:47, 1.97it/s, v_num=1, train_loss=174]Epoch 70/400: 18%|█████████████▊ | 70/400 [00:34<02:47, 1.97it/s, v_num=1, train_loss=174]Epoch 71/400: 18%|█████████████▊ | 70/400 [00:34<02:47, 1.97it/s, v_num=1, train_loss=174]Epoch 71/400: 18%|██████████████ | 71/400 [00:35<02:44, 2.01it/s, v_num=1, train_loss=174]Epoch 71/400: 18%|██████████████ | 71/400 [00:35<02:44, 2.01it/s, v_num=1, train_loss=173]Epoch 72/400: 18%|██████████████ | 71/400 [00:35<02:44, 2.01it/s, v_num=1, train_loss=173]Epoch 72/400: 18%|██████████████▏ | 72/400 [00:35<02:38, 2.07it/s, v_num=1, train_loss=173]Epoch 72/400: 18%|██████████████▏ | 72/400 [00:35<02:38, 2.07it/s, v_num=1, train_loss=173]Epoch 73/400: 18%|██████████████▏ | 72/400 [00:35<02:38, 2.07it/s, v_num=1, train_loss=173]Epoch 73/400: 18%|██████████████▍ | 73/400 [00:36<02:32, 2.15it/s, v_num=1, train_loss=173]Epoch 73/400: 18%|██████████████▍ | 73/400 [00:36<02:32, 2.15it/s, v_num=1, train_loss=173]Epoch 74/400: 18%|██████████████▍ | 73/400 [00:36<02:32, 2.15it/s, v_num=1, train_loss=173]Epoch 74/400: 18%|██████████████▌ | 74/400 [00:36<02:29, 2.19it/s, v_num=1, train_loss=173]Epoch 74/400: 18%|██████████████▌ | 74/400 [00:36<02:29, 2.19it/s, v_num=1, train_loss=173]Epoch 75/400: 18%|██████████████▌ | 74/400 [00:36<02:29, 2.19it/s, v_num=1, train_loss=173]Epoch 75/400: 19%|██████████████▊ | 75/400 [00:37<02:28, 2.19it/s, v_num=1, train_loss=173]Epoch 75/400: 19%|██████████████▊ | 75/400 [00:37<02:28, 2.19it/s, v_num=1, train_loss=173]Epoch 76/400: 19%|██████████████▊ | 75/400 [00:37<02:28, 2.19it/s, v_num=1, train_loss=173]Epoch 76/400: 19%|███████████████ | 76/400 [00:37<02:26, 2.21it/s, v_num=1, train_loss=173]Epoch 76/400: 19%|███████████████ | 76/400 [00:37<02:26, 2.21it/s, v_num=1, train_loss=173]Epoch 77/400: 19%|███████████████ | 76/400 [00:37<02:26, 2.21it/s, v_num=1, train_loss=173]Epoch 77/400: 19%|███████████████▏ | 77/400 [00:38<02:22, 2.27it/s, v_num=1, train_loss=173]Epoch 77/400: 19%|███████████████▏ | 77/400 [00:38<02:22, 2.27it/s, v_num=1, train_loss=173]Epoch 78/400: 19%|███████████████▏ | 77/400 [00:38<02:22, 2.27it/s, v_num=1, train_loss=173]Epoch 78/400: 20%|███████████████▍ | 78/400 [00:38<02:41, 2.00it/s, v_num=1, train_loss=173]Epoch 78/400: 20%|███████████████▍ | 78/400 [00:38<02:41, 2.00it/s, v_num=1, train_loss=172]Epoch 79/400: 20%|███████████████▍ | 78/400 [00:38<02:41, 2.00it/s, v_num=1, train_loss=172]Epoch 79/400: 20%|███████████████▌ | 79/400 [00:39<02:42, 1.97it/s, v_num=1, train_loss=172]Epoch 79/400: 20%|███████████████▌ | 79/400 [00:39<02:42, 1.97it/s, v_num=1, train_loss=172]Epoch 80/400: 20%|███████████████▌ | 79/400 [00:39<02:42, 1.97it/s, v_num=1, train_loss=172]Epoch 80/400: 20%|███████████████▊ | 80/400 [00:39<02:35, 2.06it/s, v_num=1, train_loss=172]Epoch 80/400: 20%|███████████████▊ | 80/400 [00:39<02:35, 2.06it/s, v_num=1, train_loss=172]Epoch 81/400: 20%|███████████████▊ | 80/400 [00:39<02:35, 2.06it/s, v_num=1, train_loss=172]Epoch 81/400: 20%|███████████████▉ | 81/400 [00:40<02:28, 2.15it/s, v_num=1, train_loss=172]Epoch 81/400: 20%|███████████████▉ | 81/400 [00:40<02:28, 2.15it/s, v_num=1, train_loss=172]Epoch 82/400: 20%|███████████████▉ | 81/400 [00:40<02:28, 2.15it/s, v_num=1, train_loss=172]Epoch 82/400: 20%|████████████████▏ | 82/400 [00:40<02:25, 2.19it/s, v_num=1, train_loss=172]Epoch 82/400: 20%|████████████████▏ | 82/400 [00:40<02:25, 2.19it/s, v_num=1, train_loss=172]Epoch 83/400: 20%|████████████████▏ | 82/400 [00:40<02:25, 2.19it/s, v_num=1, train_loss=172]Epoch 83/400: 21%|████████████████▍ | 83/400 [00:40<02:26, 2.17it/s, v_num=1, train_loss=172]Epoch 83/400: 21%|████████████████▍ | 83/400 [00:40<02:26, 2.17it/s, v_num=1, train_loss=172]Epoch 84/400: 21%|████████████████▍ | 83/400 [00:40<02:26, 2.17it/s, v_num=1, train_loss=172]Epoch 84/400: 21%|████████████████▌ | 84/400 [00:41<02:22, 2.21it/s, v_num=1, train_loss=172]Epoch 84/400: 21%|████████████████▌ | 84/400 [00:41<02:22, 2.21it/s, v_num=1, train_loss=171]Epoch 85/400: 21%|████████████████▌ | 84/400 [00:41<02:22, 2.21it/s, v_num=1, train_loss=171]Epoch 85/400: 21%|████████████████▊ | 85/400 [00:41<02:21, 2.22it/s, v_num=1, train_loss=171]Epoch 85/400: 21%|████████████████▊ | 85/400 [00:41<02:21, 2.22it/s, v_num=1, train_loss=171]Epoch 86/400: 21%|████████████████▊ | 85/400 [00:41<02:21, 2.22it/s, v_num=1, train_loss=171]Epoch 86/400: 22%|████████████████▉ | 86/400 [00:42<02:18, 2.27it/s, v_num=1, train_loss=171]Epoch 86/400: 22%|████████████████▉ | 86/400 [00:42<02:18, 2.27it/s, v_num=1, train_loss=171]Epoch 87/400: 22%|████████████████▉ | 86/400 [00:42<02:18, 2.27it/s, v_num=1, train_loss=171]Epoch 87/400: 22%|█████████████████▏ | 87/400 [00:42<02:18, 2.27it/s, v_num=1, train_loss=171]Epoch 87/400: 22%|█████████████████▏ | 87/400 [00:42<02:18, 2.27it/s, v_num=1, train_loss=171]Epoch 88/400: 22%|█████████████████▏ | 87/400 [00:42<02:18, 2.27it/s, v_num=1, train_loss=171]Epoch 88/400: 22%|█████████████████▍ | 88/400 [00:43<02:18, 2.25it/s, v_num=1, train_loss=171]Epoch 88/400: 22%|█████████████████▍ | 88/400 [00:43<02:18, 2.25it/s, v_num=1, train_loss=171]Epoch 89/400: 22%|█████████████████▍ | 88/400 [00:43<02:18, 2.25it/s, v_num=1, train_loss=171]Epoch 89/400: 22%|█████████████████▌ | 89/400 [00:43<02:20, 2.22it/s, v_num=1, train_loss=171]Epoch 89/400: 22%|█████████████████▌ | 89/400 [00:43<02:20, 2.22it/s, v_num=1, train_loss=171]Epoch 90/400: 22%|█████████████████▌ | 89/400 [00:43<02:20, 2.22it/s, v_num=1, train_loss=171]Epoch 90/400: 22%|█████████████████▊ | 90/400 [00:44<02:23, 2.16it/s, v_num=1, train_loss=171]Epoch 90/400: 22%|█████████████████▊ | 90/400 [00:44<02:23, 2.16it/s, v_num=1, train_loss=171]Epoch 91/400: 22%|█████████████████▊ | 90/400 [00:44<02:23, 2.16it/s, v_num=1, train_loss=171]Epoch 91/400: 23%|█████████████████▉ | 91/400 [00:44<02:20, 2.20it/s, v_num=1, train_loss=171]Epoch 91/400: 23%|█████████████████▉ | 91/400 [00:44<02:20, 2.20it/s, v_num=1, train_loss=170]Epoch 92/400: 23%|█████████████████▉ | 91/400 [00:44<02:20, 2.20it/s, v_num=1, train_loss=170]Epoch 92/400: 23%|██████████████████▏ | 92/400 [00:45<02:22, 2.17it/s, v_num=1, train_loss=170]Epoch 92/400: 23%|██████████████████▏ | 92/400 [00:45<02:22, 2.17it/s, v_num=1, train_loss=170]Epoch 93/400: 23%|██████████████████▏ | 92/400 [00:45<02:22, 2.17it/s, v_num=1, train_loss=170]Epoch 93/400: 23%|██████████████████▎ | 93/400 [00:45<02:18, 2.22it/s, v_num=1, train_loss=170]Epoch 93/400: 23%|██████████████████▎ | 93/400 [00:45<02:18, 2.22it/s, v_num=1, train_loss=170]Epoch 94/400: 23%|██████████████████▎ | 93/400 [00:45<02:18, 2.22it/s, v_num=1, train_loss=170]Epoch 94/400: 24%|██████████████████▌ | 94/400 [00:45<02:17, 2.22it/s, v_num=1, train_loss=170]Epoch 94/400: 24%|██████████████████▌ | 94/400 [00:45<02:17, 2.22it/s, v_num=1, train_loss=170]Epoch 95/400: 24%|██████████████████▌ | 94/400 [00:45<02:17, 2.22it/s, v_num=1, train_loss=170]Epoch 95/400: 24%|██████████████████▊ | 95/400 [00:46<02:21, 2.16it/s, v_num=1, train_loss=170]Epoch 95/400: 24%|██████████████████▊ | 95/400 [00:46<02:21, 2.16it/s, v_num=1, train_loss=170]Epoch 96/400: 24%|██████████████████▊ | 95/400 [00:46<02:21, 2.16it/s, v_num=1, train_loss=170]Epoch 96/400: 24%|██████████████████▉ | 96/400 [00:46<02:21, 2.15it/s, v_num=1, train_loss=170]Epoch 96/400: 24%|██████████████████▉ | 96/400 [00:46<02:21, 2.15it/s, v_num=1, train_loss=170]Epoch 97/400: 24%|██████████████████▉ | 96/400 [00:46<02:21, 2.15it/s, v_num=1, train_loss=170]Epoch 97/400: 24%|███████████████████▏ | 97/400 [00:47<02:22, 2.13it/s, v_num=1, train_loss=170]Epoch 97/400: 24%|███████████████████▏ | 97/400 [00:47<02:22, 2.13it/s, v_num=1, train_loss=170]Epoch 98/400: 24%|███████████████████▏ | 97/400 [00:47<02:22, 2.13it/s, v_num=1, train_loss=170]Epoch 98/400: 24%|███████████████████▎ | 98/400 [00:47<02:21, 2.13it/s, v_num=1, train_loss=170]Epoch 98/400: 24%|███████████████████▎ | 98/400 [00:47<02:21, 2.13it/s, v_num=1, train_loss=170]Epoch 99/400: 24%|███████████████████▎ | 98/400 [00:47<02:21, 2.13it/s, v_num=1, train_loss=170]Epoch 99/400: 25%|███████████████████▌ | 99/400 [00:48<02:18, 2.17it/s, v_num=1, train_loss=170]Epoch 99/400: 25%|███████████████████▌ | 99/400 [00:48<02:18, 2.17it/s, v_num=1, train_loss=170]Epoch 100/400: 25%|███████████████████▎ | 99/400 [00:48<02:18, 2.17it/s, v_num=1, train_loss=170]Epoch 100/400: 25%|███████████████████▎ | 100/400 [00:48<02:17, 2.18it/s, v_num=1, train_loss=170]Epoch 100/400: 25%|███████████████████▎ | 100/400 [00:48<02:17, 2.18it/s, v_num=1, train_loss=169]Epoch 101/400: 25%|███████████████████▎ | 100/400 [00:48<02:17, 2.18it/s, v_num=1, train_loss=169]Epoch 101/400: 25%|███████████████████▍ | 101/400 [00:49<02:18, 2.16it/s, v_num=1, train_loss=169]Epoch 101/400: 25%|███████████████████▍ | 101/400 [00:49<02:18, 2.16it/s, v_num=1, train_loss=169]Epoch 102/400: 25%|███████████████████▍ | 101/400 [00:49<02:18, 2.16it/s, v_num=1, train_loss=169]Epoch 102/400: 26%|███████████████████▋ | 102/400 [00:49<02:14, 2.21it/s, v_num=1, train_loss=169]Epoch 102/400: 26%|███████████████████▋ | 102/400 [00:49<02:14, 2.21it/s, v_num=1, train_loss=169]Epoch 103/400: 26%|███████████████████▋ | 102/400 [00:49<02:14, 2.21it/s, v_num=1, train_loss=169]Epoch 103/400: 26%|███████████████████▊ | 103/400 [00:50<02:18, 2.15it/s, v_num=1, train_loss=169]Epoch 103/400: 26%|███████████████████▊ | 103/400 [00:50<02:18, 2.15it/s, v_num=1, train_loss=169]Epoch 104/400: 26%|███████████████████▊ | 103/400 [00:50<02:18, 2.15it/s, v_num=1, train_loss=169]Epoch 104/400: 26%|████████████████████ | 104/400 [00:50<02:14, 2.20it/s, v_num=1, train_loss=169]Epoch 104/400: 26%|████████████████████ | 104/400 [00:50<02:14, 2.20it/s, v_num=1, train_loss=169]Epoch 105/400: 26%|████████████████████ | 104/400 [00:50<02:14, 2.20it/s, v_num=1, train_loss=169]Epoch 105/400: 26%|████████████████████▏ | 105/400 [00:50<02:11, 2.24it/s, v_num=1, train_loss=169]Epoch 105/400: 26%|████████████████████▏ | 105/400 [00:50<02:11, 2.24it/s, v_num=1, train_loss=169]Epoch 106/400: 26%|████████████████████▏ | 105/400 [00:50<02:11, 2.24it/s, v_num=1, train_loss=169]Epoch 106/400: 26%|████████████████████▍ | 106/400 [00:51<02:09, 2.27it/s, v_num=1, train_loss=169]Epoch 106/400: 26%|████████████████████▍ | 106/400 [00:51<02:09, 2.27it/s, v_num=1, train_loss=169]Epoch 107/400: 26%|████████████████████▍ | 106/400 [00:51<02:09, 2.27it/s, v_num=1, train_loss=169]Epoch 107/400: 27%|████████████████████▌ | 107/400 [00:51<02:10, 2.24it/s, v_num=1, train_loss=169]Epoch 107/400: 27%|████████████████████▌ | 107/400 [00:51<02:10, 2.24it/s, v_num=1, train_loss=169]Epoch 108/400: 27%|████████████████████▌ | 107/400 [00:51<02:10, 2.24it/s, v_num=1, train_loss=169]Epoch 108/400: 27%|████████████████████▊ | 108/400 [00:52<02:09, 2.26it/s, v_num=1, train_loss=169]Epoch 108/400: 27%|████████████████████▊ | 108/400 [00:52<02:09, 2.26it/s, v_num=1, train_loss=169]Epoch 109/400: 27%|████████████████████▊ | 108/400 [00:52<02:09, 2.26it/s, v_num=1, train_loss=169]Epoch 109/400: 27%|████████████████████▉ | 109/400 [00:52<02:14, 2.16it/s, v_num=1, train_loss=169]Epoch 109/400: 27%|████████████████████▉ | 109/400 [00:52<02:14, 2.16it/s, v_num=1, train_loss=168]Epoch 110/400: 27%|████████████████████▉ | 109/400 [00:52<02:14, 2.16it/s, v_num=1, train_loss=168]Epoch 110/400: 28%|█████████████████████▏ | 110/400 [00:53<02:24, 2.00it/s, v_num=1, train_loss=168]Epoch 110/400: 28%|█████████████████████▏ | 110/400 [00:53<02:24, 2.00it/s, v_num=1, train_loss=168]Epoch 111/400: 28%|█████████████████████▏ | 110/400 [00:53<02:24, 2.00it/s, v_num=1, train_loss=168]Epoch 111/400: 28%|█████████████████████▎ | 111/400 [00:53<02:28, 1.95it/s, v_num=1, train_loss=168]Epoch 111/400: 28%|█████████████████████▎ | 111/400 [00:53<02:28, 1.95it/s, v_num=1, train_loss=168]Epoch 112/400: 28%|█████████████████████▎ | 111/400 [00:53<02:28, 1.95it/s, v_num=1, train_loss=168]Epoch 112/400: 28%|█████████████████████▌ | 112/400 [00:54<02:32, 1.89it/s, v_num=1, train_loss=168]Epoch 112/400: 28%|█████████████████████▌ | 112/400 [00:54<02:32, 1.89it/s, v_num=1, train_loss=168]Epoch 113/400: 28%|█████████████████████▌ | 112/400 [00:54<02:32, 1.89it/s, v_num=1, train_loss=168]Epoch 113/400: 28%|█████████████████████▊ | 113/400 [00:54<02:24, 1.98it/s, v_num=1, train_loss=168]Epoch 113/400: 28%|█████████████████████▊ | 113/400 [00:54<02:24, 1.98it/s, v_num=1, train_loss=168]Epoch 114/400: 28%|█████████████████████▊ | 113/400 [00:54<02:24, 1.98it/s, v_num=1, train_loss=168]Epoch 114/400: 28%|█████████████████████▉ | 114/400 [00:55<02:20, 2.03it/s, v_num=1, train_loss=168]Epoch 114/400: 28%|█████████████████████▉ | 114/400 [00:55<02:20, 2.03it/s, v_num=1, train_loss=168]Epoch 115/400: 28%|█████████████████████▉ | 114/400 [00:55<02:20, 2.03it/s, v_num=1, train_loss=168]Epoch 115/400: 29%|██████████████████████▏ | 115/400 [00:55<02:18, 2.06it/s, v_num=1, train_loss=168]Epoch 115/400: 29%|██████████████████████▏ | 115/400 [00:55<02:18, 2.06it/s, v_num=1, train_loss=168]Epoch 116/400: 29%|██████████████████████▏ | 115/400 [00:55<02:18, 2.06it/s, v_num=1, train_loss=168]Epoch 116/400: 29%|██████████████████████▎ | 116/400 [00:56<02:24, 1.97it/s, v_num=1, train_loss=168]Epoch 116/400: 29%|██████████████████████▎ | 116/400 [00:56<02:24, 1.97it/s, v_num=1, train_loss=168]Epoch 117/400: 29%|██████████████████████▎ | 116/400 [00:56<02:24, 1.97it/s, v_num=1, train_loss=168]Epoch 117/400: 29%|██████████████████████▌ | 117/400 [00:56<02:24, 1.95it/s, v_num=1, train_loss=168]Epoch 117/400: 29%|██████████████████████▌ | 117/400 [00:56<02:24, 1.95it/s, v_num=1, train_loss=168]Epoch 118/400: 29%|██████████████████████▌ | 117/400 [00:56<02:24, 1.95it/s, v_num=1, train_loss=168]Epoch 118/400: 30%|██████████████████████▋ | 118/400 [00:57<02:27, 1.91it/s, v_num=1, train_loss=168]Epoch 118/400: 30%|██████████████████████▋ | 118/400 [00:57<02:27, 1.91it/s, v_num=1, train_loss=168]Epoch 119/400: 30%|██████████████████████▋ | 118/400 [00:57<02:27, 1.91it/s, v_num=1, train_loss=168]Epoch 119/400: 30%|██████████████████████▉ | 119/400 [00:58<02:24, 1.95it/s, v_num=1, train_loss=168]Epoch 119/400: 30%|██████████████████████▉ | 119/400 [00:58<02:24, 1.95it/s, v_num=1, train_loss=167]Epoch 120/400: 30%|██████████████████████▉ | 119/400 [00:58<02:24, 1.95it/s, v_num=1, train_loss=167]Epoch 120/400: 30%|███████████████████████ | 120/400 [00:58<02:17, 2.04it/s, v_num=1, train_loss=167]Epoch 120/400: 30%|███████████████████████ | 120/400 [00:58<02:17, 2.04it/s, v_num=1, train_loss=167]Epoch 121/400: 30%|███████████████████████ | 120/400 [00:58<02:17, 2.04it/s, v_num=1, train_loss=167]Epoch 121/400: 30%|███████████████████████▎ | 121/400 [00:59<02:25, 1.91it/s, v_num=1, train_loss=167]Epoch 121/400: 30%|███████████████████████▎ | 121/400 [00:59<02:25, 1.91it/s, v_num=1, train_loss=168]Epoch 122/400: 30%|███████████████████████▎ | 121/400 [00:59<02:25, 1.91it/s, v_num=1, train_loss=168]Epoch 122/400: 30%|███████████████████████▍ | 122/400 [00:59<02:34, 1.80it/s, v_num=1, train_loss=168]Epoch 122/400: 30%|███████████████████████▍ | 122/400 [00:59<02:34, 1.80it/s, v_num=1, train_loss=167]Epoch 123/400: 30%|███████████████████████▍ | 122/400 [00:59<02:34, 1.80it/s, v_num=1, train_loss=167]Epoch 123/400: 31%|███████████████████████▋ | 123/400 [01:00<02:33, 1.81it/s, v_num=1, train_loss=167]Epoch 123/400: 31%|███████████████████████▋ | 123/400 [01:00<02:33, 1.81it/s, v_num=1, train_loss=167]Epoch 124/400: 31%|███████████████████████▋ | 123/400 [01:00<02:33, 1.81it/s, v_num=1, train_loss=167]Epoch 124/400: 31%|███████████████████████▊ | 124/400 [01:00<02:28, 1.86it/s, v_num=1, train_loss=167]Epoch 124/400: 31%|███████████████████████▊ | 124/400 [01:00<02:28, 1.86it/s, v_num=1, train_loss=167]Epoch 125/400: 31%|███████████████████████▊ | 124/400 [01:00<02:28, 1.86it/s, v_num=1, train_loss=167]Epoch 125/400: 31%|████████████████████████ | 125/400 [01:01<02:22, 1.93it/s, v_num=1, train_loss=167]Epoch 125/400: 31%|████████████████████████ | 125/400 [01:01<02:22, 1.93it/s, v_num=1, train_loss=167]Epoch 126/400: 31%|████████████████████████ | 125/400 [01:01<02:22, 1.93it/s, v_num=1, train_loss=167]Epoch 126/400: 32%|████████████████████████▎ | 126/400 [01:01<02:19, 1.97it/s, v_num=1, train_loss=167]Epoch 126/400: 32%|████████████████████████▎ | 126/400 [01:01<02:19, 1.97it/s, v_num=1, train_loss=167]Epoch 127/400: 32%|████████████████████████▎ | 126/400 [01:01<02:19, 1.97it/s, v_num=1, train_loss=167]Epoch 127/400: 32%|████████████████████████▍ | 127/400 [01:02<02:13, 2.04it/s, v_num=1, train_loss=167]Epoch 127/400: 32%|████████████████████████▍ | 127/400 [01:02<02:13, 2.04it/s, v_num=1, train_loss=167]Epoch 128/400: 32%|████████████████████████▍ | 127/400 [01:02<02:13, 2.04it/s, v_num=1, train_loss=167]Epoch 128/400: 32%|████████████████████████▋ | 128/400 [01:02<02:09, 2.10it/s, v_num=1, train_loss=167]Epoch 128/400: 32%|████████████████████████▋ | 128/400 [01:02<02:09, 2.10it/s, v_num=1, train_loss=167]Epoch 129/400: 32%|████████████████████████▋ | 128/400 [01:02<02:09, 2.10it/s, v_num=1, train_loss=167]Epoch 129/400: 32%|████████████████████████▊ | 129/400 [01:03<02:06, 2.14it/s, v_num=1, train_loss=167]Epoch 129/400: 32%|████████████████████████▊ | 129/400 [01:03<02:06, 2.14it/s, v_num=1, train_loss=167]Epoch 130/400: 32%|████████████████████████▊ | 129/400 [01:03<02:06, 2.14it/s, v_num=1, train_loss=167]Epoch 130/400: 32%|█████████████████████████ | 130/400 [01:03<02:06, 2.13it/s, v_num=1, train_loss=167]Epoch 130/400: 32%|█████████████████████████ | 130/400 [01:03<02:06, 2.13it/s, v_num=1, train_loss=167]Epoch 131/400: 32%|█████████████████████████ | 130/400 [01:03<02:06, 2.13it/s, v_num=1, train_loss=167]Epoch 131/400: 33%|█████████████████████████▏ | 131/400 [01:03<02:06, 2.13it/s, v_num=1, train_loss=167]Epoch 131/400: 33%|█████████████████████████▏ | 131/400 [01:03<02:06, 2.13it/s, v_num=1, train_loss=167]Epoch 132/400: 33%|█████████████████████████▏ | 131/400 [01:03<02:06, 2.13it/s, v_num=1, train_loss=167]Epoch 132/400: 33%|█████████████████████████▍ | 132/400 [01:04<02:04, 2.16it/s, v_num=1, train_loss=167]Epoch 132/400: 33%|█████████████████████████▍ | 132/400 [01:04<02:04, 2.16it/s, v_num=1, train_loss=167]Epoch 133/400: 33%|█████████████████████████▍ | 132/400 [01:04<02:04, 2.16it/s, v_num=1, train_loss=167]Epoch 133/400: 33%|█████████████████████████▌ | 133/400 [01:04<02:07, 2.09it/s, v_num=1, train_loss=167]Epoch 133/400: 33%|█████████████████████████▌ | 133/400 [01:04<02:07, 2.09it/s, v_num=1, train_loss=166]Epoch 134/400: 33%|█████████████████████████▌ | 133/400 [01:04<02:07, 2.09it/s, v_num=1, train_loss=166]Epoch 134/400: 34%|█████████████████████████▊ | 134/400 [01:05<02:03, 2.15it/s, v_num=1, train_loss=166]Epoch 134/400: 34%|█████████████████████████▊ | 134/400 [01:05<02:03, 2.15it/s, v_num=1, train_loss=166]Epoch 135/400: 34%|█████████████████████████▊ | 134/400 [01:05<02:03, 2.15it/s, v_num=1, train_loss=166]Epoch 135/400: 34%|█████████████████████████▉ | 135/400 [01:05<01:59, 2.22it/s, v_num=1, train_loss=166]Epoch 135/400: 34%|█████████████████████████▉ | 135/400 [01:05<01:59, 2.22it/s, v_num=1, train_loss=166]Epoch 136/400: 34%|█████████████████████████▉ | 135/400 [01:05<01:59, 2.22it/s, v_num=1, train_loss=166]Epoch 136/400: 34%|██████████████████████████▏ | 136/400 [01:06<01:57, 2.25it/s, v_num=1, train_loss=166]Epoch 136/400: 34%|██████████████████████████▏ | 136/400 [01:06<01:57, 2.25it/s, v_num=1, train_loss=166]Epoch 137/400: 34%|██████████████████████████▏ | 136/400 [01:06<01:57, 2.25it/s, v_num=1, train_loss=166]Epoch 137/400: 34%|██████████████████████████▎ | 137/400 [01:06<01:55, 2.28it/s, v_num=1, train_loss=166]Epoch 137/400: 34%|██████████████████████████▎ | 137/400 [01:06<01:55, 2.28it/s, v_num=1, train_loss=166]Epoch 138/400: 34%|██████████████████████████▎ | 137/400 [01:06<01:55, 2.28it/s, v_num=1, train_loss=166]Epoch 138/400: 34%|██████████████████████████▌ | 138/400 [01:07<01:58, 2.21it/s, v_num=1, train_loss=166]Epoch 138/400: 34%|██████████████████████████▌ | 138/400 [01:07<01:58, 2.21it/s, v_num=1, train_loss=166]Epoch 139/400: 34%|██████████████████████████▌ | 138/400 [01:07<01:58, 2.21it/s, v_num=1, train_loss=166]Epoch 139/400: 35%|██████████████████████████▊ | 139/400 [01:07<02:02, 2.13it/s, v_num=1, train_loss=166]Epoch 139/400: 35%|██████████████████████████▊ | 139/400 [01:07<02:02, 2.13it/s, v_num=1, train_loss=166]Epoch 140/400: 35%|██████████████████████████▊ | 139/400 [01:07<02:02, 2.13it/s, v_num=1, train_loss=166]Epoch 140/400: 35%|██████████████████████████▉ | 140/400 [01:08<02:03, 2.10it/s, v_num=1, train_loss=166]Epoch 140/400: 35%|██████████████████████████▉ | 140/400 [01:08<02:03, 2.10it/s, v_num=1, train_loss=166]Epoch 141/400: 35%|██████████████████████████▉ | 140/400 [01:08<02:03, 2.10it/s, v_num=1, train_loss=166]Epoch 141/400: 35%|███████████████████████████▏ | 141/400 [01:08<02:00, 2.15it/s, v_num=1, train_loss=166]Epoch 141/400: 35%|███████████████████████████▏ | 141/400 [01:08<02:00, 2.15it/s, v_num=1, train_loss=166]Epoch 142/400: 35%|███████████████████████████▏ | 141/400 [01:08<02:00, 2.15it/s, v_num=1, train_loss=166]Epoch 142/400: 36%|███████████████████████████▎ | 142/400 [01:09<02:02, 2.11it/s, v_num=1, train_loss=166]Epoch 142/400: 36%|███████████████████████████▎ | 142/400 [01:09<02:02, 2.11it/s, v_num=1, train_loss=166]Epoch 143/400: 36%|███████████████████████████▎ | 142/400 [01:09<02:02, 2.11it/s, v_num=1, train_loss=166]Epoch 143/400: 36%|███████████████████████████▌ | 143/400 [01:09<02:05, 2.04it/s, v_num=1, train_loss=166]Epoch 143/400: 36%|███████████████████████████▌ | 143/400 [01:09<02:05, 2.04it/s, v_num=1, train_loss=166]Epoch 144/400: 36%|███████████████████████████▌ | 143/400 [01:09<02:05, 2.04it/s, v_num=1, train_loss=166]Epoch 144/400: 36%|███████████████████████████▋ | 144/400 [01:10<02:07, 2.01it/s, v_num=1, train_loss=166]Epoch 144/400: 36%|███████████████████████████▋ | 144/400 [01:10<02:07, 2.01it/s, v_num=1, train_loss=166]Epoch 145/400: 36%|███████████████████████████▋ | 144/400 [01:10<02:07, 2.01it/s, v_num=1, train_loss=166]Epoch 145/400: 36%|███████████████████████████▉ | 145/400 [01:10<02:04, 2.05it/s, v_num=1, train_loss=166]Epoch 145/400: 36%|███████████████████████████▉ | 145/400 [01:10<02:04, 2.05it/s, v_num=1, train_loss=166]Epoch 146/400: 36%|███████████████████████████▉ | 145/400 [01:10<02:04, 2.05it/s, v_num=1, train_loss=166]Epoch 146/400: 36%|████████████████████████████ | 146/400 [01:10<02:00, 2.11it/s, v_num=1, train_loss=166]Epoch 146/400: 36%|████████████████████████████ | 146/400 [01:11<02:00, 2.11it/s, v_num=1, train_loss=166]Epoch 147/400: 36%|████████████████████████████ | 146/400 [01:11<02:00, 2.11it/s, v_num=1, train_loss=166]Epoch 147/400: 37%|████████████████████████████▎ | 147/400 [01:11<01:56, 2.17it/s, v_num=1, train_loss=166]Epoch 147/400: 37%|████████████████████████████▎ | 147/400 [01:11<01:56, 2.17it/s, v_num=1, train_loss=166]Epoch 148/400: 37%|████████████████████████████▎ | 147/400 [01:11<01:56, 2.17it/s, v_num=1, train_loss=166]Epoch 148/400: 37%|████████████████████████████▍ | 148/400 [01:11<01:54, 2.20it/s, v_num=1, train_loss=166]Epoch 148/400: 37%|████████████████████████████▍ | 148/400 [01:11<01:54, 2.20it/s, v_num=1, train_loss=166]Epoch 149/400: 37%|████████████████████████████▍ | 148/400 [01:11<01:54, 2.20it/s, v_num=1, train_loss=166]Epoch 149/400: 37%|████████████████████████████▋ | 149/400 [01:12<01:53, 2.22it/s, v_num=1, train_loss=166]Epoch 149/400: 37%|████████████████████████████▋ | 149/400 [01:12<01:53, 2.22it/s, v_num=1, train_loss=166]Epoch 150/400: 37%|████████████████████████████▋ | 149/400 [01:12<01:53, 2.22it/s, v_num=1, train_loss=166]Epoch 150/400: 38%|████████████████████████████▉ | 150/400 [01:12<01:50, 2.26it/s, v_num=1, train_loss=166]Epoch 150/400: 38%|████████████████████████████▉ | 150/400 [01:12<01:50, 2.26it/s, v_num=1, train_loss=166]Epoch 151/400: 38%|████████████████████████████▉ | 150/400 [01:12<01:50, 2.26it/s, v_num=1, train_loss=166]Epoch 151/400: 38%|█████████████████████████████ | 151/400 [01:13<01:57, 2.12it/s, v_num=1, train_loss=166]Epoch 151/400: 38%|█████████████████████████████ | 151/400 [01:13<01:57, 2.12it/s, v_num=1, train_loss=166]Epoch 152/400: 38%|█████████████████████████████ | 151/400 [01:13<01:57, 2.12it/s, v_num=1, train_loss=166]Epoch 152/400: 38%|█████████████████████████████▎ | 152/400 [01:13<02:03, 2.00it/s, v_num=1, train_loss=166]Epoch 152/400: 38%|█████████████████████████████▎ | 152/400 [01:13<02:03, 2.00it/s, v_num=1, train_loss=165]Epoch 153/400: 38%|█████████████████████████████▎ | 152/400 [01:13<02:03, 2.00it/s, v_num=1, train_loss=165]Epoch 153/400: 38%|█████████████████████████████▍ | 153/400 [01:14<02:01, 2.03it/s, v_num=1, train_loss=165]Epoch 153/400: 38%|█████████████████████████████▍ | 153/400 [01:14<02:01, 2.03it/s, v_num=1, train_loss=166]Epoch 154/400: 38%|█████████████████████████████▍ | 153/400 [01:14<02:01, 2.03it/s, v_num=1, train_loss=166]Epoch 154/400: 38%|█████████████████████████████▋ | 154/400 [01:14<01:59, 2.06it/s, v_num=1, train_loss=166]Epoch 154/400: 38%|█████████████████████████████▋ | 154/400 [01:14<01:59, 2.06it/s, v_num=1, train_loss=165]Epoch 155/400: 38%|█████████████████████████████▋ | 154/400 [01:14<01:59, 2.06it/s, v_num=1, train_loss=165]Epoch 155/400: 39%|█████████████████████████████▊ | 155/400 [01:15<01:57, 2.09it/s, v_num=1, train_loss=165]Epoch 155/400: 39%|█████████████████████████████▊ | 155/400 [01:15<01:57, 2.09it/s, v_num=1, train_loss=165]Epoch 156/400: 39%|█████████████████████████████▊ | 155/400 [01:15<01:57, 2.09it/s, v_num=1, train_loss=165]Epoch 156/400: 39%|██████████████████████████████ | 156/400 [01:15<01:58, 2.05it/s, v_num=1, train_loss=165]Epoch 156/400: 39%|██████████████████████████████ | 156/400 [01:15<01:58, 2.05it/s, v_num=1, train_loss=165]Epoch 157/400: 39%|██████████████████████████████ | 156/400 [01:15<01:58, 2.05it/s, v_num=1, train_loss=165]Epoch 157/400: 39%|██████████████████████████████▏ | 157/400 [01:16<02:00, 2.02it/s, v_num=1, train_loss=165]Epoch 157/400: 39%|██████████████████████████████▏ | 157/400 [01:16<02:00, 2.02it/s, v_num=1, train_loss=165]Epoch 158/400: 39%|██████████████████████████████▏ | 157/400 [01:16<02:00, 2.02it/s, v_num=1, train_loss=165]Epoch 158/400: 40%|██████████████████████████████▍ | 158/400 [01:16<01:56, 2.08it/s, v_num=1, train_loss=165]Epoch 158/400: 40%|██████████████████████████████▍ | 158/400 [01:16<01:56, 2.08it/s, v_num=1, train_loss=165]Epoch 159/400: 40%|██████████████████████████████▍ | 158/400 [01:16<01:56, 2.08it/s, v_num=1, train_loss=165]Epoch 159/400: 40%|██████████████████████████████▌ | 159/400 [01:17<01:51, 2.16it/s, v_num=1, train_loss=165]Epoch 159/400: 40%|██████████████████████████████▌ | 159/400 [01:17<01:51, 2.16it/s, v_num=1, train_loss=165]Epoch 160/400: 40%|██████████████████████████████▌ | 159/400 [01:17<01:51, 2.16it/s, v_num=1, train_loss=165]Epoch 160/400: 40%|██████████████████████████████▊ | 160/400 [01:17<01:51, 2.16it/s, v_num=1, train_loss=165]Epoch 160/400: 40%|██████████████████████████████▊ | 160/400 [01:17<01:51, 2.16it/s, v_num=1, train_loss=165]Epoch 161/400: 40%|██████████████████████████████▊ | 160/400 [01:17<01:51, 2.16it/s, v_num=1, train_loss=165]Epoch 161/400: 40%|██████████████████████████████▉ | 161/400 [01:18<01:53, 2.11it/s, v_num=1, train_loss=165]Epoch 161/400: 40%|██████████████████████████████▉ | 161/400 [01:18<01:53, 2.11it/s, v_num=1, train_loss=165]Epoch 162/400: 40%|██████████████████████████████▉ | 161/400 [01:18<01:53, 2.11it/s, v_num=1, train_loss=165]Epoch 162/400: 40%|███████████████████████████████▏ | 162/400 [01:18<01:49, 2.18it/s, v_num=1, train_loss=165]Epoch 162/400: 40%|███████████████████████████████▏ | 162/400 [01:18<01:49, 2.18it/s, v_num=1, train_loss=165]Epoch 163/400: 40%|███████████████████████████████▏ | 162/400 [01:18<01:49, 2.18it/s, v_num=1, train_loss=165]Epoch 163/400: 41%|███████████████████████████████▍ | 163/400 [01:19<01:54, 2.07it/s, v_num=1, train_loss=165]Epoch 163/400: 41%|███████████████████████████████▍ | 163/400 [01:19<01:54, 2.07it/s, v_num=1, train_loss=165]Epoch 164/400: 41%|███████████████████████████████▍ | 163/400 [01:19<01:54, 2.07it/s, v_num=1, train_loss=165]Epoch 164/400: 41%|███████████████████████████████▌ | 164/400 [01:19<01:51, 2.12it/s, v_num=1, train_loss=165]Epoch 164/400: 41%|███████████████████████████████▌ | 164/400 [01:19<01:51, 2.12it/s, v_num=1, train_loss=165]Epoch 165/400: 41%|███████████████████████████████▌ | 164/400 [01:19<01:51, 2.12it/s, v_num=1, train_loss=165]Epoch 165/400: 41%|███████████████████████████████▊ | 165/400 [01:20<01:56, 2.01it/s, v_num=1, train_loss=165]Epoch 165/400: 41%|███████████████████████████████▊ | 165/400 [01:20<01:56, 2.01it/s, v_num=1, train_loss=165]Epoch 166/400: 41%|███████████████████████████████▊ | 165/400 [01:20<01:56, 2.01it/s, v_num=1, train_loss=165]Epoch 166/400: 42%|███████████████████████████████▉ | 166/400 [01:20<01:52, 2.08it/s, v_num=1, train_loss=165]Epoch 166/400: 42%|███████████████████████████████▉ | 166/400 [01:20<01:52, 2.08it/s, v_num=1, train_loss=165]Epoch 167/400: 42%|███████████████████████████████▉ | 166/400 [01:20<01:52, 2.08it/s, v_num=1, train_loss=165]Epoch 167/400: 42%|████████████████████████████████▏ | 167/400 [01:20<01:49, 2.13it/s, v_num=1, train_loss=165]Epoch 167/400: 42%|████████████████████████████████▏ | 167/400 [01:20<01:49, 2.13it/s, v_num=1, train_loss=165]Epoch 168/400: 42%|████████████████████████████████▏ | 167/400 [01:20<01:49, 2.13it/s, v_num=1, train_loss=165]Epoch 168/400: 42%|████████████████████████████████▎ | 168/400 [01:21<01:46, 2.17it/s, v_num=1, train_loss=165]Epoch 168/400: 42%|████████████████████████████████▎ | 168/400 [01:21<01:46, 2.17it/s, v_num=1, train_loss=165]Epoch 169/400: 42%|████████████████████████████████▎ | 168/400 [01:21<01:46, 2.17it/s, v_num=1, train_loss=165]Epoch 169/400: 42%|████████████████████████████████▌ | 169/400 [01:21<01:43, 2.23it/s, v_num=1, train_loss=165]Epoch 169/400: 42%|████████████████████████████████▌ | 169/400 [01:21<01:43, 2.23it/s, v_num=1, train_loss=165]Epoch 170/400: 42%|████████████████████████████████▌ | 169/400 [01:21<01:43, 2.23it/s, v_num=1, train_loss=165]Epoch 170/400: 42%|████████████████████████████████▋ | 170/400 [01:22<01:46, 2.16it/s, v_num=1, train_loss=165]Epoch 170/400: 42%|████████████████████████████████▋ | 170/400 [01:22<01:46, 2.16it/s, v_num=1, train_loss=165]Epoch 171/400: 42%|████████████████████████████████▋ | 170/400 [01:22<01:46, 2.16it/s, v_num=1, train_loss=165]Epoch 171/400: 43%|████████████████████████████████▉ | 171/400 [01:22<01:44, 2.19it/s, v_num=1, train_loss=165]Epoch 171/400: 43%|████████████████████████████████▉ | 171/400 [01:22<01:44, 2.19it/s, v_num=1, train_loss=165]Epoch 172/400: 43%|████████████████████████████████▉ | 171/400 [01:22<01:44, 2.19it/s, v_num=1, train_loss=165]Epoch 172/400: 43%|█████████████████████████████████ | 172/400 [01:23<01:42, 2.22it/s, v_num=1, train_loss=165]Epoch 172/400: 43%|█████████████████████████████████ | 172/400 [01:23<01:42, 2.22it/s, v_num=1, train_loss=165]Epoch 173/400: 43%|█████████████████████████████████ | 172/400 [01:23<01:42, 2.22it/s, v_num=1, train_loss=165]Epoch 173/400: 43%|█████████████████████████████████▎ | 173/400 [01:23<01:41, 2.25it/s, v_num=1, train_loss=165]Epoch 173/400: 43%|█████████████████████████████████▎ | 173/400 [01:23<01:41, 2.25it/s, v_num=1, train_loss=165]Epoch 174/400: 43%|█████████████████████████████████▎ | 173/400 [01:23<01:41, 2.25it/s, v_num=1, train_loss=165]Epoch 174/400: 44%|█████████████████████████████████▍ | 174/400 [01:24<01:42, 2.21it/s, v_num=1, train_loss=165]Epoch 174/400: 44%|█████████████████████████████████▍ | 174/400 [01:24<01:42, 2.21it/s, v_num=1, train_loss=165]Epoch 175/400: 44%|█████████████████████████████████▍ | 174/400 [01:24<01:42, 2.21it/s, v_num=1, train_loss=165]Epoch 175/400: 44%|█████████████████████████████████▋ | 175/400 [01:24<01:45, 2.13it/s, v_num=1, train_loss=165]Epoch 175/400: 44%|█████████████████████████████████▋ | 175/400 [01:24<01:45, 2.13it/s, v_num=1, train_loss=165]Epoch 176/400: 44%|█████████████████████████████████▋ | 175/400 [01:24<01:45, 2.13it/s, v_num=1, train_loss=165]Epoch 176/400: 44%|█████████████████████████████████▉ | 176/400 [01:25<01:41, 2.20it/s, v_num=1, train_loss=165]Epoch 176/400: 44%|█████████████████████████████████▉ | 176/400 [01:25<01:41, 2.20it/s, v_num=1, train_loss=165]Epoch 177/400: 44%|█████████████████████████████████▉ | 176/400 [01:25<01:41, 2.20it/s, v_num=1, train_loss=165]Epoch 177/400: 44%|██████████████████████████████████ | 177/400 [01:25<01:45, 2.11it/s, v_num=1, train_loss=165]Epoch 177/400: 44%|██████████████████████████████████ | 177/400 [01:25<01:45, 2.11it/s, v_num=1, train_loss=165]Epoch 178/400: 44%|██████████████████████████████████ | 177/400 [01:25<01:45, 2.11it/s, v_num=1, train_loss=165]Epoch 178/400: 44%|██████████████████████████████████▎ | 178/400 [01:26<01:45, 2.11it/s, v_num=1, train_loss=165]Epoch 178/400: 44%|██████████████████████████████████▎ | 178/400 [01:26<01:45, 2.11it/s, v_num=1, train_loss=164]Epoch 179/400: 44%|██████████████████████████████████▎ | 178/400 [01:26<01:45, 2.11it/s, v_num=1, train_loss=164]Epoch 179/400: 45%|██████████████████████████████████▍ | 179/400 [01:26<01:45, 2.09it/s, v_num=1, train_loss=164]Epoch 179/400: 45%|██████████████████████████████████▍ | 179/400 [01:26<01:45, 2.09it/s, v_num=1, train_loss=165]Epoch 180/400: 45%|██████████████████████████████████▍ | 179/400 [01:26<01:45, 2.09it/s, v_num=1, train_loss=165]Epoch 180/400: 45%|██████████████████████████████████▋ | 180/400 [01:26<01:43, 2.12it/s, v_num=1, train_loss=165]Epoch 180/400: 45%|██████████████████████████████████▋ | 180/400 [01:26<01:43, 2.12it/s, v_num=1, train_loss=165]Epoch 181/400: 45%|██████████████████████████████████▋ | 180/400 [01:26<01:43, 2.12it/s, v_num=1, train_loss=165]Epoch 181/400: 45%|██████████████████████████████████▊ | 181/400 [01:27<01:42, 2.14it/s, v_num=1, train_loss=165]Epoch 181/400: 45%|██████████████████████████████████▊ | 181/400 [01:27<01:42, 2.14it/s, v_num=1, train_loss=164]Epoch 182/400: 45%|██████████████████████████████████▊ | 181/400 [01:27<01:42, 2.14it/s, v_num=1, train_loss=164]Epoch 182/400: 46%|███████████████████████████████████ | 182/400 [01:27<01:40, 2.18it/s, v_num=1, train_loss=164]Epoch 182/400: 46%|███████████████████████████████████ | 182/400 [01:27<01:40, 2.18it/s, v_num=1, train_loss=165]Epoch 183/400: 46%|███████████████████████████████████ | 182/400 [01:27<01:40, 2.18it/s, v_num=1, train_loss=165]Epoch 183/400: 46%|███████████████████████████████████▏ | 183/400 [01:28<01:41, 2.14it/s, v_num=1, train_loss=165]Epoch 183/400: 46%|███████████████████████████████████▏ | 183/400 [01:28<01:41, 2.14it/s, v_num=1, train_loss=164]Epoch 184/400: 46%|███████████████████████████████████▏ | 183/400 [01:28<01:41, 2.14it/s, v_num=1, train_loss=164]Epoch 184/400: 46%|███████████████████████████████████▍ | 184/400 [01:28<01:46, 2.04it/s, v_num=1, train_loss=164]Epoch 184/400: 46%|███████████████████████████████████▍ | 184/400 [01:28<01:46, 2.04it/s, v_num=1, train_loss=164]Epoch 185/400: 46%|███████████████████████████████████▍ | 184/400 [01:28<01:46, 2.04it/s, v_num=1, train_loss=164]Epoch 185/400: 46%|███████████████████████████████████▌ | 185/400 [01:29<01:47, 2.01it/s, v_num=1, train_loss=164]Epoch 185/400: 46%|███████████████████████████████████▌ | 185/400 [01:29<01:47, 2.01it/s, v_num=1, train_loss=164]Epoch 186/400: 46%|███████████████████████████████████▌ | 185/400 [01:29<01:47, 2.01it/s, v_num=1, train_loss=164]Epoch 186/400: 46%|███████████████████████████████████▊ | 186/400 [01:29<01:49, 1.96it/s, v_num=1, train_loss=164]Epoch 186/400: 46%|███████████████████████████████████▊ | 186/400 [01:29<01:49, 1.96it/s, v_num=1, train_loss=164]Epoch 187/400: 46%|███████████████████████████████████▊ | 186/400 [01:29<01:49, 1.96it/s, v_num=1, train_loss=164]Epoch 187/400: 47%|███████████████████████████████████▉ | 187/400 [01:30<01:46, 1.99it/s, v_num=1, train_loss=164]Epoch 187/400: 47%|███████████████████████████████████▉ | 187/400 [01:30<01:46, 1.99it/s, v_num=1, train_loss=164]Epoch 188/400: 47%|███████████████████████████████████▉ | 187/400 [01:30<01:46, 1.99it/s, v_num=1, train_loss=164]Epoch 188/400: 47%|████████████████████████████████████▏ | 188/400 [01:30<01:46, 1.99it/s, v_num=1, train_loss=164]Epoch 188/400: 47%|████████████████████████████████████▏ | 188/400 [01:30<01:46, 1.99it/s, v_num=1, train_loss=164]Epoch 189/400: 47%|████████████████████████████████████▏ | 188/400 [01:30<01:46, 1.99it/s, v_num=1, train_loss=164]Epoch 189/400: 47%|████████████████████████████████████▍ | 189/400 [01:31<01:43, 2.03it/s, v_num=1, train_loss=164]Epoch 189/400: 47%|████████████████████████████████████▍ | 189/400 [01:31<01:43, 2.03it/s, v_num=1, train_loss=164]Epoch 190/400: 47%|████████████████████████████████████▍ | 189/400 [01:31<01:43, 2.03it/s, v_num=1, train_loss=164]Epoch 190/400: 48%|████████████████████████████████████▌ | 190/400 [01:31<01:39, 2.11it/s, v_num=1, train_loss=164]Epoch 190/400: 48%|████████████████████████████████████▌ | 190/400 [01:31<01:39, 2.11it/s, v_num=1, train_loss=164]Epoch 191/400: 48%|████████████████████████████████████▌ | 190/400 [01:31<01:39, 2.11it/s, v_num=1, train_loss=164]Epoch 191/400: 48%|████████████████████████████████████▊ | 191/400 [01:32<01:37, 2.14it/s, v_num=1, train_loss=164]Epoch 191/400: 48%|████████████████████████████████████▊ | 191/400 [01:32<01:37, 2.14it/s, v_num=1, train_loss=164]Epoch 192/400: 48%|████████████████████████████████████▊ | 191/400 [01:32<01:37, 2.14it/s, v_num=1, train_loss=164]Epoch 192/400: 48%|████████████████████████████████████▉ | 192/400 [01:32<01:34, 2.20it/s, v_num=1, train_loss=164]Epoch 192/400: 48%|████████████████████████████████████▉ | 192/400 [01:32<01:34, 2.20it/s, v_num=1, train_loss=164]Epoch 193/400: 48%|████████████████████████████████████▉ | 192/400 [01:32<01:34, 2.20it/s, v_num=1, train_loss=164]Epoch 193/400: 48%|█████████████████████████████████████▏ | 193/400 [01:33<01:34, 2.20it/s, v_num=1, train_loss=164]Epoch 193/400: 48%|█████████████████████████████████████▏ | 193/400 [01:33<01:34, 2.20it/s, v_num=1, train_loss=164]Epoch 194/400: 48%|█████████████████████████████████████▏ | 193/400 [01:33<01:34, 2.20it/s, v_num=1, train_loss=164]Epoch 194/400: 48%|█████████████████████████████████████▎ | 194/400 [01:33<01:32, 2.23it/s, v_num=1, train_loss=164]Epoch 194/400: 48%|█████████████████████████████████████▎ | 194/400 [01:33<01:32, 2.23it/s, v_num=1, train_loss=164]Epoch 195/400: 48%|█████████████████████████████████████▎ | 194/400 [01:33<01:32, 2.23it/s, v_num=1, train_loss=164]Epoch 195/400: 49%|█████████████████████████████████████▌ | 195/400 [01:34<01:34, 2.18it/s, v_num=1, train_loss=164]Epoch 195/400: 49%|█████████████████████████████████████▌ | 195/400 [01:34<01:34, 2.18it/s, v_num=1, train_loss=164]Epoch 196/400: 49%|█████████████████████████████████████▌ | 195/400 [01:34<01:34, 2.18it/s, v_num=1, train_loss=164]Epoch 196/400: 49%|█████████████████████████████████████▋ | 196/400 [01:34<01:33, 2.19it/s, v_num=1, train_loss=164]Epoch 196/400: 49%|█████████████████████████████████████▋ | 196/400 [01:34<01:33, 2.19it/s, v_num=1, train_loss=164]Epoch 197/400: 49%|█████████████████████████████████████▋ | 196/400 [01:34<01:33, 2.19it/s, v_num=1, train_loss=164]Epoch 197/400: 49%|█████████████████████████████████████▉ | 197/400 [01:34<01:32, 2.20it/s, v_num=1, train_loss=164]Epoch 197/400: 49%|█████████████████████████████████████▉ | 197/400 [01:34<01:32, 2.20it/s, v_num=1, train_loss=164]Epoch 198/400: 49%|█████████████████████████████████████▉ | 197/400 [01:34<01:32, 2.20it/s, v_num=1, train_loss=164]Epoch 198/400: 50%|██████████████████████████████████████ | 198/400 [01:35<01:30, 2.23it/s, v_num=1, train_loss=164]Epoch 198/400: 50%|██████████████████████████████████████ | 198/400 [01:35<01:30, 2.23it/s, v_num=1, train_loss=164]Epoch 199/400: 50%|██████████████████████████████████████ | 198/400 [01:35<01:30, 2.23it/s, v_num=1, train_loss=164]Epoch 199/400: 50%|██████████████████████████████████████▎ | 199/400 [01:35<01:29, 2.25it/s, v_num=1, train_loss=164]Epoch 199/400: 50%|██████████████████████████████████████▎ | 199/400 [01:35<01:29, 2.25it/s, v_num=1, train_loss=164]Epoch 200/400: 50%|██████████████████████████████████████▎ | 199/400 [01:35<01:29, 2.25it/s, v_num=1, train_loss=164]Epoch 200/400: 50%|██████████████████████████████████████▌ | 200/400 [01:36<01:33, 2.13it/s, v_num=1, train_loss=164]Epoch 200/400: 50%|██████████████████████████████████████▌ | 200/400 [01:36<01:33, 2.13it/s, v_num=1, train_loss=164]Epoch 201/400: 50%|██████████████████████████████████████▌ | 200/400 [01:36<01:33, 2.13it/s, v_num=1, train_loss=164]Epoch 201/400: 50%|██████████████████████████████████████▋ | 201/400 [01:36<01:33, 2.13it/s, v_num=1, train_loss=164]Epoch 201/400: 50%|██████████████████████████████████████▋ | 201/400 [01:36<01:33, 2.13it/s, v_num=1, train_loss=164]Epoch 202/400: 50%|██████████████████████████████████████▋ | 201/400 [01:36<01:33, 2.13it/s, v_num=1, train_loss=164]Epoch 202/400: 50%|██████████████████████████████████████▉ | 202/400 [01:37<01:33, 2.12it/s, v_num=1, train_loss=164]Epoch 202/400: 50%|██████████████████████████████████████▉ | 202/400 [01:37<01:33, 2.12it/s, v_num=1, train_loss=164]Epoch 203/400: 50%|██████████████████████████████████████▉ | 202/400 [01:37<01:33, 2.12it/s, v_num=1, train_loss=164]Epoch 203/400: 51%|███████████████████████████████████████ | 203/400 [01:37<01:34, 2.10it/s, v_num=1, train_loss=164]Epoch 203/400: 51%|███████████████████████████████████████ | 203/400 [01:37<01:34, 2.10it/s, v_num=1, train_loss=164]Epoch 204/400: 51%|███████████████████████████████████████ | 203/400 [01:37<01:34, 2.10it/s, v_num=1, train_loss=164]Epoch 204/400: 51%|███████████████████████████████████████▎ | 204/400 [01:38<01:35, 2.05it/s, v_num=1, train_loss=164]Epoch 204/400: 51%|███████████████████████████████████████▎ | 204/400 [01:38<01:35, 2.05it/s, v_num=1, train_loss=164]Epoch 205/400: 51%|███████████████████████████████████████▎ | 204/400 [01:38<01:35, 2.05it/s, v_num=1, train_loss=164]Epoch 205/400: 51%|███████████████████████████████████████▍ | 205/400 [01:38<01:33, 2.10it/s, v_num=1, train_loss=164]Epoch 205/400: 51%|███████████████████████████████████████▍ | 205/400 [01:38<01:33, 2.10it/s, v_num=1, train_loss=164]Epoch 206/400: 51%|███████████████████████████████████████▍ | 205/400 [01:38<01:33, 2.10it/s, v_num=1, train_loss=164]Epoch 206/400: 52%|███████████████████████████████████████▋ | 206/400 [01:39<01:32, 2.10it/s, v_num=1, train_loss=164]Epoch 206/400: 52%|███████████████████████████████████████▋ | 206/400 [01:39<01:32, 2.10it/s, v_num=1, train_loss=164]Epoch 207/400: 52%|███████████████████████████████████████▋ | 206/400 [01:39<01:32, 2.10it/s, v_num=1, train_loss=164]Epoch 207/400: 52%|███████████████████████████████████████▊ | 207/400 [01:39<01:30, 2.14it/s, v_num=1, train_loss=164]Epoch 207/400: 52%|███████████████████████████████████████▊ | 207/400 [01:39<01:30, 2.14it/s, v_num=1, train_loss=164]Epoch 208/400: 52%|███████████████████████████████████████▊ | 207/400 [01:39<01:30, 2.14it/s, v_num=1, train_loss=164]Epoch 208/400: 52%|████████████████████████████████████████ | 208/400 [01:40<01:27, 2.19it/s, v_num=1, train_loss=164]Epoch 208/400: 52%|████████████████████████████████████████ | 208/400 [01:40<01:27, 2.19it/s, v_num=1, train_loss=164]Epoch 209/400: 52%|████████████████████████████████████████ | 208/400 [01:40<01:27, 2.19it/s, v_num=1, train_loss=164]Epoch 209/400: 52%|████████████████████████████████████████▏ | 209/400 [01:40<01:29, 2.13it/s, v_num=1, train_loss=164]Epoch 209/400: 52%|████████████████████████████████████████▏ | 209/400 [01:40<01:29, 2.13it/s, v_num=1, train_loss=164]Epoch 210/400: 52%|████████████████████████████████████████▏ | 209/400 [01:40<01:29, 2.13it/s, v_num=1, train_loss=164]Epoch 210/400: 52%|████████████████████████████████████████▍ | 210/400 [01:41<01:31, 2.07it/s, v_num=1, train_loss=164]Epoch 210/400: 52%|████████████████████████████████████████▍ | 210/400 [01:41<01:31, 2.07it/s, v_num=1, train_loss=164]Epoch 211/400: 52%|████████████████████████████████████████▍ | 210/400 [01:41<01:31, 2.07it/s, v_num=1, train_loss=164]Epoch 211/400: 53%|████████████████████████████████████████▌ | 211/400 [01:41<01:33, 2.02it/s, v_num=1, train_loss=164]Epoch 211/400: 53%|████████████████████████████████████████▌ | 211/400 [01:41<01:33, 2.02it/s, v_num=1, train_loss=164]Epoch 212/400: 53%|████████████████████████████████████████▌ | 211/400 [01:41<01:33, 2.02it/s, v_num=1, train_loss=164]Epoch 212/400: 53%|████████████████████████████████████████▊ | 212/400 [01:42<01:29, 2.10it/s, v_num=1, train_loss=164]Epoch 212/400: 53%|████████████████████████████████████████▊ | 212/400 [01:42<01:29, 2.10it/s, v_num=1, train_loss=164]Epoch 213/400: 53%|████████████████████████████████████████▊ | 212/400 [01:42<01:29, 2.10it/s, v_num=1, train_loss=164]Epoch 213/400: 53%|█████████████████████████████████████████ | 213/400 [01:42<01:27, 2.14it/s, v_num=1, train_loss=164]Epoch 213/400: 53%|█████████████████████████████████████████ | 213/400 [01:42<01:27, 2.14it/s, v_num=1, train_loss=164]Epoch 214/400: 53%|█████████████████████████████████████████ | 213/400 [01:42<01:27, 2.14it/s, v_num=1, train_loss=164]Epoch 214/400: 54%|█████████████████████████████████████████▏ | 214/400 [01:42<01:25, 2.18it/s, v_num=1, train_loss=164]Epoch 214/400: 54%|█████████████████████████████████████████▏ | 214/400 [01:42<01:25, 2.18it/s, v_num=1, train_loss=164]Epoch 215/400: 54%|█████████████████████████████████████████▏ | 214/400 [01:42<01:25, 2.18it/s, v_num=1, train_loss=164]Epoch 215/400: 54%|█████████████████████████████████████████▍ | 215/400 [01:43<01:24, 2.20it/s, v_num=1, train_loss=164]Epoch 215/400: 54%|█████████████████████████████████████████▍ | 215/400 [01:43<01:24, 2.20it/s, v_num=1, train_loss=164]Epoch 216/400: 54%|█████████████████████████████████████████▍ | 215/400 [01:43<01:24, 2.20it/s, v_num=1, train_loss=164]Epoch 216/400: 54%|█████████████████████████████████████████▌ | 216/400 [01:43<01:22, 2.24it/s, v_num=1, train_loss=164]Epoch 216/400: 54%|█████████████████████████████████████████▌ | 216/400 [01:43<01:22, 2.24it/s, v_num=1, train_loss=164]Epoch 217/400: 54%|█████████████████████████████████████████▌ | 216/400 [01:43<01:22, 2.24it/s, v_num=1, train_loss=164]Epoch 217/400: 54%|█████████████████████████████████████████▊ | 217/400 [01:44<01:21, 2.24it/s, v_num=1, train_loss=164]Epoch 217/400: 54%|█████████████████████████████████████████▊ | 217/400 [01:44<01:21, 2.24it/s, v_num=1, train_loss=164]Epoch 218/400: 54%|█████████████████████████████████████████▊ | 217/400 [01:44<01:21, 2.24it/s, v_num=1, train_loss=164]Epoch 218/400: 55%|█████████████████████████████████████████▉ | 218/400 [01:44<01:20, 2.26it/s, v_num=1, train_loss=164]Epoch 218/400: 55%|█████████████████████████████████████████▉ | 218/400 [01:44<01:20, 2.26it/s, v_num=1, train_loss=164]Epoch 219/400: 55%|█████████████████████████████████████████▉ | 218/400 [01:44<01:20, 2.26it/s, v_num=1, train_loss=164]Epoch 219/400: 55%|██████████████████████████████████████████▏ | 219/400 [01:45<01:20, 2.25it/s, v_num=1, train_loss=164]Epoch 219/400: 55%|██████████████████████████████████████████▏ | 219/400 [01:45<01:20, 2.25it/s, v_num=1, train_loss=164]Epoch 220/400: 55%|██████████████████████████████████████████▏ | 219/400 [01:45<01:20, 2.25it/s, v_num=1, train_loss=164]Epoch 220/400: 55%|██████████████████████████████████████████▎ | 220/400 [01:45<01:20, 2.25it/s, v_num=1, train_loss=164]Epoch 220/400: 55%|██████████████████████████████████████████▎ | 220/400 [01:45<01:20, 2.25it/s, v_num=1, train_loss=164]Epoch 221/400: 55%|██████████████████████████████████████████▎ | 220/400 [01:45<01:20, 2.25it/s, v_num=1, train_loss=164]Epoch 221/400: 55%|██████████████████████████████████████████▌ | 221/400 [01:46<01:21, 2.20it/s, v_num=1, train_loss=164]Epoch 221/400: 55%|██████████████████████████████████████████▌ | 221/400 [01:46<01:21, 2.20it/s, v_num=1, train_loss=164]Epoch 222/400: 55%|██████████████████████████████████████████▌ | 221/400 [01:46<01:21, 2.20it/s, v_num=1, train_loss=164]Epoch 222/400: 56%|██████████████████████████████████████████▋ | 222/400 [01:46<01:24, 2.11it/s, v_num=1, train_loss=164]Epoch 222/400: 56%|██████████████████████████████████████████▋ | 222/400 [01:46<01:24, 2.11it/s, v_num=1, train_loss=164]Epoch 223/400: 56%|██████████████████████████████████████████▋ | 222/400 [01:46<01:24, 2.11it/s, v_num=1, train_loss=164]Epoch 223/400: 56%|██████████████████████████████████████████▉ | 223/400 [01:47<01:22, 2.14it/s, v_num=1, train_loss=164]Epoch 223/400: 56%|██████████████████████████████████████████▉ | 223/400 [01:47<01:22, 2.14it/s, v_num=1, train_loss=164]Epoch 224/400: 56%|██████████████████████████████████████████▉ | 223/400 [01:47<01:22, 2.14it/s, v_num=1, train_loss=164]Epoch 224/400: 56%|███████████████████████████████████████████ | 224/400 [01:47<01:22, 2.13it/s, v_num=1, train_loss=164]Epoch 224/400: 56%|███████████████████████████████████████████ | 224/400 [01:47<01:22, 2.13it/s, v_num=1, train_loss=164]Epoch 225/400: 56%|███████████████████████████████████████████ | 224/400 [01:47<01:22, 2.13it/s, v_num=1, train_loss=164]Epoch 225/400: 56%|███████████████████████████████████████████▎ | 225/400 [01:47<01:19, 2.19it/s, v_num=1, train_loss=164]Epoch 225/400: 56%|███████████████████████████████████████████▎ | 225/400 [01:47<01:19, 2.19it/s, v_num=1, train_loss=164]Epoch 226/400: 56%|███████████████████████████████████████████▎ | 225/400 [01:47<01:19, 2.19it/s, v_num=1, train_loss=164]Epoch 226/400: 56%|███████████████████████████████████████████▌ | 226/400 [01:48<01:18, 2.21it/s, v_num=1, train_loss=164]Epoch 226/400: 56%|███████████████████████████████████████████▌ | 226/400 [01:48<01:18, 2.21it/s, v_num=1, train_loss=164]Epoch 227/400: 56%|███████████████████████████████████████████▌ | 226/400 [01:48<01:18, 2.21it/s, v_num=1, train_loss=164]Epoch 227/400: 57%|███████████████████████████████████████████▋ | 227/400 [01:48<01:17, 2.23it/s, v_num=1, train_loss=164]Epoch 227/400: 57%|███████████████████████████████████████████▋ | 227/400 [01:48<01:17, 2.23it/s, v_num=1, train_loss=164]Epoch 228/400: 57%|███████████████████████████████████████████▋ | 227/400 [01:48<01:17, 2.23it/s, v_num=1, train_loss=164]Epoch 228/400: 57%|███████████████████████████████████████████▉ | 228/400 [01:49<01:17, 2.23it/s, v_num=1, train_loss=164]Epoch 228/400: 57%|███████████████████████████████████████████▉ | 228/400 [01:49<01:17, 2.23it/s, v_num=1, train_loss=164]Epoch 229/400: 57%|███████████████████████████████████████████▉ | 228/400 [01:49<01:17, 2.23it/s, v_num=1, train_loss=164]Epoch 229/400: 57%|████████████████████████████████████████████ | 229/400 [01:49<01:15, 2.27it/s, v_num=1, train_loss=164]Epoch 229/400: 57%|████████████████████████████████████████████ | 229/400 [01:49<01:15, 2.27it/s, v_num=1, train_loss=164]Epoch 230/400: 57%|████████████████████████████████████████████ | 229/400 [01:49<01:15, 2.27it/s, v_num=1, train_loss=164]Epoch 230/400: 57%|████████████████████████████████████████████▎ | 230/400 [01:50<01:17, 2.19it/s, v_num=1, train_loss=164]Epoch 230/400: 57%|████████████████████████████████████████████▎ | 230/400 [01:50<01:17, 2.19it/s, v_num=1, train_loss=164]Epoch 231/400: 57%|████████████████████████████████████████████▎ | 230/400 [01:50<01:17, 2.19it/s, v_num=1, train_loss=164]Epoch 231/400: 58%|████████████████████████████████████████████▍ | 231/400 [01:50<01:17, 2.19it/s, v_num=1, train_loss=164]Epoch 231/400: 58%|████████████████████████████████████████████▍ | 231/400 [01:50<01:17, 2.19it/s, v_num=1, train_loss=164]Epoch 232/400: 58%|████████████████████████████████████████████▍ | 231/400 [01:50<01:17, 2.19it/s, v_num=1, train_loss=164]Epoch 232/400: 58%|████████████████████████████████████████████▋ | 232/400 [01:51<01:15, 2.23it/s, v_num=1, train_loss=164]Epoch 232/400: 58%|████████████████████████████████████████████▋ | 232/400 [01:51<01:15, 2.23it/s, v_num=1, train_loss=164]Epoch 233/400: 58%|████████████████████████████████████████████▋ | 232/400 [01:51<01:15, 2.23it/s, v_num=1, train_loss=164]Epoch 233/400: 58%|████████████████████████████████████████████▊ | 233/400 [01:51<01:17, 2.15it/s, v_num=1, train_loss=164]Epoch 233/400: 58%|████████████████████████████████████████████▊ | 233/400 [01:51<01:17, 2.15it/s, v_num=1, train_loss=164]Epoch 234/400: 58%|████████████████████████████████████████████▊ | 233/400 [01:51<01:17, 2.15it/s, v_num=1, train_loss=164]Epoch 234/400: 58%|█████████████████████████████████████████████ | 234/400 [01:52<01:17, 2.14it/s, v_num=1, train_loss=164]Epoch 234/400: 58%|█████████████████████████████████████████████ | 234/400 [01:52<01:17, 2.14it/s, v_num=1, train_loss=164]Epoch 235/400: 58%|█████████████████████████████████████████████ | 234/400 [01:52<01:17, 2.14it/s, v_num=1, train_loss=164]Epoch 235/400: 59%|█████████████████████████████████████████████▏ | 235/400 [01:52<01:17, 2.13it/s, v_num=1, train_loss=164]Epoch 235/400: 59%|█████████████████████████████████████████████▏ | 235/400 [01:52<01:17, 2.13it/s, v_num=1, train_loss=164]Epoch 236/400: 59%|█████████████████████████████████████████████▏ | 235/400 [01:52<01:17, 2.13it/s, v_num=1, train_loss=164]Epoch 236/400: 59%|█████████████████████████████████████████████▍ | 236/400 [01:53<01:16, 2.16it/s, v_num=1, train_loss=164]Epoch 236/400: 59%|█████████████████████████████████████████████▍ | 236/400 [01:53<01:16, 2.16it/s, v_num=1, train_loss=164]Epoch 237/400: 59%|█████████████████████████████████████████████▍ | 236/400 [01:53<01:16, 2.16it/s, v_num=1, train_loss=164]Epoch 237/400: 59%|█████████████████████████████████████████████▌ | 237/400 [01:53<01:22, 1.98it/s, v_num=1, train_loss=164]Epoch 237/400: 59%|█████████████████████████████████████████████▌ | 237/400 [01:53<01:22, 1.98it/s, v_num=1, train_loss=164]Epoch 238/400: 59%|█████████████████████████████████████████████▌ | 237/400 [01:53<01:22, 1.98it/s, v_num=1, train_loss=164]Epoch 238/400: 60%|█████████████████████████████████████████████▊ | 238/400 [01:54<01:24, 1.93it/s, v_num=1, train_loss=164]Epoch 238/400: 60%|█████████████████████████████████████████████▊ | 238/400 [01:54<01:24, 1.93it/s, v_num=1, train_loss=164]Epoch 239/400: 60%|█████████████████████████████████████████████▊ | 238/400 [01:54<01:24, 1.93it/s, v_num=1, train_loss=164]Epoch 239/400: 60%|██████████████████████████████████████████████ | 239/400 [01:54<01:23, 1.93it/s, v_num=1, train_loss=164]Epoch 239/400: 60%|██████████████████████████████████████████████ | 239/400 [01:54<01:23, 1.93it/s, v_num=1, train_loss=164]Epoch 240/400: 60%|██████████████████████████████████████████████ | 239/400 [01:54<01:23, 1.93it/s, v_num=1, train_loss=164]Epoch 240/400: 60%|██████████████████████████████████████████████▏ | 240/400 [01:55<01:22, 1.94it/s, v_num=1, train_loss=164]Epoch 240/400: 60%|██████████████████████████████████████████████▏ | 240/400 [01:55<01:22, 1.94it/s, v_num=1, train_loss=164]Epoch 241/400: 60%|██████████████████████████████████████████████▏ | 240/400 [01:55<01:22, 1.94it/s, v_num=1, train_loss=164]Epoch 241/400: 60%|██████████████████████████████████████████████▍ | 241/400 [01:55<01:20, 1.98it/s, v_num=1, train_loss=164]Epoch 241/400: 60%|██████████████████████████████████████████████▍ | 241/400 [01:55<01:20, 1.98it/s, v_num=1, train_loss=164]Epoch 242/400: 60%|██████████████████████████████████████████████▍ | 241/400 [01:55<01:20, 1.98it/s, v_num=1, train_loss=164]Epoch 242/400: 60%|██████████████████████████████████████████████▌ | 242/400 [01:56<01:16, 2.07it/s, v_num=1, train_loss=164]Epoch 242/400: 60%|██████████████████████████████████████████████▌ | 242/400 [01:56<01:16, 2.07it/s, v_num=1, train_loss=164]Epoch 243/400: 60%|██████████████████████████████████████████████▌ | 242/400 [01:56<01:16, 2.07it/s, v_num=1, train_loss=164]Epoch 243/400: 61%|██████████████████████████████████████████████▊ | 243/400 [01:56<01:14, 2.12it/s, v_num=1, train_loss=164]Epoch 243/400: 61%|██████████████████████████████████████████████▊ | 243/400 [01:56<01:14, 2.12it/s, v_num=1, train_loss=164]Epoch 244/400: 61%|██████████████████████████████████████████████▊ | 243/400 [01:56<01:14, 2.12it/s, v_num=1, train_loss=164]Epoch 244/400: 61%|██████████████████████████████████████████████▉ | 244/400 [01:56<01:12, 2.15it/s, v_num=1, train_loss=164]Epoch 244/400: 61%|██████████████████████████████████████████████▉ | 244/400 [01:56<01:12, 2.15it/s, v_num=1, train_loss=164]Epoch 245/400: 61%|██████████████████████████████████████████████▉ | 244/400 [01:56<01:12, 2.15it/s, v_num=1, train_loss=164]Epoch 245/400: 61%|███████████████████████████████████████████████▏ | 245/400 [01:57<01:12, 2.14it/s, v_num=1, train_loss=164]Epoch 245/400: 61%|███████████████████████████████████████████████▏ | 245/400 [01:57<01:12, 2.14it/s, v_num=1, train_loss=164]Epoch 246/400: 61%|███████████████████████████████████████████████▏ | 245/400 [01:57<01:12, 2.14it/s, v_num=1, train_loss=164]Epoch 246/400: 62%|███████████████████████████████████████████████▎ | 246/400 [01:57<01:09, 2.20it/s, v_num=1, train_loss=164]Epoch 246/400: 62%|███████████████████████████████████████████████▎ | 246/400 [01:57<01:09, 2.20it/s, v_num=1, train_loss=164]Epoch 247/400: 62%|███████████████████████████████████████████████▎ | 246/400 [01:57<01:09, 2.20it/s, v_num=1, train_loss=164]Epoch 247/400: 62%|███████████████████████████████████████████████▌ | 247/400 [01:58<01:07, 2.25it/s, v_num=1, train_loss=164]Epoch 247/400: 62%|███████████████████████████████████████████████▌ | 247/400 [01:58<01:07, 2.25it/s, v_num=1, train_loss=164]Epoch 248/400: 62%|███████████████████████████████████████████████▌ | 247/400 [01:58<01:07, 2.25it/s, v_num=1, train_loss=164]Epoch 248/400: 62%|███████████████████████████████████████████████▋ | 248/400 [01:58<01:06, 2.27it/s, v_num=1, train_loss=164]Epoch 248/400: 62%|███████████████████████████████████████████████▋ | 248/400 [01:58<01:06, 2.27it/s, v_num=1, train_loss=164]Epoch 249/400: 62%|███████████████████████████████████████████████▋ | 248/400 [01:58<01:06, 2.27it/s, v_num=1, train_loss=164]Epoch 249/400: 62%|███████████████████████████████████████████████▉ | 249/400 [01:59<01:06, 2.29it/s, v_num=1, train_loss=164]Epoch 249/400: 62%|███████████████████████████████████████████████▉ | 249/400 [01:59<01:06, 2.29it/s, v_num=1, train_loss=164]Epoch 250/400: 62%|███████████████████████████████████████████████▉ | 249/400 [01:59<01:06, 2.29it/s, v_num=1, train_loss=164]Epoch 250/400: 62%|████████████████████████████████████████████████▏ | 250/400 [01:59<01:05, 2.28it/s, v_num=1, train_loss=164]Epoch 250/400: 62%|████████████████████████████████████████████████▏ | 250/400 [01:59<01:05, 2.28it/s, v_num=1, train_loss=164]Epoch 251/400: 62%|████████████████████████████████████████████████▏ | 250/400 [01:59<01:05, 2.28it/s, v_num=1, train_loss=164]Epoch 251/400: 63%|████████████████████████████████████████████████▎ | 251/400 [02:00<01:06, 2.26it/s, v_num=1, train_loss=164]Epoch 251/400: 63%|████████████████████████████████████████████████▎ | 251/400 [02:00<01:06, 2.26it/s, v_num=1, train_loss=164]Epoch 252/400: 63%|████████████████████████████████████████████████▎ | 251/400 [02:00<01:06, 2.26it/s, v_num=1, train_loss=164]Epoch 252/400: 63%|████████████████████████████████████████████████▌ | 252/400 [02:00<01:07, 2.19it/s, v_num=1, train_loss=164]Epoch 252/400: 63%|████████████████████████████████████████████████▌ | 252/400 [02:00<01:07, 2.19it/s, v_num=1, train_loss=164]Epoch 253/400: 63%|████████████████████████████████████████████████▌ | 252/400 [02:00<01:07, 2.19it/s, v_num=1, train_loss=164]Epoch 253/400: 63%|████████████████████████████████████████████████▋ | 253/400 [02:01<01:08, 2.16it/s, v_num=1, train_loss=164]Epoch 253/400: 63%|████████████████████████████████████████████████▋ | 253/400 [02:01<01:08, 2.16it/s, v_num=1, train_loss=164]Epoch 254/400: 63%|████████████████████████████████████████████████▋ | 253/400 [02:01<01:08, 2.16it/s, v_num=1, train_loss=164]Epoch 254/400: 64%|████████████████████████████████████████████████▉ | 254/400 [02:01<01:08, 2.13it/s, v_num=1, train_loss=164]Epoch 254/400: 64%|████████████████████████████████████████████████▉ | 254/400 [02:01<01:08, 2.13it/s, v_num=1, train_loss=164]Epoch 255/400: 64%|████████████████████████████████████████████████▉ | 254/400 [02:01<01:08, 2.13it/s, v_num=1, train_loss=164]Epoch 255/400: 64%|█████████████████████████████████████████████████ | 255/400 [02:01<01:08, 2.12it/s, v_num=1, train_loss=164]Epoch 255/400: 64%|█████████████████████████████████████████████████ | 255/400 [02:01<01:08, 2.12it/s, v_num=1, train_loss=164]Epoch 256/400: 64%|█████████████████████████████████████████████████ | 255/400 [02:01<01:08, 2.12it/s, v_num=1, train_loss=164]Epoch 256/400: 64%|█████████████████████████████████████████████████▎ | 256/400 [02:02<01:07, 2.13it/s, v_num=1, train_loss=164]Epoch 256/400: 64%|█████████████████████████████████████████████████▎ | 256/400 [02:02<01:07, 2.13it/s, v_num=1, train_loss=163]Epoch 257/400: 64%|█████████████████████████████████████████████████▎ | 256/400 [02:02<01:07, 2.13it/s, v_num=1, train_loss=163]Epoch 257/400: 64%|█████████████████████████████████████████████████▍ | 257/400 [02:02<01:07, 2.13it/s, v_num=1, train_loss=163]Epoch 257/400: 64%|█████████████████████████████████████████████████▍ | 257/400 [02:02<01:07, 2.13it/s, v_num=1, train_loss=164]Epoch 258/400: 64%|█████████████████████████████████████████████████▍ | 257/400 [02:02<01:07, 2.13it/s, v_num=1, train_loss=164]Epoch 258/400: 64%|█████████████████████████████████████████████████▋ | 258/400 [02:03<01:05, 2.18it/s, v_num=1, train_loss=164]Epoch 258/400: 64%|█████████████████████████████████████████████████▋ | 258/400 [02:03<01:05, 2.18it/s, v_num=1, train_loss=164]Epoch 259/400: 64%|█████████████████████████████████████████████████▋ | 258/400 [02:03<01:05, 2.18it/s, v_num=1, train_loss=164]Epoch 259/400: 65%|█████████████████████████████████████████████████▊ | 259/400 [02:03<01:03, 2.21it/s, v_num=1, train_loss=164]Epoch 259/400: 65%|█████████████████████████████████████████████████▊ | 259/400 [02:03<01:03, 2.21it/s, v_num=1, train_loss=164]Epoch 260/400: 65%|█████████████████████████████████████████████████▊ | 259/400 [02:03<01:03, 2.21it/s, v_num=1, train_loss=164]Epoch 260/400: 65%|██████████████████████████████████████████████████ | 260/400 [02:04<01:03, 2.22it/s, v_num=1, train_loss=164]Epoch 260/400: 65%|██████████████████████████████████████████████████ | 260/400 [02:04<01:03, 2.22it/s, v_num=1, train_loss=164]Epoch 261/400: 65%|██████████████████████████████████████████████████ | 260/400 [02:04<01:03, 2.22it/s, v_num=1, train_loss=164]Epoch 261/400: 65%|██████████████████████████████████████████████████▏ | 261/400 [02:04<01:01, 2.26it/s, v_num=1, train_loss=164]Epoch 261/400: 65%|██████████████████████████████████████████████████▏ | 261/400 [02:04<01:01, 2.26it/s, v_num=1, train_loss=164]Epoch 262/400: 65%|██████████████████████████████████████████████████▏ | 261/400 [02:04<01:01, 2.26it/s, v_num=1, train_loss=164]Epoch 262/400: 66%|██████████████████████████████████████████████████▍ | 262/400 [02:05<01:01, 2.24it/s, v_num=1, train_loss=164]Epoch 262/400: 66%|██████████████████████████████████████████████████▍ | 262/400 [02:05<01:01, 2.24it/s, v_num=1, train_loss=164]Epoch 263/400: 66%|██████████████████████████████████████████████████▍ | 262/400 [02:05<01:01, 2.24it/s, v_num=1, train_loss=164]Epoch 263/400: 66%|██████████████████████████████████████████████████▋ | 263/400 [02:05<01:03, 2.15it/s, v_num=1, train_loss=164]Epoch 263/400: 66%|██████████████████████████████████████████████████▋ | 263/400 [02:05<01:03, 2.15it/s, v_num=1, train_loss=164]Epoch 264/400: 66%|██████████████████████████████████████████████████▋ | 263/400 [02:05<01:03, 2.15it/s, v_num=1, train_loss=164]Epoch 264/400: 66%|██████████████████████████████████████████████████▊ | 264/400 [02:06<01:03, 2.13it/s, v_num=1, train_loss=164]Epoch 264/400: 66%|██████████████████████████████████████████████████▊ | 264/400 [02:06<01:03, 2.13it/s, v_num=1, train_loss=164]Epoch 265/400: 66%|██████████████████████████████████████████████████▊ | 264/400 [02:06<01:03, 2.13it/s, v_num=1, train_loss=164]Epoch 265/400: 66%|███████████████████████████████████████████████████ | 265/400 [02:06<01:03, 2.13it/s, v_num=1, train_loss=164]Epoch 265/400: 66%|███████████████████████████████████████████████████ | 265/400 [02:06<01:03, 2.13it/s, v_num=1, train_loss=164]Epoch 266/400: 66%|███████████████████████████████████████████████████ | 265/400 [02:06<01:03, 2.13it/s, v_num=1, train_loss=164]Epoch 266/400: 66%|███████████████████████████████████████████████████▏ | 266/400 [02:07<01:01, 2.18it/s, v_num=1, train_loss=164]Epoch 266/400: 66%|███████████████████████████████████████████████████▏ | 266/400 [02:07<01:01, 2.18it/s, v_num=1, train_loss=164]Epoch 267/400: 66%|███████████████████████████████████████████████████▏ | 266/400 [02:07<01:01, 2.18it/s, v_num=1, train_loss=164]Epoch 267/400: 67%|███████████████████████████████████████████████████▍ | 267/400 [02:07<01:00, 2.19it/s, v_num=1, train_loss=164]Epoch 267/400: 67%|███████████████████████████████████████████████████▍ | 267/400 [02:07<01:00, 2.19it/s, v_num=1, train_loss=164]Epoch 268/400: 67%|███████████████████████████████████████████████████▍ | 267/400 [02:07<01:00, 2.19it/s, v_num=1, train_loss=164]Epoch 268/400: 67%|███████████████████████████████████████████████████▌ | 268/400 [02:07<01:00, 2.17it/s, v_num=1, train_loss=164]Epoch 268/400: 67%|███████████████████████████████████████████████████▌ | 268/400 [02:07<01:00, 2.17it/s, v_num=1, train_loss=164]Epoch 269/400: 67%|███████████████████████████████████████████████████▌ | 268/400 [02:07<01:00, 2.17it/s, v_num=1, train_loss=164]Epoch 269/400: 67%|███████████████████████████████████████████████████▊ | 269/400 [02:08<01:01, 2.14it/s, v_num=1, train_loss=164]Epoch 269/400: 67%|███████████████████████████████████████████████████▊ | 269/400 [02:08<01:01, 2.14it/s, v_num=1, train_loss=164]Epoch 270/400: 67%|███████████████████████████████████████████████████▊ | 269/400 [02:08<01:01, 2.14it/s, v_num=1, train_loss=164]Epoch 270/400: 68%|███████████████████████████████████████████████████▉ | 270/400 [02:08<01:01, 2.11it/s, v_num=1, train_loss=164]Epoch 270/400: 68%|███████████████████████████████████████████████████▉ | 270/400 [02:08<01:01, 2.11it/s, v_num=1, train_loss=164]Epoch 271/400: 68%|███████████████████████████████████████████████████▉ | 270/400 [02:08<01:01, 2.11it/s, v_num=1, train_loss=164]Epoch 271/400: 68%|████████████████████████████████████████████████████▏ | 271/400 [02:09<01:01, 2.10it/s, v_num=1, train_loss=164]Epoch 271/400: 68%|████████████████████████████████████████████████████▏ | 271/400 [02:09<01:01, 2.10it/s, v_num=1, train_loss=164]Epoch 272/400: 68%|████████████████████████████████████████████████████▏ | 271/400 [02:09<01:01, 2.10it/s, v_num=1, train_loss=164]Epoch 272/400: 68%|████████████████████████████████████████████████████▎ | 272/400 [02:09<01:01, 2.10it/s, v_num=1, train_loss=164]Epoch 272/400: 68%|████████████████████████████████████████████████████▎ | 272/400 [02:09<01:01, 2.10it/s, v_num=1, train_loss=164]Epoch 273/400: 68%|████████████████████████████████████████████████████▎ | 272/400 [02:09<01:01, 2.10it/s, v_num=1, train_loss=164]Epoch 273/400: 68%|████████████████████████████████████████████████████▌ | 273/400 [02:10<00:59, 2.12it/s, v_num=1, train_loss=164]Epoch 273/400: 68%|████████████████████████████████████████████████████▌ | 273/400 [02:10<00:59, 2.12it/s, v_num=1, train_loss=164]Epoch 274/400: 68%|████████████████████████████████████████████████████▌ | 273/400 [02:10<00:59, 2.12it/s, v_num=1, train_loss=164]Epoch 274/400: 68%|████████████████████████████████████████████████████▋ | 274/400 [02:10<00:58, 2.14it/s, v_num=1, train_loss=164]Epoch 274/400: 68%|████████████████████████████████████████████████████▋ | 274/400 [02:10<00:58, 2.14it/s, v_num=1, train_loss=164]Epoch 275/400: 68%|████████████████████████████████████████████████████▋ | 274/400 [02:10<00:58, 2.14it/s, v_num=1, train_loss=164]Epoch 275/400: 69%|████████████████████████████████████████████████████▉ | 275/400 [02:11<00:57, 2.18it/s, v_num=1, train_loss=164]Epoch 275/400: 69%|████████████████████████████████████████████████████▉ | 275/400 [02:11<00:57, 2.18it/s, v_num=1, train_loss=164]Epoch 276/400: 69%|████████████████████████████████████████████████████▉ | 275/400 [02:11<00:57, 2.18it/s, v_num=1, train_loss=164]Epoch 276/400: 69%|█████████████████████████████████████████████████████▏ | 276/400 [02:11<00:57, 2.17it/s, v_num=1, train_loss=164]Epoch 276/400: 69%|█████████████████████████████████████████████████████▏ | 276/400 [02:11<00:57, 2.17it/s, v_num=1, train_loss=164]Epoch 277/400: 69%|█████████████████████████████████████████████████████▏ | 276/400 [02:11<00:57, 2.17it/s, v_num=1, train_loss=164]Epoch 277/400: 69%|█████████████████████████████████████████████████████▎ | 277/400 [02:12<00:56, 2.17it/s, v_num=1, train_loss=164]Epoch 277/400: 69%|█████████████████████████████████████████████████████▎ | 277/400 [02:12<00:56, 2.17it/s, v_num=1, train_loss=164]Epoch 278/400: 69%|█████████████████████████████████████████████████████▎ | 277/400 [02:12<00:56, 2.17it/s, v_num=1, train_loss=164]Epoch 278/400: 70%|█████████████████████████████████████████████████████▌ | 278/400 [02:12<00:56, 2.17it/s, v_num=1, train_loss=164]Epoch 278/400: 70%|█████████████████████████████████████████████████████▌ | 278/400 [02:12<00:56, 2.17it/s, v_num=1, train_loss=164]Epoch 279/400: 70%|█████████████████████████████████████████████████████▌ | 278/400 [02:12<00:56, 2.17it/s, v_num=1, train_loss=164]Epoch 279/400: 70%|█████████████████████████████████████████████████████▋ | 279/400 [02:13<00:55, 2.20it/s, v_num=1, train_loss=164]Epoch 279/400: 70%|█████████████████████████████████████████████████████▋ | 279/400 [02:13<00:55, 2.20it/s, v_num=1, train_loss=164]Epoch 280/400: 70%|█████████████████████████████████████████████████████▋ | 279/400 [02:13<00:55, 2.20it/s, v_num=1, train_loss=164]Epoch 280/400: 70%|█████████████████████████████████████████████████████▉ | 280/400 [02:13<00:54, 2.20it/s, v_num=1, train_loss=164]Epoch 280/400: 70%|█████████████████████████████████████████████████████▉ | 280/400 [02:13<00:54, 2.20it/s, v_num=1, train_loss=164]Epoch 281/400: 70%|█████████████████████████████████████████████████████▉ | 280/400 [02:13<00:54, 2.20it/s, v_num=1, train_loss=164]Epoch 281/400: 70%|██████████████████████████████████████████████████████ | 281/400 [02:14<00:59, 2.01it/s, v_num=1, train_loss=164]Epoch 281/400: 70%|██████████████████████████████████████████████████████ | 281/400 [02:14<00:59, 2.01it/s, v_num=1, train_loss=164]Epoch 282/400: 70%|██████████████████████████████████████████████████████ | 281/400 [02:14<00:59, 2.01it/s, v_num=1, train_loss=164]Epoch 282/400: 70%|██████████████████████████████████████████████████████▎ | 282/400 [02:14<00:57, 2.06it/s, v_num=1, train_loss=164]Epoch 282/400: 70%|██████████████████████████████████████████████████████▎ | 282/400 [02:14<00:57, 2.06it/s, v_num=1, train_loss=164]Epoch 283/400: 70%|██████████████████████████████████████████████████████▎ | 282/400 [02:14<00:57, 2.06it/s, v_num=1, train_loss=164]Epoch 283/400: 71%|██████████████████████████████████████████████████████▍ | 283/400 [02:14<00:54, 2.15it/s, v_num=1, train_loss=164]Epoch 283/400: 71%|██████████████████████████████████████████████████████▍ | 283/400 [02:14<00:54, 2.15it/s, v_num=1, train_loss=164]Epoch 284/400: 71%|██████████████████████████████████████████████████████▍ | 283/400 [02:14<00:54, 2.15it/s, v_num=1, train_loss=164]Epoch 284/400: 71%|██████████████████████████████████████████████████████▋ | 284/400 [02:15<00:53, 2.19it/s, v_num=1, train_loss=164]Epoch 284/400: 71%|██████████████████████████████████████████████████████▋ | 284/400 [02:15<00:53, 2.19it/s, v_num=1, train_loss=164]Epoch 285/400: 71%|██████████████████████████████████████████████████████▋ | 284/400 [02:15<00:53, 2.19it/s, v_num=1, train_loss=164]Epoch 285/400: 71%|██████████████████████████████████████████████████████▊ | 285/400 [02:15<00:52, 2.20it/s, v_num=1, train_loss=164]Epoch 285/400: 71%|██████████████████████████████████████████████████████▊ | 285/400 [02:15<00:52, 2.20it/s, v_num=1, train_loss=164]Epoch 286/400: 71%|██████████████████████████████████████████████████████▊ | 285/400 [02:15<00:52, 2.20it/s, v_num=1, train_loss=164]Epoch 286/400: 72%|███████████████████████████████████████████████████████ | 286/400 [02:16<00:50, 2.24it/s, v_num=1, train_loss=164]Epoch 286/400: 72%|███████████████████████████████████████████████████████ | 286/400 [02:16<00:50, 2.24it/s, v_num=1, train_loss=164]Epoch 287/400: 72%|███████████████████████████████████████████████████████ | 286/400 [02:16<00:50, 2.24it/s, v_num=1, train_loss=164]Epoch 287/400: 72%|███████████████████████████████████████████████████████▏ | 287/400 [02:16<00:50, 2.25it/s, v_num=1, train_loss=164]Epoch 287/400: 72%|███████████████████████████████████████████████████████▏ | 287/400 [02:16<00:50, 2.25it/s, v_num=1, train_loss=164]Epoch 288/400: 72%|███████████████████████████████████████████████████████▏ | 287/400 [02:16<00:50, 2.25it/s, v_num=1, train_loss=164]Epoch 288/400: 72%|███████████████████████████████████████████████████████▍ | 288/400 [02:17<00:49, 2.25it/s, v_num=1, train_loss=164]Epoch 288/400: 72%|███████████████████████████████████████████████████████▍ | 288/400 [02:17<00:49, 2.25it/s, v_num=1, train_loss=164]Epoch 289/400: 72%|███████████████████████████████████████████████████████▍ | 288/400 [02:17<00:49, 2.25it/s, v_num=1, train_loss=164]Epoch 289/400: 72%|███████████████████████████████████████████████████████▋ | 289/400 [02:17<00:51, 2.15it/s, v_num=1, train_loss=164]Epoch 289/400: 72%|███████████████████████████████████████████████████████▋ | 289/400 [02:17<00:51, 2.15it/s, v_num=1, train_loss=164]Epoch 290/400: 72%|███████████████████████████████████████████████████████▋ | 289/400 [02:17<00:51, 2.15it/s, v_num=1, train_loss=164]Epoch 290/400: 72%|███████████████████████████████████████████████████████▊ | 290/400 [02:18<00:51, 2.12it/s, v_num=1, train_loss=164]Epoch 290/400: 72%|███████████████████████████████████████████████████████▊ | 290/400 [02:18<00:51, 2.12it/s, v_num=1, train_loss=164]Epoch 291/400: 72%|███████████████████████████████████████████████████████▊ | 290/400 [02:18<00:51, 2.12it/s, v_num=1, train_loss=164]Epoch 291/400: 73%|████████████████████████████████████████████████████████ | 291/400 [02:18<00:54, 1.99it/s, v_num=1, train_loss=164]Epoch 291/400: 73%|████████████████████████████████████████████████████████ | 291/400 [02:18<00:54, 1.99it/s, v_num=1, train_loss=164]Epoch 292/400: 73%|████████████████████████████████████████████████████████ | 291/400 [02:18<00:54, 1.99it/s, v_num=1, train_loss=164]Epoch 292/400: 73%|████████████████████████████████████████████████████████▏ | 292/400 [02:19<00:57, 1.87it/s, v_num=1, train_loss=164]Epoch 292/400: 73%|████████████████████████████████████████████████████████▏ | 292/400 [02:19<00:57, 1.87it/s, v_num=1, train_loss=164]Epoch 293/400: 73%|████████████████████████████████████████████████████████▏ | 292/400 [02:19<00:57, 1.87it/s, v_num=1, train_loss=164]Epoch 293/400: 73%|████████████████████████████████████████████████████████▍ | 293/400 [02:19<00:55, 1.93it/s, v_num=1, train_loss=164]Epoch 293/400: 73%|████████████████████████████████████████████████████████▍ | 293/400 [02:19<00:55, 1.93it/s, v_num=1, train_loss=164]Epoch 294/400: 73%|████████████████████████████████████████████████████████▍ | 293/400 [02:19<00:55, 1.93it/s, v_num=1, train_loss=164]Epoch 294/400: 74%|████████████████████████████████████████████████████████▌ | 294/400 [02:20<00:52, 2.01it/s, v_num=1, train_loss=164]Epoch 294/400: 74%|████████████████████████████████████████████████████████▌ | 294/400 [02:20<00:52, 2.01it/s, v_num=1, train_loss=164]Epoch 295/400: 74%|████████████████████████████████████████████████████████▌ | 294/400 [02:20<00:52, 2.01it/s, v_num=1, train_loss=164]Epoch 295/400: 74%|████████████████████████████████████████████████████████▊ | 295/400 [02:20<00:50, 2.08it/s, v_num=1, train_loss=164]Epoch 295/400: 74%|████████████████████████████████████████████████████████▊ | 295/400 [02:20<00:50, 2.08it/s, v_num=1, train_loss=164]Epoch 296/400: 74%|████████████████████████████████████████████████████████▊ | 295/400 [02:20<00:50, 2.08it/s, v_num=1, train_loss=164]Epoch 296/400: 74%|████████████████████████████████████████████████████████▉ | 296/400 [02:21<00:51, 2.01it/s, v_num=1, train_loss=164]Epoch 296/400: 74%|████████████████████████████████████████████████████████▉ | 296/400 [02:21<00:51, 2.01it/s, v_num=1, train_loss=164]Epoch 297/400: 74%|████████████████████████████████████████████████████████▉ | 296/400 [02:21<00:51, 2.01it/s, v_num=1, train_loss=164]Epoch 297/400: 74%|█████████████████████████████████████████████████████████▏ | 297/400 [02:21<00:51, 1.99it/s, v_num=1, train_loss=164]Epoch 297/400: 74%|█████████████████████████████████████████████████████████▏ | 297/400 [02:21<00:51, 1.99it/s, v_num=1, train_loss=164]Epoch 298/400: 74%|█████████████████████████████████████████████████████████▏ | 297/400 [02:21<00:51, 1.99it/s, v_num=1, train_loss=164]Epoch 298/400: 74%|█████████████████████████████████████████████████████████▎ | 298/400 [02:22<00:50, 2.01it/s, v_num=1, train_loss=164]Epoch 298/400: 74%|█████████████████████████████████████████████████████████▎ | 298/400 [02:22<00:50, 2.01it/s, v_num=1, train_loss=164]Epoch 299/400: 74%|█████████████████████████████████████████████████████████▎ | 298/400 [02:22<00:50, 2.01it/s, v_num=1, train_loss=164]Epoch 299/400: 75%|█████████████████████████████████████████████████████████▌ | 299/400 [02:22<00:49, 2.04it/s, v_num=1, train_loss=164]Epoch 299/400: 75%|█████████████████████████████████████████████████████████▌ | 299/400 [02:22<00:49, 2.04it/s, v_num=1, train_loss=164]Epoch 300/400: 75%|█████████████████████████████████████████████████████████▌ | 299/400 [02:22<00:49, 2.04it/s, v_num=1, train_loss=164]Epoch 300/400: 75%|█████████████████████████████████████████████████████████▊ | 300/400 [02:23<00:48, 2.06it/s, v_num=1, train_loss=164]Epoch 300/400: 75%|█████████████████████████████████████████████████████████▊ | 300/400 [02:23<00:48, 2.06it/s, v_num=1, train_loss=164]Epoch 301/400: 75%|█████████████████████████████████████████████████████████▊ | 300/400 [02:23<00:48, 2.06it/s, v_num=1, train_loss=164]Epoch 301/400: 75%|█████████████████████████████████████████████████████████▉ | 301/400 [02:23<00:48, 2.03it/s, v_num=1, train_loss=164]Epoch 301/400: 75%|█████████████████████████████████████████████████████████▉ | 301/400 [02:23<00:48, 2.03it/s, v_num=1, train_loss=164]Epoch 302/400: 75%|█████████████████████████████████████████████████████████▉ | 301/400 [02:23<00:48, 2.03it/s, v_num=1, train_loss=164]Epoch 302/400: 76%|██████████████████████████████████████████████████████████▏ | 302/400 [02:24<00:50, 1.94it/s, v_num=1, train_loss=164]Epoch 302/400: 76%|██████████████████████████████████████████████████████████▏ | 302/400 [02:24<00:50, 1.94it/s, v_num=1, train_loss=164]Epoch 303/400: 76%|██████████████████████████████████████████████████████████▏ | 302/400 [02:24<00:50, 1.94it/s, v_num=1, train_loss=164]Epoch 303/400: 76%|██████████████████████████████████████████████████████████▎ | 303/400 [02:24<00:48, 1.98it/s, v_num=1, train_loss=164]Epoch 303/400: 76%|██████████████████████████████████████████████████████████▎ | 303/400 [02:24<00:48, 1.98it/s, v_num=1, train_loss=164]Epoch 304/400: 76%|██████████████████████████████████████████████████████████▎ | 303/400 [02:24<00:48, 1.98it/s, v_num=1, train_loss=164]Epoch 304/400: 76%|██████████████████████████████████████████████████████████▌ | 304/400 [02:25<00:47, 2.04it/s, v_num=1, train_loss=164]Epoch 304/400: 76%|██████████████████████████████████████████████████████████▌ | 304/400 [02:25<00:47, 2.04it/s, v_num=1, train_loss=164]Epoch 305/400: 76%|██████████████████████████████████████████████████████████▌ | 304/400 [02:25<00:47, 2.04it/s, v_num=1, train_loss=164]Epoch 305/400: 76%|██████████████████████████████████████████████████████████▋ | 305/400 [02:25<00:48, 1.95it/s, v_num=1, train_loss=164]Epoch 305/400: 76%|██████████████████████████████████████████████████████████▋ | 305/400 [02:25<00:48, 1.95it/s, v_num=1, train_loss=164]Epoch 306/400: 76%|██████████████████████████████████████████████████████████▋ | 305/400 [02:25<00:48, 1.95it/s, v_num=1, train_loss=164]Epoch 306/400: 76%|██████████████████████████████████████████████████████████▉ | 306/400 [02:26<00:47, 1.99it/s, v_num=1, train_loss=164]Epoch 306/400: 76%|██████████████████████████████████████████████████████████▉ | 306/400 [02:26<00:47, 1.99it/s, v_num=1, train_loss=164]Epoch 307/400: 76%|██████████████████████████████████████████████████████████▉ | 306/400 [02:26<00:47, 1.99it/s, v_num=1, train_loss=164]Epoch 307/400: 77%|███████████████████████████████████████████████████████████ | 307/400 [02:26<00:46, 1.99it/s, v_num=1, train_loss=164]Epoch 307/400: 77%|███████████████████████████████████████████████████████████ | 307/400 [02:26<00:46, 1.99it/s, v_num=1, train_loss=164]Epoch 308/400: 77%|███████████████████████████████████████████████████████████ | 307/400 [02:26<00:46, 1.99it/s, v_num=1, train_loss=164]Epoch 308/400: 77%|███████████████████████████████████████████████████████████▎ | 308/400 [02:27<00:45, 2.02it/s, v_num=1, train_loss=164]Epoch 308/400: 77%|███████████████████████████████████████████████████████████▎ | 308/400 [02:27<00:45, 2.02it/s, v_num=1, train_loss=164]Epoch 309/400: 77%|███████████████████████████████████████████████████████████▎ | 308/400 [02:27<00:45, 2.02it/s, v_num=1, train_loss=164]Epoch 309/400: 77%|███████████████████████████████████████████████████████████▍ | 309/400 [02:27<00:45, 1.99it/s, v_num=1, train_loss=164]Epoch 309/400: 77%|███████████████████████████████████████████████████████████▍ | 309/400 [02:27<00:45, 1.99it/s, v_num=1, train_loss=164]Epoch 310/400: 77%|███████████████████████████████████████████████████████████▍ | 309/400 [02:27<00:45, 1.99it/s, v_num=1, train_loss=164]Epoch 310/400: 78%|███████████████████████████████████████████████████████████▋ | 310/400 [02:28<00:42, 2.09it/s, v_num=1, train_loss=164]Epoch 310/400: 78%|███████████████████████████████████████████████████████████▋ | 310/400 [02:28<00:42, 2.09it/s, v_num=1, train_loss=164]Epoch 311/400: 78%|███████████████████████████████████████████████████████████▋ | 310/400 [02:28<00:42, 2.09it/s, v_num=1, train_loss=164]Epoch 311/400: 78%|███████████████████████████████████████████████████████████▊ | 311/400 [02:28<00:41, 2.16it/s, v_num=1, train_loss=164]Epoch 311/400: 78%|███████████████████████████████████████████████████████████▊ | 311/400 [02:28<00:41, 2.16it/s, v_num=1, train_loss=164]Epoch 312/400: 78%|███████████████████████████████████████████████████████████▊ | 311/400 [02:28<00:41, 2.16it/s, v_num=1, train_loss=164]Epoch 312/400: 78%|████████████████████████████████████████████████████████████ | 312/400 [02:29<00:39, 2.21it/s, v_num=1, train_loss=164]Epoch 312/400: 78%|████████████████████████████████████████████████████████████ | 312/400 [02:29<00:39, 2.21it/s, v_num=1, train_loss=164]Epoch 313/400: 78%|████████████████████████████████████████████████████████████ | 312/400 [02:29<00:39, 2.21it/s, v_num=1, train_loss=164]Epoch 313/400: 78%|████████████████████████████████████████████████████████████▎ | 313/400 [02:29<00:39, 2.22it/s, v_num=1, train_loss=164]Epoch 313/400: 78%|████████████████████████████████████████████████████████████▎ | 313/400 [02:29<00:39, 2.22it/s, v_num=1, train_loss=164]Epoch 314/400: 78%|████████████████████████████████████████████████████████████▎ | 313/400 [02:29<00:39, 2.22it/s, v_num=1, train_loss=164]Epoch 314/400: 78%|████████████████████████████████████████████████████████████▍ | 314/400 [02:30<00:40, 2.12it/s, v_num=1, train_loss=164]Epoch 314/400: 78%|████████████████████████████████████████████████████████████▍ | 314/400 [02:30<00:40, 2.12it/s, v_num=1, train_loss=164]Epoch 315/400: 78%|████████████████████████████████████████████████████████████▍ | 314/400 [02:30<00:40, 2.12it/s, v_num=1, train_loss=164]Epoch 315/400: 79%|████████████████████████████████████████████████████████████▋ | 315/400 [02:30<00:41, 2.07it/s, v_num=1, train_loss=164]Epoch 315/400: 79%|████████████████████████████████████████████████████████████▋ | 315/400 [02:30<00:41, 2.07it/s, v_num=1, train_loss=164]Epoch 316/400: 79%|████████████████████████████████████████████████████████████▋ | 315/400 [02:30<00:41, 2.07it/s, v_num=1, train_loss=164]Epoch 316/400: 79%|████████████████████████████████████████████████████████████▊ | 316/400 [02:30<00:39, 2.13it/s, v_num=1, train_loss=164]Epoch 316/400: 79%|████████████████████████████████████████████████████████████▊ | 316/400 [02:30<00:39, 2.13it/s, v_num=1, train_loss=164]Epoch 317/400: 79%|████████████████████████████████████████████████████████████▊ | 316/400 [02:30<00:39, 2.13it/s, v_num=1, train_loss=164]Epoch 317/400: 79%|█████████████████████████████████████████████████████████████ | 317/400 [02:31<00:37, 2.19it/s, v_num=1, train_loss=164]Epoch 317/400: 79%|█████████████████████████████████████████████████████████████ | 317/400 [02:31<00:37, 2.19it/s, v_num=1, train_loss=164]Epoch 318/400: 79%|█████████████████████████████████████████████████████████████ | 317/400 [02:31<00:37, 2.19it/s, v_num=1, train_loss=164]Epoch 318/400: 80%|█████████████████████████████████████████████████████████████▏ | 318/400 [02:31<00:38, 2.15it/s, v_num=1, train_loss=164]Epoch 318/400: 80%|█████████████████████████████████████████████████████████████▏ | 318/400 [02:31<00:38, 2.15it/s, v_num=1, train_loss=164]Epoch 319/400: 80%|█████████████████████████████████████████████████████████████▏ | 318/400 [02:31<00:38, 2.15it/s, v_num=1, train_loss=164]Epoch 319/400: 80%|█████████████████████████████████████████████████████████████▍ | 319/400 [02:32<00:36, 2.21it/s, v_num=1, train_loss=164]Epoch 319/400: 80%|█████████████████████████████████████████████████████████████▍ | 319/400 [02:32<00:36, 2.21it/s, v_num=1, train_loss=164]Epoch 320/400: 80%|█████████████████████████████████████████████████████████████▍ | 319/400 [02:32<00:36, 2.21it/s, v_num=1, train_loss=164]Epoch 320/400: 80%|█████████████████████████████████████████████████████████████▌ | 320/400 [02:32<00:36, 2.22it/s, v_num=1, train_loss=164]Epoch 320/400: 80%|█████████████████████████████████████████████████████████████▌ | 320/400 [02:32<00:36, 2.22it/s, v_num=1, train_loss=164]Epoch 321/400: 80%|█████████████████████████████████████████████████████████████▌ | 320/400 [02:32<00:36, 2.22it/s, v_num=1, train_loss=164]Epoch 321/400: 80%|█████████████████████████████████████████████████████████████▊ | 321/400 [02:33<00:35, 2.24it/s, v_num=1, train_loss=164]Epoch 321/400: 80%|█████████████████████████████████████████████████████████████▊ | 321/400 [02:33<00:35, 2.24it/s, v_num=1, train_loss=164]Epoch 322/400: 80%|█████████████████████████████████████████████████████████████▊ | 321/400 [02:33<00:35, 2.24it/s, v_num=1, train_loss=164]Epoch 322/400: 80%|█████████████████████████████████████████████████████████████▉ | 322/400 [02:33<00:34, 2.23it/s, v_num=1, train_loss=164]Epoch 322/400: 80%|█████████████████████████████████████████████████████████████▉ | 322/400 [02:33<00:34, 2.23it/s, v_num=1, train_loss=164]Epoch 323/400: 80%|█████████████████████████████████████████████████████████████▉ | 322/400 [02:33<00:34, 2.23it/s, v_num=1, train_loss=164]Epoch 323/400: 81%|██████████████████████████████████████████████████████████████▏ | 323/400 [02:34<00:35, 2.16it/s, v_num=1, train_loss=164]Epoch 323/400: 81%|██████████████████████████████████████████████████████████████▏ | 323/400 [02:34<00:35, 2.16it/s, v_num=1, train_loss=164]Epoch 324/400: 81%|██████████████████████████████████████████████████████████████▏ | 323/400 [02:34<00:35, 2.16it/s, v_num=1, train_loss=164]Epoch 324/400: 81%|██████████████████████████████████████████████████████████████▎ | 324/400 [02:34<00:34, 2.20it/s, v_num=1, train_loss=164]Epoch 324/400: 81%|██████████████████████████████████████████████████████████████▎ | 324/400 [02:34<00:34, 2.20it/s, v_num=1, train_loss=164]Epoch 325/400: 81%|██████████████████████████████████████████████████████████████▎ | 324/400 [02:34<00:34, 2.20it/s, v_num=1, train_loss=164]Epoch 325/400: 81%|██████████████████████████████████████████████████████████████▌ | 325/400 [02:35<00:34, 2.20it/s, v_num=1, train_loss=164]Epoch 325/400: 81%|██████████████████████████████████████████████████████████████▌ | 325/400 [02:35<00:34, 2.20it/s, v_num=1, train_loss=164]Epoch 326/400: 81%|██████████████████████████████████████████████████████████████▌ | 325/400 [02:35<00:34, 2.20it/s, v_num=1, train_loss=164]Epoch 326/400: 82%|██████████████████████████████████████████████████████████████▊ | 326/400 [02:35<00:37, 1.95it/s, v_num=1, train_loss=164]Epoch 326/400: 82%|██████████████████████████████████████████████████████████████▊ | 326/400 [02:35<00:37, 1.95it/s, v_num=1, train_loss=164]Epoch 327/400: 82%|██████████████████████████████████████████████████████████████▊ | 326/400 [02:35<00:37, 1.95it/s, v_num=1, train_loss=164]Epoch 327/400: 82%|██████████████████████████████████████████████████████████████▉ | 327/400 [02:36<00:36, 1.99it/s, v_num=1, train_loss=164]Epoch 327/400: 82%|██████████████████████████████████████████████████████████████▉ | 327/400 [02:36<00:36, 1.99it/s, v_num=1, train_loss=164]Epoch 328/400: 82%|██████████████████████████████████████████████████████████████▉ | 327/400 [02:36<00:36, 1.99it/s, v_num=1, train_loss=164]Epoch 328/400: 82%|███████████████████████████████████████████████████████████████▏ | 328/400 [02:36<00:34, 2.07it/s, v_num=1, train_loss=164]Epoch 328/400: 82%|███████████████████████████████████████████████████████████████▏ | 328/400 [02:36<00:34, 2.07it/s, v_num=1, train_loss=164]Epoch 329/400: 82%|███████████████████████████████████████████████████████████████▏ | 328/400 [02:36<00:34, 2.07it/s, v_num=1, train_loss=164]Epoch 329/400: 82%|███████████████████████████████████████████████████████████████▎ | 329/400 [02:37<00:33, 2.11it/s, v_num=1, train_loss=164]Epoch 329/400: 82%|███████████████████████████████████████████████████████████████▎ | 329/400 [02:37<00:33, 2.11it/s, v_num=1, train_loss=164]Epoch 330/400: 82%|███████████████████████████████████████████████████████████████▎ | 329/400 [02:37<00:33, 2.11it/s, v_num=1, train_loss=164]Epoch 330/400: 82%|███████████████████████████████████████████████████████████████▌ | 330/400 [02:37<00:32, 2.17it/s, v_num=1, train_loss=164]Epoch 330/400: 82%|███████████████████████████████████████████████████████████████▌ | 330/400 [02:37<00:32, 2.17it/s, v_num=1, train_loss=164]Epoch 331/400: 82%|███████████████████████████████████████████████████████████████▌ | 330/400 [02:37<00:32, 2.17it/s, v_num=1, train_loss=164]Epoch 331/400: 83%|███████████████████████████████████████████████████████████████▋ | 331/400 [02:37<00:32, 2.15it/s, v_num=1, train_loss=164]Epoch 331/400: 83%|███████████████████████████████████████████████████████████████▋ | 331/400 [02:37<00:32, 2.15it/s, v_num=1, train_loss=164]Epoch 332/400: 83%|███████████████████████████████████████████████████████████████▋ | 331/400 [02:37<00:32, 2.15it/s, v_num=1, train_loss=164]Epoch 332/400: 83%|███████████████████████████████████████████████████████████████▉ | 332/400 [02:38<00:31, 2.19it/s, v_num=1, train_loss=164]Epoch 332/400: 83%|███████████████████████████████████████████████████████████████▉ | 332/400 [02:38<00:31, 2.19it/s, v_num=1, train_loss=164]Epoch 333/400: 83%|███████████████████████████████████████████████████████████████▉ | 332/400 [02:38<00:31, 2.19it/s, v_num=1, train_loss=164]Epoch 333/400: 83%|████████████████████████████████████████████████████████████████ | 333/400 [02:38<00:30, 2.20it/s, v_num=1, train_loss=164]Epoch 333/400: 83%|████████████████████████████████████████████████████████████████ | 333/400 [02:38<00:30, 2.20it/s, v_num=1, train_loss=164]Epoch 334/400: 83%|████████████████████████████████████████████████████████████████ | 333/400 [02:38<00:30, 2.20it/s, v_num=1, train_loss=164]Epoch 334/400: 84%|████████████████████████████████████████████████████████████████▎ | 334/400 [02:39<00:30, 2.19it/s, v_num=1, train_loss=164]Epoch 334/400: 84%|████████████████████████████████████████████████████████████████▎ | 334/400 [02:39<00:30, 2.19it/s, v_num=1, train_loss=164]Epoch 335/400: 84%|████████████████████████████████████████████████████████████████▎ | 334/400 [02:39<00:30, 2.19it/s, v_num=1, train_loss=164]Epoch 335/400: 84%|████████████████████████████████████████████████████████████████▍ | 335/400 [02:39<00:29, 2.17it/s, v_num=1, train_loss=164]Epoch 335/400: 84%|████████████████████████████████████████████████████████████████▍ | 335/400 [02:39<00:29, 2.17it/s, v_num=1, train_loss=164]Epoch 336/400: 84%|████████████████████████████████████████████████████████████████▍ | 335/400 [02:39<00:29, 2.17it/s, v_num=1, train_loss=164]Epoch 336/400: 84%|████████████████████████████████████████████████████████████████▋ | 336/400 [02:40<00:30, 2.09it/s, v_num=1, train_loss=164]Epoch 336/400: 84%|████████████████████████████████████████████████████████████████▋ | 336/400 [02:40<00:30, 2.09it/s, v_num=1, train_loss=164]Epoch 337/400: 84%|████████████████████████████████████████████████████████████████▋ | 336/400 [02:40<00:30, 2.09it/s, v_num=1, train_loss=164]Epoch 337/400: 84%|████████████████████████████████████████████████████████████████▊ | 337/400 [02:40<00:30, 2.08it/s, v_num=1, train_loss=164]Epoch 337/400: 84%|████████████████████████████████████████████████████████████████▊ | 337/400 [02:40<00:30, 2.08it/s, v_num=1, train_loss=164]Epoch 338/400: 84%|████████████████████████████████████████████████████████████████▊ | 337/400 [02:40<00:30, 2.08it/s, v_num=1, train_loss=164]Epoch 338/400: 84%|█████████████████████████████████████████████████████████████████ | 338/400 [02:41<00:30, 2.05it/s, v_num=1, train_loss=164]Epoch 338/400: 84%|█████████████████████████████████████████████████████████████████ | 338/400 [02:41<00:30, 2.05it/s, v_num=1, train_loss=164]Epoch 339/400: 84%|█████████████████████████████████████████████████████████████████ | 338/400 [02:41<00:30, 2.05it/s, v_num=1, train_loss=164]Epoch 339/400: 85%|█████████████████████████████████████████████████████████████████▎ | 339/400 [02:41<00:29, 2.06it/s, v_num=1, train_loss=164]Epoch 339/400: 85%|█████████████████████████████████████████████████████████████████▎ | 339/400 [02:41<00:29, 2.06it/s, v_num=1, train_loss=164]Epoch 340/400: 85%|█████████████████████████████████████████████████████████████████▎ | 339/400 [02:41<00:29, 2.06it/s, v_num=1, train_loss=164]Epoch 340/400: 85%|█████████████████████████████████████████████████████████████████▍ | 340/400 [02:42<00:28, 2.09it/s, v_num=1, train_loss=164]Epoch 340/400: 85%|█████████████████████████████████████████████████████████████████▍ | 340/400 [02:42<00:28, 2.09it/s, v_num=1, train_loss=164]Epoch 341/400: 85%|█████████████████████████████████████████████████████████████████▍ | 340/400 [02:42<00:28, 2.09it/s, v_num=1, train_loss=164]Epoch 341/400: 85%|█████████████████████████████████████████████████████████████████▋ | 341/400 [02:42<00:27, 2.13it/s, v_num=1, train_loss=164]Epoch 341/400: 85%|█████████████████████████████████████████████████████████████████▋ | 341/400 [02:42<00:27, 2.13it/s, v_num=1, train_loss=164]Epoch 342/400: 85%|█████████████████████████████████████████████████████████████████▋ | 341/400 [02:42<00:27, 2.13it/s, v_num=1, train_loss=164]Epoch 342/400: 86%|█████████████████████████████████████████████████████████████████▊ | 342/400 [02:43<00:27, 2.13it/s, v_num=1, train_loss=164]Epoch 342/400: 86%|█████████████████████████████████████████████████████████████████▊ | 342/400 [02:43<00:27, 2.13it/s, v_num=1, train_loss=164]Epoch 343/400: 86%|█████████████████████████████████████████████████████████████████▊ | 342/400 [02:43<00:27, 2.13it/s, v_num=1, train_loss=164]Epoch 343/400: 86%|██████████████████████████████████████████████████████████████████ | 343/400 [02:43<00:26, 2.19it/s, v_num=1, train_loss=164]Epoch 343/400: 86%|██████████████████████████████████████████████████████████████████ | 343/400 [02:43<00:26, 2.19it/s, v_num=1, train_loss=164]Epoch 344/400: 86%|██████████████████████████████████████████████████████████████████ | 343/400 [02:43<00:26, 2.19it/s, v_num=1, train_loss=164]Epoch 344/400: 86%|██████████████████████████████████████████████████████████████████▏ | 344/400 [02:44<00:26, 2.13it/s, v_num=1, train_loss=164]Epoch 344/400: 86%|██████████████████████████████████████████████████████████████████▏ | 344/400 [02:44<00:26, 2.13it/s, v_num=1, train_loss=164]Epoch 345/400: 86%|██████████████████████████████████████████████████████████████████▏ | 344/400 [02:44<00:26, 2.13it/s, v_num=1, train_loss=164]Epoch 345/400: 86%|██████████████████████████████████████████████████████████████████▍ | 345/400 [02:44<00:26, 2.04it/s, v_num=1, train_loss=164]Epoch 345/400: 86%|██████████████████████████████████████████████████████████████████▍ | 345/400 [02:44<00:26, 2.04it/s, v_num=1, train_loss=164]Epoch 346/400: 86%|██████████████████████████████████████████████████████████████████▍ | 345/400 [02:44<00:26, 2.04it/s, v_num=1, train_loss=164]Epoch 346/400: 86%|██████████████████████████████████████████████████████████████████▌ | 346/400 [02:45<00:26, 2.06it/s, v_num=1, train_loss=164]Epoch 346/400: 86%|██████████████████████████████████████████████████████████████████▌ | 346/400 [02:45<00:26, 2.06it/s, v_num=1, train_loss=164]Epoch 347/400: 86%|██████████████████████████████████████████████████████████████████▌ | 346/400 [02:45<00:26, 2.06it/s, v_num=1, train_loss=164]Epoch 347/400: 87%|██████████████████████████████████████████████████████████████████▊ | 347/400 [02:45<00:25, 2.08it/s, v_num=1, train_loss=164]Epoch 347/400: 87%|██████████████████████████████████████████████████████████████████▊ | 347/400 [02:45<00:25, 2.08it/s, v_num=1, train_loss=164]Epoch 348/400: 87%|██████████████████████████████████████████████████████████████████▊ | 347/400 [02:45<00:25, 2.08it/s, v_num=1, train_loss=164]Epoch 348/400: 87%|██████████████████████████████████████████████████████████████████▉ | 348/400 [02:46<00:25, 2.08it/s, v_num=1, train_loss=164]Epoch 348/400: 87%|██████████████████████████████████████████████████████████████████▉ | 348/400 [02:46<00:25, 2.08it/s, v_num=1, train_loss=164]Epoch 349/400: 87%|██████████████████████████████████████████████████████████████████▉ | 348/400 [02:46<00:25, 2.08it/s, v_num=1, train_loss=164]Epoch 349/400: 87%|███████████████████████████████████████████████████████████████████▏ | 349/400 [02:46<00:25, 2.02it/s, v_num=1, train_loss=164]Epoch 349/400: 87%|███████████████████████████████████████████████████████████████████▏ | 349/400 [02:46<00:25, 2.02it/s, v_num=1, train_loss=164]Epoch 350/400: 87%|███████████████████████████████████████████████████████████████████▏ | 349/400 [02:46<00:25, 2.02it/s, v_num=1, train_loss=164]Epoch 350/400: 88%|███████████████████████████████████████████████████████████████████▍ | 350/400 [02:47<00:24, 2.04it/s, v_num=1, train_loss=164]Epoch 350/400: 88%|███████████████████████████████████████████████████████████████████▍ | 350/400 [02:47<00:24, 2.04it/s, v_num=1, train_loss=164]Epoch 351/400: 88%|███████████████████████████████████████████████████████████████████▍ | 350/400 [02:47<00:24, 2.04it/s, v_num=1, train_loss=164]Epoch 351/400: 88%|███████████████████████████████████████████████████████████████████▌ | 351/400 [02:47<00:23, 2.06it/s, v_num=1, train_loss=164]Epoch 351/400: 88%|███████████████████████████████████████████████████████████████████▌ | 351/400 [02:47<00:23, 2.06it/s, v_num=1, train_loss=164]Epoch 352/400: 88%|███████████████████████████████████████████████████████████████████▌ | 351/400 [02:47<00:23, 2.06it/s, v_num=1, train_loss=164]Epoch 352/400: 88%|███████████████████████████████████████████████████████████████████▊ | 352/400 [02:47<00:23, 2.06it/s, v_num=1, train_loss=164]Epoch 352/400: 88%|███████████████████████████████████████████████████████████████████▊ | 352/400 [02:47<00:23, 2.06it/s, v_num=1, train_loss=164]Epoch 353/400: 88%|███████████████████████████████████████████████████████████████████▊ | 352/400 [02:47<00:23, 2.06it/s, v_num=1, train_loss=164]Epoch 353/400: 88%|███████████████████████████████████████████████████████████████████▉ | 353/400 [02:48<00:22, 2.11it/s, v_num=1, train_loss=164]Epoch 353/400: 88%|███████████████████████████████████████████████████████████████████▉ | 353/400 [02:48<00:22, 2.11it/s, v_num=1, train_loss=164]Epoch 354/400: 88%|███████████████████████████████████████████████████████████████████▉ | 353/400 [02:48<00:22, 2.11it/s, v_num=1, train_loss=164]Epoch 354/400: 88%|████████████████████████████████████████████████████████████████████▏ | 354/400 [02:48<00:21, 2.09it/s, v_num=1, train_loss=164]Epoch 354/400: 88%|████████████████████████████████████████████████████████████████████▏ | 354/400 [02:48<00:21, 2.09it/s, v_num=1, train_loss=164]Epoch 355/400: 88%|████████████████████████████████████████████████████████████████████▏ | 354/400 [02:48<00:21, 2.09it/s, v_num=1, train_loss=164]Epoch 355/400: 89%|████████████████████████████████████████████████████████████████████▎ | 355/400 [02:49<00:21, 2.10it/s, v_num=1, train_loss=164]Epoch 355/400: 89%|████████████████████████████████████████████████████████████████████▎ | 355/400 [02:49<00:21, 2.10it/s, v_num=1, train_loss=164]Epoch 356/400: 89%|████████████████████████████████████████████████████████████████████▎ | 355/400 [02:49<00:21, 2.10it/s, v_num=1, train_loss=164]Epoch 356/400: 89%|████████████████████████████████████████████████████████████████████▌ | 356/400 [02:49<00:21, 2.05it/s, v_num=1, train_loss=164]Epoch 356/400: 89%|████████████████████████████████████████████████████████████████████▌ | 356/400 [02:49<00:21, 2.05it/s, v_num=1, train_loss=164]Epoch 357/400: 89%|████████████████████████████████████████████████████████████████████▌ | 356/400 [02:49<00:21, 2.05it/s, v_num=1, train_loss=164]Epoch 357/400: 89%|████████████████████████████████████████████████████████████████████▋ | 357/400 [02:50<00:20, 2.05it/s, v_num=1, train_loss=164]Epoch 357/400: 89%|████████████████████████████████████████████████████████████████████▋ | 357/400 [02:50<00:20, 2.05it/s, v_num=1, train_loss=164]Epoch 358/400: 89%|████████████████████████████████████████████████████████████████████▋ | 357/400 [02:50<00:20, 2.05it/s, v_num=1, train_loss=164]Epoch 358/400: 90%|████████████████████████████████████████████████████████████████████▉ | 358/400 [02:50<00:20, 2.08it/s, v_num=1, train_loss=164]Epoch 358/400: 90%|████████████████████████████████████████████████████████████████████▉ | 358/400 [02:50<00:20, 2.08it/s, v_num=1, train_loss=164]Epoch 359/400: 90%|████████████████████████████████████████████████████████████████████▉ | 358/400 [02:50<00:20, 2.08it/s, v_num=1, train_loss=164]Epoch 359/400: 90%|█████████████████████████████████████████████████████████████████████ | 359/400 [02:51<00:20, 1.97it/s, v_num=1, train_loss=164]Epoch 359/400: 90%|█████████████████████████████████████████████████████████████████████ | 359/400 [02:51<00:20, 1.97it/s, v_num=1, train_loss=164]Epoch 360/400: 90%|█████████████████████████████████████████████████████████████████████ | 359/400 [02:51<00:20, 1.97it/s, v_num=1, train_loss=164]Epoch 360/400: 90%|█████████████████████████████████████████████████████████████████████▎ | 360/400 [02:51<00:20, 1.93it/s, v_num=1, train_loss=164]Epoch 360/400: 90%|█████████████████████████████████████████████████████████████████████▎ | 360/400 [02:51<00:20, 1.93it/s, v_num=1, train_loss=164]Epoch 361/400: 90%|█████████████████████████████████████████████████████████████████████▎ | 360/400 [02:51<00:20, 1.93it/s, v_num=1, train_loss=164]Epoch 361/400: 90%|█████████████████████████████████████████████████████████████████████▍ | 361/400 [02:52<00:19, 1.98it/s, v_num=1, train_loss=164]Epoch 361/400: 90%|█████████████████████████████████████████████████████████████████████▍ | 361/400 [02:52<00:19, 1.98it/s, v_num=1, train_loss=164]Epoch 362/400: 90%|█████████████████████████████████████████████████████████████████████▍ | 361/400 [02:52<00:19, 1.98it/s, v_num=1, train_loss=164]Epoch 362/400: 90%|█████████████████████████████████████████████████████████████████████▋ | 362/400 [02:52<00:19, 1.92it/s, v_num=1, train_loss=164]Epoch 362/400: 90%|█████████████████████████████████████████████████████████████████████▋ | 362/400 [02:52<00:19, 1.92it/s, v_num=1, train_loss=165]Epoch 363/400: 90%|█████████████████████████████████████████████████████████████████████▋ | 362/400 [02:53<00:19, 1.92it/s, v_num=1, train_loss=165]Epoch 363/400: 91%|█████████████████████████████████████████████████████████████████████▉ | 363/400 [02:53<00:19, 1.92it/s, v_num=1, train_loss=165]Epoch 363/400: 91%|█████████████████████████████████████████████████████████████████████▉ | 363/400 [02:53<00:19, 1.92it/s, v_num=1, train_loss=164]Epoch 364/400: 91%|█████████████████████████████████████████████████████████████████████▉ | 363/400 [02:53<00:19, 1.92it/s, v_num=1, train_loss=164]Epoch 364/400: 91%|██████████████████████████████████████████████████████████████████████ | 364/400 [02:54<00:19, 1.88it/s, v_num=1, train_loss=164]Epoch 364/400: 91%|██████████████████████████████████████████████████████████████████████ | 364/400 [02:54<00:19, 1.88it/s, v_num=1, train_loss=164]Epoch 365/400: 91%|██████████████████████████████████████████████████████████████████████ | 364/400 [02:54<00:19, 1.88it/s, v_num=1, train_loss=164]Epoch 365/400: 91%|██████████████████████████████████████████████████████████████████████▎ | 365/400 [02:54<00:18, 1.86it/s, v_num=1, train_loss=164]Epoch 365/400: 91%|██████████████████████████████████████████████████████████████████████▎ | 365/400 [02:54<00:18, 1.86it/s, v_num=1, train_loss=164]Epoch 366/400: 91%|██████████████████████████████████████████████████████████████████████▎ | 365/400 [02:54<00:18, 1.86it/s, v_num=1, train_loss=164]Epoch 366/400: 92%|██████████████████████████████████████████████████████████████████████▍ | 366/400 [02:55<00:18, 1.88it/s, v_num=1, train_loss=164]Epoch 366/400: 92%|██████████████████████████████████████████████████████████████████████▍ | 366/400 [02:55<00:18, 1.88it/s, v_num=1, train_loss=164]Epoch 367/400: 92%|██████████████████████████████████████████████████████████████████████▍ | 366/400 [02:55<00:18, 1.88it/s, v_num=1, train_loss=164]Epoch 367/400: 92%|██████████████████████████████████████████████████████████████████████▋ | 367/400 [02:55<00:17, 1.87it/s, v_num=1, train_loss=164]Epoch 367/400: 92%|██████████████████████████████████████████████████████████████████████▋ | 367/400 [02:55<00:17, 1.87it/s, v_num=1, train_loss=164]Epoch 368/400: 92%|██████████████████████████████████████████████████████████████████████▋ | 367/400 [02:55<00:17, 1.87it/s, v_num=1, train_loss=164]Epoch 368/400: 92%|██████████████████████████████████████████████████████████████████████▊ | 368/400 [02:56<00:17, 1.83it/s, v_num=1, train_loss=164]Epoch 368/400: 92%|██████████████████████████████████████████████████████████████████████▊ | 368/400 [02:56<00:17, 1.83it/s, v_num=1, train_loss=164]Epoch 369/400: 92%|██████████████████████████████████████████████████████████████████████▊ | 368/400 [02:56<00:17, 1.83it/s, v_num=1, train_loss=164]Epoch 369/400: 92%|███████████████████████████████████████████████████████████████████████ | 369/400 [02:56<00:17, 1.82it/s, v_num=1, train_loss=164]Epoch 369/400: 92%|███████████████████████████████████████████████████████████████████████ | 369/400 [02:56<00:17, 1.82it/s, v_num=1, train_loss=165]Epoch 370/400: 92%|███████████████████████████████████████████████████████████████████████ | 369/400 [02:56<00:17, 1.82it/s, v_num=1, train_loss=165]Epoch 370/400: 92%|███████████████████████████████████████████████████████████████████████▏ | 370/400 [02:57<00:16, 1.86it/s, v_num=1, train_loss=165]Epoch 370/400: 92%|███████████████████████████████████████████████████████████████████████▏ | 370/400 [02:57<00:16, 1.86it/s, v_num=1, train_loss=165]Epoch 371/400: 92%|███████████████████████████████████████████████████████████████████████▏ | 370/400 [02:57<00:16, 1.86it/s, v_num=1, train_loss=165]Epoch 371/400: 93%|███████████████████████████████████████████████████████████████████████▍ | 371/400 [02:57<00:14, 1.93it/s, v_num=1, train_loss=165]Epoch 371/400: 93%|███████████████████████████████████████████████████████████████████████▍ | 371/400 [02:57<00:14, 1.93it/s, v_num=1, train_loss=165]Epoch 372/400: 93%|███████████████████████████████████████████████████████████████████████▍ | 371/400 [02:57<00:14, 1.93it/s, v_num=1, train_loss=165]Epoch 372/400: 93%|███████████████████████████████████████████████████████████████████████▌ | 372/400 [02:58<00:14, 1.91it/s, v_num=1, train_loss=165]Epoch 372/400: 93%|███████████████████████████████████████████████████████████████████████▌ | 372/400 [02:58<00:14, 1.91it/s, v_num=1, train_loss=164]Epoch 373/400: 93%|███████████████████████████████████████████████████████████████████████▌ | 372/400 [02:58<00:14, 1.91it/s, v_num=1, train_loss=164]Epoch 373/400: 93%|███████████████████████████████████████████████████████████████████████▊ | 373/400 [02:58<00:13, 1.93it/s, v_num=1, train_loss=164]Epoch 373/400: 93%|███████████████████████████████████████████████████████████████████████▊ | 373/400 [02:58<00:13, 1.93it/s, v_num=1, train_loss=164]Epoch 374/400: 93%|███████████████████████████████████████████████████████████████████████▊ | 373/400 [02:58<00:13, 1.93it/s, v_num=1, train_loss=164]Epoch 374/400: 94%|███████████████████████████████████████████████████████████████████████▉ | 374/400 [02:59<00:14, 1.86it/s, v_num=1, train_loss=164]Epoch 374/400: 94%|███████████████████████████████████████████████████████████████████████▉ | 374/400 [02:59<00:14, 1.86it/s, v_num=1, train_loss=164]Epoch 375/400: 94%|███████████████████████████████████████████████████████████████████████▉ | 374/400 [02:59<00:14, 1.86it/s, v_num=1, train_loss=164]Epoch 375/400: 94%|████████████████████████████████████████████████████████████████████████▏ | 375/400 [02:59<00:13, 1.85it/s, v_num=1, train_loss=164]Epoch 375/400: 94%|████████████████████████████████████████████████████████████████████████▏ | 375/400 [02:59<00:13, 1.85it/s, v_num=1, train_loss=165]Epoch 376/400: 94%|████████████████████████████████████████████████████████████████████████▏ | 375/400 [02:59<00:13, 1.85it/s, v_num=1, train_loss=165]Epoch 376/400: 94%|████████████████████████████████████████████████████████████████████████▍ | 376/400 [03:00<00:13, 1.74it/s, v_num=1, train_loss=165]Epoch 376/400: 94%|████████████████████████████████████████████████████████████████████████▍ | 376/400 [03:00<00:13, 1.74it/s, v_num=1, train_loss=165]Epoch 377/400: 94%|████████████████████████████████████████████████████████████████████████▍ | 376/400 [03:00<00:13, 1.74it/s, v_num=1, train_loss=165]Epoch 377/400: 94%|████████████████████████████████████████████████████████████████████████▌ | 377/400 [03:01<00:13, 1.71it/s, v_num=1, train_loss=165]Epoch 377/400: 94%|████████████████████████████████████████████████████████████████████████▌ | 377/400 [03:01<00:13, 1.71it/s, v_num=1, train_loss=165]Epoch 378/400: 94%|████████████████████████████████████████████████████████████████████████▌ | 377/400 [03:01<00:13, 1.71it/s, v_num=1, train_loss=165]Epoch 378/400: 94%|████████████████████████████████████████████████████████████████████████▊ | 378/400 [03:01<00:12, 1.70it/s, v_num=1, train_loss=165]Epoch 378/400: 94%|████████████████████████████████████████████████████████████████████████▊ | 378/400 [03:01<00:12, 1.70it/s, v_num=1, train_loss=165]Epoch 379/400: 94%|████████████████████████████████████████████████████████████████████████▊ | 378/400 [03:01<00:12, 1.70it/s, v_num=1, train_loss=165]Epoch 379/400: 95%|████████████████████████████████████████████████████████████████████████▉ | 379/400 [03:02<00:12, 1.72it/s, v_num=1, train_loss=165]Epoch 379/400: 95%|████████████████████████████████████████████████████████████████████████▉ | 379/400 [03:02<00:12, 1.72it/s, v_num=1, train_loss=165]Epoch 380/400: 95%|████████████████████████████████████████████████████████████████████████▉ | 379/400 [03:02<00:12, 1.72it/s, v_num=1, train_loss=165]Epoch 380/400: 95%|█████████████████████████████████████████████████████████████████████████▏ | 380/400 [03:03<00:11, 1.69it/s, v_num=1, train_loss=165]Epoch 380/400: 95%|█████████████████████████████████████████████████████████████████████████▏ | 380/400 [03:03<00:11, 1.69it/s, v_num=1, train_loss=165]Epoch 381/400: 95%|█████████████████████████████████████████████████████████████████████████▏ | 380/400 [03:03<00:11, 1.69it/s, v_num=1, train_loss=165]Epoch 381/400: 95%|█████████████████████████████████████████████████████████████████████████▎ | 381/400 [03:03<00:11, 1.64it/s, v_num=1, train_loss=165]Epoch 381/400: 95%|█████████████████████████████████████████████████████████████████████████▎ | 381/400 [03:03<00:11, 1.64it/s, v_num=1, train_loss=165]Epoch 382/400: 95%|█████████████████████████████████████████████████████████████████████████▎ | 381/400 [03:03<00:11, 1.64it/s, v_num=1, train_loss=165]Epoch 382/400: 96%|█████████████████████████████████████████████████████████████████████████▌ | 382/400 [03:04<00:10, 1.67it/s, v_num=1, train_loss=165]Epoch 382/400: 96%|█████████████████████████████████████████████████████████████████████████▌ | 382/400 [03:04<00:10, 1.67it/s, v_num=1, train_loss=165]Epoch 383/400: 96%|█████████████████████████████████████████████████████████████████████████▌ | 382/400 [03:04<00:10, 1.67it/s, v_num=1, train_loss=165]Epoch 383/400: 96%|█████████████████████████████████████████████████████████████████████████▋ | 383/400 [03:04<00:09, 1.72it/s, v_num=1, train_loss=165]Epoch 383/400: 96%|█████████████████████████████████████████████████████████████████████████▋ | 383/400 [03:04<00:09, 1.72it/s, v_num=1, train_loss=165]Epoch 384/400: 96%|█████████████████████████████████████████████████████████████████████████▋ | 383/400 [03:04<00:09, 1.72it/s, v_num=1, train_loss=165]Epoch 384/400: 96%|█████████████████████████████████████████████████████████████████████████▉ | 384/400 [03:05<00:09, 1.71it/s, v_num=1, train_loss=165]Epoch 384/400: 96%|█████████████████████████████████████████████████████████████████████████▉ | 384/400 [03:05<00:09, 1.71it/s, v_num=1, train_loss=165]Epoch 385/400: 96%|█████████████████████████████████████████████████████████████████████████▉ | 384/400 [03:05<00:09, 1.71it/s, v_num=1, train_loss=165]Epoch 385/400: 96%|██████████████████████████████████████████████████████████████████████████ | 385/400 [03:06<00:09, 1.64it/s, v_num=1, train_loss=165]Epoch 385/400: 96%|██████████████████████████████████████████████████████████████████████████ | 385/400 [03:06<00:09, 1.64it/s, v_num=1, train_loss=165]Epoch 386/400: 96%|██████████████████████████████████████████████████████████████████████████ | 385/400 [03:06<00:09, 1.64it/s, v_num=1, train_loss=165]Epoch 386/400: 96%|██████████████████████████████████████████████████████████████████████████▎ | 386/400 [03:06<00:08, 1.68it/s, v_num=1, train_loss=165]Epoch 386/400: 96%|██████████████████████████████████████████████████████████████████████████▎ | 386/400 [03:06<00:08, 1.68it/s, v_num=1, train_loss=164]Epoch 387/400: 96%|██████████████████████████████████████████████████████████████████████████▎ | 386/400 [03:06<00:08, 1.68it/s, v_num=1, train_loss=164]Epoch 387/400: 97%|██████████████████████████████████████████████████████████████████████████▍ | 387/400 [03:07<00:07, 1.67it/s, v_num=1, train_loss=164]Epoch 387/400: 97%|██████████████████████████████████████████████████████████████████████████▍ | 387/400 [03:07<00:07, 1.67it/s, v_num=1, train_loss=165]Epoch 388/400: 97%|██████████████████████████████████████████████████████████████████████████▍ | 387/400 [03:07<00:07, 1.67it/s, v_num=1, train_loss=165]Epoch 388/400: 97%|██████████████████████████████████████████████████████████████████████████▋ | 388/400 [03:07<00:06, 1.73it/s, v_num=1, train_loss=165]Epoch 388/400: 97%|██████████████████████████████████████████████████████████████████████████▋ | 388/400 [03:07<00:06, 1.73it/s, v_num=1, train_loss=165]Epoch 389/400: 97%|██████████████████████████████████████████████████████████████████████████▋ | 388/400 [03:07<00:06, 1.73it/s, v_num=1, train_loss=165]Epoch 389/400: 97%|██████████████████████████████████████████████████████████████████████████▉ | 389/400 [03:08<00:06, 1.70it/s, v_num=1, train_loss=165]Epoch 389/400: 97%|██████████████████████████████████████████████████████████████████████████▉ | 389/400 [03:08<00:06, 1.70it/s, v_num=1, train_loss=164]Epoch 390/400: 97%|██████████████████████████████████████████████████████████████████████████▉ | 389/400 [03:08<00:06, 1.70it/s, v_num=1, train_loss=164]Epoch 390/400: 98%|███████████████████████████████████████████████████████████████████████████ | 390/400 [03:08<00:05, 1.77it/s, v_num=1, train_loss=164]Epoch 390/400: 98%|███████████████████████████████████████████████████████████████████████████ | 390/400 [03:08<00:05, 1.77it/s, v_num=1, train_loss=164]Epoch 391/400: 98%|███████████████████████████████████████████████████████████████████████████ | 390/400 [03:08<00:05, 1.77it/s, v_num=1, train_loss=164]Epoch 391/400: 98%|███████████████████████████████████████████████████████████████████████████▎ | 391/400 [03:09<00:05, 1.79it/s, v_num=1, train_loss=164]Epoch 391/400: 98%|███████████████████████████████████████████████████████████████████████████▎ | 391/400 [03:09<00:05, 1.79it/s, v_num=1, train_loss=165]Epoch 392/400: 98%|███████████████████████████████████████████████████████████████████████████▎ | 391/400 [03:09<00:05, 1.79it/s, v_num=1, train_loss=165]Epoch 392/400: 98%|███████████████████████████████████████████████████████████████████████████▍ | 392/400 [03:09<00:04, 1.79it/s, v_num=1, train_loss=165]Epoch 392/400: 98%|███████████████████████████████████████████████████████████████████████████▍ | 392/400 [03:09<00:04, 1.79it/s, v_num=1, train_loss=165]Epoch 393/400: 98%|███████████████████████████████████████████████████████████████████████████▍ | 392/400 [03:09<00:04, 1.79it/s, v_num=1, train_loss=165]Epoch 393/400: 98%|███████████████████████████████████████████████████████████████████████████▋ | 393/400 [03:10<00:03, 1.84it/s, v_num=1, train_loss=165]Epoch 393/400: 98%|███████████████████████████████████████████████████████████████████████████▋ | 393/400 [03:10<00:03, 1.84it/s, v_num=1, train_loss=165]Epoch 394/400: 98%|███████████████████████████████████████████████████████████████████████████▋ | 393/400 [03:10<00:03, 1.84it/s, v_num=1, train_loss=165]Epoch 394/400: 98%|███████████████████████████████████████████████████████████████████████████▊ | 394/400 [03:11<00:03, 1.85it/s, v_num=1, train_loss=165]Epoch 394/400: 98%|███████████████████████████████████████████████████████████████████████████▊ | 394/400 [03:11<00:03, 1.85it/s, v_num=1, train_loss=165]Epoch 395/400: 98%|███████████████████████████████████████████████████████████████████████████▊ | 394/400 [03:11<00:03, 1.85it/s, v_num=1, train_loss=165]Epoch 395/400: 99%|████████████████████████████████████████████████████████████████████████████ | 395/400 [03:11<00:02, 1.84it/s, v_num=1, train_loss=165]Epoch 395/400: 99%|████████████████████████████████████████████████████████████████████████████ | 395/400 [03:11<00:02, 1.84it/s, v_num=1, train_loss=165]Epoch 396/400: 99%|████████████████████████████████████████████████████████████████████████████ | 395/400 [03:11<00:02, 1.84it/s, v_num=1, train_loss=165]Epoch 396/400: 99%|████████████████████████████████████████████████████████████████████████████▏| 396/400 [03:12<00:02, 1.83it/s, v_num=1, train_loss=165]Epoch 396/400: 99%|████████████████████████████████████████████████████████████████████████████▏| 396/400 [03:12<00:02, 1.83it/s, v_num=1, train_loss=165]Epoch 397/400: 99%|████████████████████████████████████████████████████████████████████████████▏| 396/400 [03:12<00:02, 1.83it/s, v_num=1, train_loss=165]Epoch 397/400: 99%|████████████████████████████████████████████████████████████████████████████▍| 397/400 [03:12<00:01, 1.85it/s, v_num=1, train_loss=165]Epoch 397/400: 99%|████████████████████████████████████████████████████████████████████████████▍| 397/400 [03:12<00:01, 1.85it/s, v_num=1, train_loss=165]Epoch 398/400: 99%|████████████████████████████████████████████████████████████████████████████▍| 397/400 [03:12<00:01, 1.85it/s, v_num=1, train_loss=165]Epoch 398/400: 100%|████████████████████████████████████████████████████████████████████████████▌| 398/400 [03:13<00:01, 1.84it/s, v_num=1, train_loss=165]Epoch 398/400: 100%|████████████████████████████████████████████████████████████████████████████▌| 398/400 [03:13<00:01, 1.84it/s, v_num=1, train_loss=165]Epoch 399/400: 100%|████████████████████████████████████████████████████████████████████████████▌| 398/400 [03:13<00:01, 1.84it/s, v_num=1, train_loss=165]Epoch 399/400: 100%|████████████████████████████████████████████████████████████████████████████▊| 399/400 [03:13<00:00, 1.86it/s, v_num=1, train_loss=165]Epoch 399/400: 100%|████████████████████████████████████████████████████████████████████████████▊| 399/400 [03:13<00:00, 1.86it/s, v_num=1, train_loss=165]Epoch 400/400: 100%|████████████████████████████████████████████████████████████████████████████▊| 399/400 [03:13<00:00, 1.86it/s, v_num=1, train_loss=165]Epoch 400/400: 100%|█████████████████████████████████████████████████████████████████████████████| 400/400 [03:14<00:00, 1.85it/s, v_num=1, train_loss=165]Epoch 400/400: 100%|█████████████████████████████████████████████████████████████████████████████| 400/400 [03:14<00:00, 1.85it/s, v_num=1, train_loss=165]Epoch 400/400: 100%|█████████████████████████████████████████████████████████████████████████████| 400/400 [03:14<00:00, 2.06it/s, v_num=1, train_loss=165]
#> Iteration 1/10
#> Iteration 2/10
#> Iteration 3/10
#> Iteration 4/10
#> Iteration 5/10
#> Iteration 6/10
#> Iteration 7/10
#> Iteration 8/10
#> Iteration 9/10
#> Iteration 10/10
#> Optimal number of clusters: 5data$cellcharter_clusters = as.character(data$cellcharter_clusters)
KanPlot(data, feature = 'cellcharter_clusters',pt.size=2) + labs(fill = 'CellCharter cluster')
8) Co-localisation / dispersion analysis
Kandinsky can test whether pairs of cell/spot types tend to
co-localise or spatially avoid each other, using a join-count based
statistic implemented in jc_coloc(). Positive z-values
indicate co-localisation, negative values indicate dispersion:
| Joincount | Expected | Variance | z-value | Type1 | Type2 | oe_log2ratio | oe_ratio | |
|---|---|---|---|---|---|---|---|---|
| Layer1:Layer1 | 686 | 60.46109 | 51.94399 | 86.79340 | Layer1 | Layer1 | 3.504130 | 11.346141 |
| Layer2:Layer2 | 541 | 51.91172 | 45.11390 | 72.81688 | Layer2 | Layer2 | 3.381496 | 10.421538 |
| Layer3:Layer3 | 2734 | 795.60523 | 430.78783 | 93.39218 | Layer3 | Layer3 | 1.780889 | 3.436378 |
| Layer4:Layer4 | 457 | 38.51772 | 34.14773 | 71.61370 | Layer4 | Layer4 | 3.568600 | 11.864668 |
| Layer5:Layer5 | 1824 | 368.23811 | 247.70635 | 92.49575 | Layer5 | Layer5 | 2.308395 | 4.953317 |
| Layer6:Layer6 | 1920 | 389.33956 | 258.67294 | 95.17072 | Layer6 | Layer6 | 2.302005 | 4.931428 |
As expected for a layered tissue, each layer strongly co-localises with itself (high positive diagonal z-values), reflecting the spatial contiguity of the simulated cortical layers. Missing dots in the final colocalisation diagonal heatmap (e.g. Layer1-WM pair) reflects the complete absence of any cells/spots of two groups lying in each other’s neighbourhood.
Alternative co-localisation methods
jc_coloc() is Kandinsky’s native join-count based
colocalisation test, but the package also provides wrappers around
widely-used external tools sharing the same
label/return.mat interface and producing a
comparable heatmap, useful for benchmarking:
hoodscanR_coloc(): computes colocalisation from neighbourhood probability profiles using the Bioconductor packagehoodscanR(pure R, no Python required).giotto_coloc(): cell-proximity enrichment test from theGiottopackage.squidpy_coloc()/cellcharter_coloc(): neighbourhood enrichment tests from the Python packagessquidpy/CellCharter, called throughreticulate(again usingpython_path = pypath).
# Requires hoodscanR to be installed
coloc_hs = hoodscanR_coloc(data, label = 'label', k=50, return.mat = TRUE)
coloc_hs$plot9) Hot and cold expression areas
Kandinsky can identify spatial hot-spots (locally
high values) and cold-spots (locally low values) for
any numeric feature - a gene, a protein, or a signature score - using
Getis-Ord Gi* statistics via hotspot_analysis(). Continuing
with MOBP (ENSG00000168314), we expect hot-spots
in the WM region:
data = hotspot_analysis(data, feature = 'ENSG00000168314', layer = 'data',
sim = 499, padj.thresh = 0.05)
table(data$ENSG00000168314_clust)
#>
#> Cold Hot NS
#> 0 464 3147KanPlot(data, feature = 'ENSG00000168314_clust',pt.size=2,
palette_discrete = c('Cold' = 'dodgerblue3', 'Hot' = 'firebrick3', 'NS' = '#D9D9D9')) +
labs(fill = 'MOBP area')
We can cross-tabulate hot/cold status against the simulated layer annotation to confirm that MOBP hot-spots are indeed concentrated in the white matter:
table(data$label, data$ENSG00000168314_clust)
#>
#> Cold Hot NS
#> Layer1 0 0 273
#> Layer2 0 0 253
#> Layer3 0 0 989
#> Layer4 0 0 218
#> Layer5 0 0 673
#> Layer6 0 36 656
#> WM 0 428 85If more than one hot/cold annotation is defined within the same
dataset, hotspot_overlap() can be used to create a combined
annotation describing the overlap between the two.
10) Changing the neighbourhood definition
The neighbourhood definition used throughout the analysis is not
fixed: it can be changed at any point using nb_update().
For instance, we can switch from a Delaunay definition
(soi=T) to a centroid-distance based definition, connecting
all spots within 20 distance units of each other
(nb.method = 'C', d.max = 20), corresponding
to ~3 layers of surrounding spots:
Any downstream function relying on the neighbourhood network
(nnMat(), nn_query(),
nbCluster(), hotspot_analysis(), …) will now
use this updated definition.
Switching to a distance-based method typically increases and
unbalances neighbourhood sizes, as confirmed by
nb_summary(), although this becomes more evident for
single-cell resolution data:
nb_summary(data)
#> Neighbour method: Centroid distance (C), d.max = 20
#> Total cells/spots: 3611
#> Total neighbour links: 122660
#> Minimum number of neighbours per cell/spot: 1
#> Mean number of neighbours per cell/spot: 33.97
#> Median number of neighbours per cell/spot: 36
#> Max number of neighbours per cell/spot: 36
#> Number of cells/spots with no neighbours: 011) Downsizing dense neighbourhoods
Distance-based neighbour definitions
(nb.method = 'C'/'M'/'Q') can
produce very large, unevenly-sized c/s-NBs in densely packed regions of
a sample, which can slow down or bias downstream analyses
(e.g. nnMat(), nbCluster()).
nb_downsize() addresses this by randomly subsampling each
c/s-NB down to an expected maximum size (exp_links), while
trying to preserve the overall network structure:
data_ds = nb_downsize(data, exp_links = 6)
#> Current median number of neighbour links per cell is 36. Trying to reduce to 6...
#> Enforcing symmetric neighbourhoods...
#> Trying to trim degrees > 6...
#> Returning updated neighour network with a median number of 7 links per cell
nb_summary(data_ds)
#> Neighbour method: Centroid distance (C), d.max = 20
#> Total cells/spots: 3611
#> Total neighbour links: 26300
#> Minimum number of neighbours per cell/spot: 1
#> Mean number of neighbours per cell/spot: 7.28
#> Median number of neighbours per cell/spot: 7
#> Max number of neighbours per cell/spot: 13
#> Number of cells/spots with no neighbours: 0Neighbour relationships created with symmetric methods
('Q'/'C'/'M') are, by default,
kept symmetric after downsizing (keep_symmetric = TRUE) -
meaning if spot A is downsized to keep spot B as a neighbour, spot B
will also keep spot A - so some neighbourhoods may still end up slightly
larger than exp_links after this correction. This behaviour
can be disabled by setting keep_symmetric = FALSE.
Next steps
This vignette covered the core Kandinsky building blocks:
initializing spatial neighbourhoods, visualizing data and tissue region
masks, summarizing and querying neighbourhood composition,
neighbourhood-based clustering, testing co-localisation/dispersion,
detecting hot/cold expression areas, and reshaping the neighbourhood
network itself (nb_update(),
nb_downsize()).
Kandinsky also provides:
- platform-specific loaders (
prepare_cosmx_seurat(),prepare_xenium_seurat(),prepare_seurat_other(), …) for CosMx, Xenium, MERSCOPE, IMC and other tabular spatial data - see themodC/modD/modEvignettes for full worked examples on real datasets. - optional wrappers around external, non-native clustering
(
cellcharter_clust(),spatialleiden_clust(),mesa_clust(),spatopic_clust(),hoodscanR_clust(),giotto_clust()) and colocalisation (squidpy_coloc(),cellcharter_coloc(),hoodscanR_coloc(),giotto_coloc()) methods, for users who want to benchmark Kandinsky against other tools. The Python-based ones require a Python environment with the relevant dependencies installed and accessible viareticulate(this is what thepypathvariable at the top of this vignette would be used for).
sessionInfo()
#> R version 4.5.3 (2026-03-11)
#> Platform: x86_64-conda-linux-gnu
#> Running under: CentOS Linux 7 (Core)
#>
#> Matrix products: default
#> BLAS/LAPACK: /SAN/colcc/ITH2/Pietro/miniconda3/envs/kandinskybench/lib/libopenblasp-r0.3.33.so; LAPACK version 3.12.0
#>
#> locale:
#> [1] LC_CTYPE=en_GB.UTF-8 LC_NUMERIC=C
#> [3] LC_TIME=en_GB.UTF-8 LC_COLLATE=en_GB.UTF-8
#> [5] LC_MONETARY=en_GB.UTF-8 LC_MESSAGES=en_GB.UTF-8
#> [7] LC_PAPER=en_GB.UTF-8 LC_NAME=C
#> [9] LC_ADDRESS=C LC_TELEPHONE=C
#> [11] LC_MEASUREMENT=en_GB.UTF-8 LC_IDENTIFICATION=C
#>
#> time zone: Europe/London
#> tzcode source: system (glibc)
#>
#> attached base packages:
#> [1] stats graphics grDevices utils datasets methods base
#>
#> other attached packages:
#> [1] presto_1.0.0 data.table_1.17.8 Rcpp_1.1.1-1.1 ggplot2_4.0.3
#> [5] spdep_1.4-2 spData_2.3.5 sf_1.1-0 dplyr_1.2.1
#> [9] Kandinsky_1.1.0 Seurat_5.5.0 SeuratObject_5.4.0 sp_2.2-1
#>
#> loaded via a namespace (and not attached):
#> [1] RColorBrewer_1.1-3 jsonlite_2.0.0 wk_0.9.5
#> [4] magrittr_2.0.5 TH.data_1.1-5 spatstat.utils_3.2-3
#> [7] farver_2.1.2 rmarkdown_2.31 vctrs_0.7.3
#> [10] ROCR_1.0-12 spatstat.explore_3.8-1 terra_1.9-34
#> [13] htmltools_0.5.9 curl_7.1.0 s2_1.1.9
#> [16] LearnBayes_2.15.2 sass_0.4.10 sctransform_0.4.3
#> [19] parallelly_1.47.0 KernSmooth_2.23-26 bslib_0.11.0
#> [22] htmlwidgets_1.6.4 ica_1.0-3 sandwich_3.1-1
#> [25] plyr_1.8.9 plotly_4.12.0 zoo_1.8-15
#> [28] cachem_1.1.0 igraph_2.3.2 mime_0.13
#> [31] lifecycle_1.0.5 pkgconfig_2.0.3 anndataR_1.0.2
#> [34] Matrix_1.7-5 R6_2.6.1 fastmap_1.2.0
#> [37] fitdistrplus_1.2-6 future_1.70.0 shiny_1.13.0
#> [40] digest_0.6.39 patchwork_1.3.2 tensor_1.5.1
#> [43] RSpectra_0.16-2 irlba_2.3.7 labeling_0.4.3
#> [46] progressr_0.19.0 spatstat.sparse_3.2-0 spatialreg_1.4-3
#> [49] httr_1.4.8 polyclip_1.10-7 abind_1.4-8
#> [52] compiler_4.5.3 proxy_0.4-29 withr_3.0.2
#> [55] bit64_4.8.2 backports_1.5.1 S7_0.2.2
#> [58] viridis_0.6.5 DBI_1.3.0 fastDummies_1.7.6
#> [61] MASS_7.3-65 concaveman_1.2.0 classInt_0.4-11
#> [64] tripack_1.3-9.4 tools_4.5.3 units_1.0-1
#> [67] lmtest_0.9-40 otel_0.2.0 httpuv_1.6.17
#> [70] future.apply_1.20.2 goftest_1.2-3 glue_1.8.1
#> [73] dbscan_1.2.5 nlme_3.1-169 promises_1.5.0
#> [76] grid_4.5.3 Rtsne_0.17 cluster_2.1.8.2
#> [79] reshape2_1.4.5 generics_0.1.4 gtable_0.3.6
#> [82] spatstat.data_3.1-9 class_7.3-23 tidyr_1.3.2
#> [85] spatstat.geom_3.8-1 RcppAnnoy_0.0.23 ggrepel_0.9.8
#> [88] RANN_2.6.2 pillar_1.11.1 stringr_1.6.0
#> [91] limma_3.66.0 spam_2.11-4 RcppHNSW_0.7.0
#> [94] later_1.4.8 splines_4.5.3 lattice_0.22-9
#> [97] bit_4.6.0 survival_3.8-6 deldir_2.0-4
#> [100] tidyselect_1.2.1 sfheaders_0.4.5 miniUI_0.1.2
#> [103] pbapply_1.7-4 knitr_1.51 gridExtra_2.3
#> [106] V8_8.2.0 scattermore_1.2 xfun_0.57
#> [109] statmod_1.5.2 matrixStats_1.5.0 stringi_1.8.7
#> [112] lazyeval_0.2.3 yaml_2.3.12 boot_1.3-32
#> [115] evaluate_1.0.5 codetools_0.2-20 tibble_3.3.1
#> [118] cli_3.6.6 uwot_0.2.4 arrow_24.0.0
#> [121] xtable_1.8-8 reticulate_1.46.0 jquerylib_0.1.4
#> [124] globals_0.19.1 spatstat.random_3.5-0 coda_0.19-4.1
#> [127] png_0.1-9 spatstat.univar_3.2-0 parallel_4.5.3
#> [130] assertthat_0.2.1 mclust_6.1.2 dotCall64_1.2
#> [133] marginaleffects_0.32.0 listenv_0.10.1 mvtnorm_1.4-1
#> [136] viridisLite_0.4.3 scales_1.4.0 ggridges_0.5.7
#> [139] e1071_1.7-17 purrr_1.2.2 rlang_1.2.0
#> [142] multcomp_1.4-30 cowplot_1.2.0