Amanarora's workspace
Runs
70
Name
69 visualized
auc_score.max
0.89768
0.88347
0.8821
0.8815
0.87082
0.86974
0.86617
0.86082
0.85744
0.85532
0.85251
0.85188
0.8496
0.84917
0.84848
0.84622
0.84593
0.84522
0.84481
0.84424
State
Notes
User
Tags
Created
Runtime
Sweep
accumulation_steps
arch_name
cc
device
epochs
freeze_cnn
kfold
learning_rate
loss
metric
model_name
pretrained
sz
train_batch_size
train_data_dir
training_folds_csv
tta
use_metadata
valid_batch_size
image size
learning rate
model name
save_mode
training augmentation
valid augmentation
valid_sz
auc_score
learning_rate
train_loss
valid_loss
valid_loss.min
LR
Finished
The best performing experiment with Weighted Focal Loss
amanarora
22m 9s
-
1
efficientnet-b0
false
cuda
25
false
0
0.0005
weighted_focal_loss
auc
efficient_net
imagenet
224
64
/home/arora/git_repos/melanoma_wandb/data/usr/resized_train_256_cc/
/home/arora/git_repos/melanoma_wandb/data/train_folds.csv
false
false
128
-
-
-
best
-
-
224
-
0.00014878
0.0035516
-
0.0049962
-
Finished
Uses "EfficientNet-b1" instead of "EfficientNet-B0"
amanarora
10m 2s
3
efficientnet-b1
false
cuda
19
false
0
0.0011085
weighted_focal_loss
auc
efficient_net
imagenet
224
37
/home/arora/git_repos/melanoma_wandb/data/usr/resized_train_256
/home/arora/git_repos/melanoma_wandb/data/train_folds.csv
false
false
41
-
-
-
all
-
-
210
-
0.000031421
0.001655
-
0.0065297
-
Finished
-
amanarora
6m 5s
1
efficientnet-b1
false
cuda
17
false
0
0.00046097
bce
auc
efficient_net
imagenet
224
49
/home/arora/git_repos/melanoma_wandb/data/usr/resized_train_256_cc/
/home/arora/git_repos/melanoma_wandb/data/train_folds.csv
false
false
44
-
-
-
all
-
-
216
-
0.00019345
0.047863
-
0.10731
-
Finished
With Color Constancy
amanarora
4m 13s
-
1
efficientnet-b0
false
cuda
10
false
0
0.0005
weighted_focal_loss
auc
efficient_net
imagenet
224
64
/home/arora/git_repos/melanoma_wandb/data/usr/resized_train_256_cc/
/home/arora/git_repos/melanoma_wandb/data/train_folds.csv
false
false
64
-
-
-
-
-
-
-
-
0.0000078125
0.002919
-
0.035516
-
Finished
Uses Cosine LR
amanarora
4m 15s
-
1
efficientnet-b0
false
cuda
10
false
0
0.0005
weighted_focal_loss
auc
efficient_net
imagenet
224
64
/home/arora/git_repos/melanoma_wandb/data/usr/resized_train_256_cc/
/home/arora/git_repos/melanoma_wandb/data/train_folds.csv
false
false
64
-
-
-
best
-
-
224
-
0.000016418
0.0024614
-
0.031646
-
Finished
-
amanarora
6m 28s
1
efficientnet-b1
false
cuda
14
false
0
0.0013668
bce
auc
efficient_net
imagenet
224
57
/home/arora/git_repos/melanoma_wandb/data/usr/resized_train_256
/home/arora/git_repos/melanoma_wandb/data/train_folds.csv
false
false
39
-
-
-
all
-
-
204
-
0.00018402
0.054863
-
0.069788
-
Finished
-
amanarora
5m 57s
2
efficientnet-b0
false
cuda
17
false
0
0.00011759
weighted_focal_loss
auc
efficient_net
imagenet
224
37
/home/arora/git_repos/melanoma_wandb/data/usr/resized_train_256
/home/arora/git_repos/melanoma_wandb/data/train_folds.csv
false
false
50
-
-
-
all
-
-
212
-
0.000026083
0.0024678
-
0.0060893
-
Finished
-
amanarora
6m 14s
2
efficientnet-b2
false
cuda
13
false
0
0.000095461
weighted_focal_loss
auc
efficient_net
imagenet
224
45
/home/arora/git_repos/melanoma_wandb/data/usr/resized_train_256
/home/arora/git_repos/melanoma_wandb/data/train_folds.csv
false
false
43
-
-
-
all
-
-
222
-
0.000014409
0.0034336
-
0.01153
-
Finished
-
amanarora
4m 13s
4
efficientnet-b1
false
cuda
7
false
0
0.0016424
bce
auc
efficient_net
imagenet
224
62
/home/arora/git_repos/melanoma_wandb/data/usr/resized_train_256_cc/
/home/arora/git_repos/melanoma_wandb/data/train_folds.csv
false
false
56
-
-
-
all
-
-
223
-
4.6563e-7
0.055076
-
0.081324
-
Finished
-
amanarora
6m 41s
2
efficientnet-b0
false
cuda
17
false
0
0.0017803
weighted_focal_loss
auc
efficient_net
imagenet
192
38
/home/arora/git_repos/melanoma_wandb/data/usr/resized_train_256
/home/arora/git_repos/melanoma_wandb/data/train_folds.csv
false
false
42
-
-
-
all
-
-
199
-
0.000064573
0.0041493
-
0.0086715
-
Finished
-
amanarora
3m 27s
-
1
efficientnet-b0
false
cuda
10
false
0
0.001
weighted_focal_loss
auc
efficient_net
imagenet
224
64
/home/arora/git_repos/melanoma_wandb/data/usr/resized_train_256_cc/
/home/arora/git_repos/melanoma_wandb/data/train_folds.csv
false
false
64
-
-
-
best
-
-
-
-
0.0000625
0.0040871
-
0.031249
-
Finished
-
amanarora
5m 47s
1
efficientnet-b2
false
cuda
20
false
0
0.00064031
weighted_focal_loss
auc
efficient_net
imagenet
224
49
/home/arora/git_repos/melanoma_wandb/data/usr/resized_train_256
/home/arora/git_repos/melanoma_wandb/data/train_folds.csv
false
false
38
-
-
-
all
-
-
210
-
0.00041207
0.0043047
-
0.0088484
-
Finished
-
amanarora
3m 43s
-
1
efficientnet-b1
false
cuda
10
false
0
0.0005
weighted_focal_loss
auc
efficient_net
imagenet
224
64
/home/arora/git_repos/melanoma_wandb/data/usr/resized_train_256_cc/
/home/arora/git_repos/melanoma_wandb/data/train_folds.csv
false
false
64
-
-
-
-
-
-
-
-
0.0000625
0.0043238
-
0.031384
-
Finished
-
amanarora
3m 30s
1
efficientnet-b1
false
cuda
6
false
0
0.00081906
bce
auc
efficient_net
imagenet
192
45
/home/arora/git_repos/melanoma_wandb/data/usr/resized_train_256
/home/arora/git_repos/melanoma_wandb/data/train_folds.csv
false
false
42
-
-
-
all
-
-
215
-
1.1034e-7
0.065052
-
0.087239
-
Finished
-
amanarora
4m 28s
2
efficientnet-b0
false
cuda
20
false
0
0.00021285
weighted_focal_loss
auc
efficient_net
imagenet
224
46
/home/arora/git_repos/melanoma_wandb/data/usr/resized_train_256_cc/
/home/arora/git_repos/melanoma_wandb/data/train_folds.csv
false
false
33
-
-
-
all
-
-
216
-
0.00013613
0.002806
-
0.014068
-
Finished
Without Color Constancy
amanarora
4m 39s
-
1
efficientnet-b0
false
cuda
10
false
0
0.0005
weighted_focal_loss
auc
efficient_net
imagenet
224
64
/home/arora/git_repos/melanoma_wandb/data/usr/resized_train_256
/home/arora/git_repos/melanoma_wandb/data/train_folds.csv
false
false
64
-
-
-
-
-
-
-
-
0.0000039063
0.0029619
-
0.031862
-
Finished
-
amanarora
3m 21s
2
efficientnet-b0
false
cuda
12
false
0
0.00053574
weighted_focal_loss
auc
efficient_net
imagenet
192
58
/home/arora/git_repos/melanoma_wandb/data/usr/resized_train_256_cc/
/home/arora/git_repos/melanoma_wandb/data/train_folds.csv
false
false
42
-
-
-
all
-
-
222
-
0.00025591
0.0046418
-
0.0086452
-
Finished
-
amanarora
5m 28s
1
efficientnet-b0
false
cuda
13
false
0
0.0015817
weighted_focal_loss
auc
efficient_net
imagenet
224
45
/home/arora/git_repos/melanoma_wandb/data/usr/resized_train_256
/home/arora/git_repos/melanoma_wandb/data/train_folds.csv
false
false
56
-
-
-
all
-
-
203
-
0.00010666
0.0048675
-
0.006537
-
Finished
-
amanarora
3m 50s
-
1
efficientnet-b0
false
cuda
10
false
0
0.0005
weighted_focal_loss
auc
efficient_net
imagenet
224
64
/home/arora/git_repos/melanoma_wandb/data/usr/resized_train_256_cc/
/home/arora/git_repos/melanoma_wandb/data/train_folds.csv
false
false
64
-
-
-
best
-
-
224
-
-
0.0033602
-
0.019335
0.000069379
Finished
-
amanarora
2m 47s
2
efficientnet-b0
false
cuda
5
false
0
0.00077138
weighted_focal_loss
auc
efficient_net
imagenet
224
44
/home/arora/git_repos/melanoma_wandb/data/usr/resized_train_256_cc/
/home/arora/git_repos/melanoma_wandb/data/train_folds.csv
false
false
40
-
-
-
all
-
-
224
-
1.1090e-7
0.0040504
-
0.0067895
-
1-20
of 70