Converts DALI into a DataModule, improves it and cleans stuff a bit.

README.md

@@ -15,6 +15,7 @@ While the library is self-contained, it is possible to use the models outside of
 ---
 
 ## News
+* **[May 02 2022]**: :diamond_shape_with_a_dot_inside: Wrapped Dali with a DataModule, added auto resume for linear eval and Wandb run resume.
 * **[Apr 12 2022]**: :rainbow: Improved design of models and added support to train with a fraction of data.
 * **[Apr 01 2022]**: :mag: Added the option to use [channel last conversion](https://pytorch.org/tutorials/intermediate/memory_format_tutorial.html#converting-existing-models) which considerably decreases training times.
 * **[Feb 04 2022]**: :partying_face: Paper got accepted to JMLR.

bash_files/knn/imagenet-100/knn.sh

@@ -1,4 +1,4 @@
-python3 ../../../main_knn.py \
+python3 main_knn.py \
     --dataset imagenet100 \
     --data_dir /datasets \
     --train_dir imagenet-100/train \

bash_files/linear/imagenet-100/barlow_linear.sh

@@ -1,4 +1,4 @@
-python3 ../../../main_linear.py \
+python3 main_linear.py \
     --dataset imagenet100 \
     --backbone resnet18 \
     --data_dir /datasets \
@@ -20,4 +20,6 @@ python3 ../../../main_linear.py \
     --pretrained_feature_extractor PATH \
     --project solo-learn \
     --entity unitn-mhug \
-    --wandb
+    --wandb \
+    --save_checkpoint \
+    --auto_resume

bash_files/linear/imagenet-100/byol_linear.sh

@@ -1,4 +1,4 @@
-python3 ../../../main_linear.py \
+python3 main_linear.py \
     --dataset imagenet100 \
     --backbone resnet18 \
     --data_dir /datasets \
@@ -20,4 +20,6 @@ python3 ../../../main_linear.py \
     --pretrained_feature_extractor PATH \
     --project solo-learn \
     --entity unitn-mhug \
-    --wandb
+    --wandb \
+    --save_checkpoint \
+    --auto_resume

bash_files/linear/imagenet-100/deepclusterv2_linear.sh

@@ -1,4 +1,4 @@
-python3 ../../../main_linear.py \
+python3 main_linear.py \
     --dataset imagenet100 \
     --backbone resnet18 \
     --data_dir /data/datasets \
@@ -19,4 +19,6 @@ python3 ../../../main_linear.py \
     --name deepclusterv2-imagenet100-linear-eval \
     --pretrained_feature_extractor PATH --project solo-learn \
     --entity unitn-mhug \
-    --wandb
+    --wandb \
+    --save_checkpoint \
+    --auto_resume

bash_files/linear/imagenet-100/dino_linear.sh

@@ -1,4 +1,4 @@
-python3 ../../../main_linear.py \
+python3 main_linear.py \
     --dataset imagenet100 \
     --backbone resnet18 \
     --data_dir /datasets \
@@ -20,4 +20,6 @@ python3 ../../../main_linear.py \
     --pretrained_feature_extractor PATH \
     --project solo-learn \
     --entity unitn-mhug \
-    --wandb
+    --wandb \
+    --save_checkpoint \
+    --auto_resume

bash_files/linear/imagenet-100/general_linear.sh

@@ -1,4 +1,4 @@
-python3 ../../../main_linear.py \
+python3 main_linear.py \
     --dataset imagenet100 \
     --backbone resnet18 \
     --data_dir /datasets \
@@ -20,4 +20,6 @@ python3 ../../../main_linear.py \
     --name simclr-linear-eval \
     --pretrained_feature_extractor PATH \
     --project contrastive_learning \
-    --wandb
+    --wandb \
+    --save_checkpoint \
+    --auto_resume

bash_files/linear/imagenet-100/mocov2plus_linear.sh

@@ -1,4 +1,4 @@
-python3 ../../../main_linear.py \
+python3 main_linear.py \
     --dataset imagenet100 \
     --backbone resnet18 \
     --data_dir /datasets \
@@ -20,4 +20,6 @@ python3 ../../../main_linear.py \
     --pretrained_feature_extractor PATH \
     --project solo-learn \
     --entity unitn-mhug \
-    --wandb
+    --wandb \
+    --save_checkpoint \
+    --auto_resume

bash_files/linear/imagenet-100/nnclr_linear.sh

@@ -1,4 +1,4 @@
-python3 ../../../main_linear.py \
+python3 main_linear.py \
     --dataset imagenet100 \
     --backbone resnet18 \
     --data_dir /datasets \
@@ -20,4 +20,6 @@ python3 ../../../main_linear.py \
     --pretrained_feature_extractor PATH \
     --project solo-learn \
     --entity unitn-mhug \
-    --wandb
+    --wandb \
+    --save_checkpoint \
+    --auto_resume

bash_files/linear/imagenet-100/ressl_linear.sh

@@ -1,4 +1,4 @@
-python3 ../../../main_linear.py \
+python3 main_linear.py \
     --dataset imagenet100 \
     --backbone resnet18 \
     --data_dir /datasets \
@@ -20,4 +20,6 @@ python3 ../../../main_linear.py \
     --pretrained_feature_extractor PATH \
     --project solo-learn \
     --entity unitn-mhug \
-    --wandb
+    --wandb \
+    --save_checkpoint \
+    --auto_resume

bash_files/linear/imagenet-100/simclr_linear.sh

@@ -1,4 +1,4 @@
-python3 ../../../main_linear.py \
+python3 main_linear.py \
     --dataset imagenet100 \
     --backbone resnet18 \
     --data_dir /data/datasets \
@@ -20,4 +20,6 @@ python3 ../../../main_linear.py \
     --pretrained_feature_extractor PATH \
     --project solo-learn \
     --entity unitn-mhug \
-    --wandb
+    --wandb \
+    --save_checkpoint \
+    --auto_resume

bash_files/linear/imagenet-100/simsiam_linear.sh

@@ -1,4 +1,4 @@
-python3 ../../../main_linear.py \
+python3 main_linear.py \
     --dataset imagenet100 \
     --backbone resnet18 \
     --data_dir /datasets \
@@ -20,4 +20,6 @@ python3 ../../../main_linear.py \
     --pretrained_feature_extractor PATH \
     --project solo-learn \
     --entity unitn-mhug \
-    --wandb
+    --wandb \
+    --save_checkpoint \
+    --auto_resume

bash_files/linear/imagenet-100/swav_linear.sh

@@ -1,4 +1,4 @@
-python3 ../../../main_linear.py \
+python3 main_linear.py \
     --dataset imagenet100 \
     --backbone resnet18 \
     --data_dir /data/datasets \
@@ -20,4 +20,6 @@ python3 ../../../main_linear.py \
     --pretrained_feature_extractor PATH \
     --project solo-learn \
     --entity unitn-mhug \
-    --wandb
+    --wandb \
+    --save_checkpoint \
+    --auto_resume

bash_files/linear/imagenet-100/vibcreg_linear.sh

@@ -1,4 +1,4 @@
-python3 ../../../main_linear.py \
+python3 main_linear.py \
     --dataset imagenet100 \
     --backbone resnet18 \
     --data_dir /datasets \
@@ -20,4 +20,6 @@ python3 ../../../main_linear.py \
     --pretrained_feature_extractor PATH \
     --project solo-learn \
     --entity unitn-mhug \
-    --wandb
+    --wandb \
+    --save_checkpoint \
+    --auto_resume

bash_files/linear/imagenet-100/vicreg_linear.sh

@@ -1,4 +1,4 @@
-python3 ../../../main_linear.py \
+python3 main_linear.py \
     --dataset imagenet100 \
     --backbone resnet18 \
     --data_dir /datasets \
@@ -20,4 +20,6 @@ python3 ../../../main_linear.py \
     --pretrained_feature_extractor PATH \
     --project solo-learn \
     --entity unitn-mhug \
-    --wandb
+    --wandb \
+    --save_checkpoint \
+    --auto_resume

bash_files/linear/imagenet/barlow.sh

@@ -1,4 +1,4 @@
-python3 ../../../main_linear.py \
+python3 main_linear.py \
     --dataset imagenet \
     --backbone resnet50 \
     --data_dir /data/datasets \
@@ -22,4 +22,6 @@ python3 ../../../main_linear.py \
     --name barlow-resnet50-imagenet-linear-eval \
     --entity unitn-mhug \
     --project solo-learn \
-    --wandb
+    --wandb \
+    --save_checkpoint \
+    --auto_resume

bash_files/linear/imagenet/byol.sh

@@ -1,4 +1,4 @@
-python3 ../../../main_linear.py \
+python3 main_linear.py \
     --dataset imagenet \
     --backbone resnet50 \
     --data_dir /data/datasets \
@@ -22,4 +22,6 @@ python3 ../../../main_linear.py \
     --name byol-resnet50-imagenet-linear-eval \
     --entity unitn-mhug \
     --project solo-learn \
-    --wandb
+    --wandb \
+    --save_checkpoint \
+    --auto_resume

bash_files/linear/imagenet/mocov2plus.sh

@@ -21,4 +21,5 @@ python3 main_linear.py \
     --project solo-learn \
     --entity unitn-mhug \
     --wandb \
-    --save_checkpoint
+    --save_checkpoint \
+    --auto_resume

bash_files/pretrain/cifar/barlow.sh

@@ -1,4 +1,4 @@
-python3 ../../../main_pretrain.py \
+python3 main_pretrain.py \
     --dataset $1 \
     --backbone resnet18 \
     --data_dir ./datasets \
@@ -29,6 +29,7 @@ python3 ../../../main_pretrain.py \
     --entity unitn-mhug \
     --wandb \
     --save_checkpoint \
+    --auto_resume \
     --method barlow_twins \
     --proj_hidden_dim 2048 \
     --proj_output_dim 2048 \

bash_files/pretrain/cifar/byol.sh

@@ -1,4 +1,4 @@
-python3 ../../../main_pretrain.py \
+python3 main_pretrain.py \
     --dataset $1 \
     --backbone resnet18 \
     --data_dir ./datasets \
@@ -30,6 +30,7 @@ python3 ../../../main_pretrain.py \
     --entity unitn-mhug \
     --wandb \
     --save_checkpoint \
+    --auto_resume \
     --method byol \
     --proj_output_dim 256 \
     --proj_hidden_dim 4096 \

bash_files/pretrain/cifar/deepclusterv2.sh

@@ -1,4 +1,4 @@
-python3 ../../../main_pretrain.py \
+python3 main_pretrain.py \
     --dataset $1 \
     --backbone resnet18 \
     --data_dir ./datasets \
@@ -31,6 +31,7 @@ python3 ../../../main_pretrain.py \
     --entity unitn-mhug \
     --wandb \
     --save_checkpoint \
+    --auto_resume \
     --method deepclusterv2 \
     --proj_hidden_dim 2048 \
     --proj_output_dim 128 \

bash_files/pretrain/cifar/dino.sh

@@ -1,4 +1,4 @@
-python3 ../../main_pretrain.py \
+python3 main_pretrain.py \
     --dataset $1 \
     --backbone resnet18 \
     --data_dir ./datasets \
@@ -31,6 +31,7 @@ python3 ../../main_pretrain.py \
     --project solo-learn \
     --wandb \
     --save_checkpoint \
+    --auto_resume \
     --method dino \
     --proj_output_dim 256 \
     --proj_hidden_dim 2048 \

bash_files/pretrain/cifar/mocov2plus.sh

@@ -1,4 +1,4 @@
-python3 ../../../main_pretrain.py \
+python3 main_pretrain.py \
     --dataset $1 \
     --backbone resnet18 \
     --data_dir ./datasets \
@@ -25,6 +25,7 @@ python3 ../../../main_pretrain.py \
     --entity unitn-mhug \
     --wandb \
     --save_checkpoint \
+    --auto_resume \
     --method mocov2plus \
     --proj_hidden_dim 2048 \
     --queue_size 32768 \

bash_files/pretrain/cifar/nnbyol.sh

@@ -1,4 +1,4 @@
-python3 ../../../main_pretrain.py \
+python3 main_pretrain.py \
     --dataset $1 \
     --backbone resnet18 \
     --data_dir ./datasets \
@@ -30,6 +30,7 @@ python3 ../../../main_pretrain.py \
     --entity unitn-mhug \
     --wandb \
     --save_checkpoint \
+    --auto_resume \
     --method nnbyol \
     --proj_output_dim 256 \
     --proj_hidden_dim 4096 \

bash_files/pretrain/cifar/nnclr.sh

@@ -1,4 +1,4 @@
-python3 ../../../main_pretrain.py \
+python3 main_pretrain.py \
     --dataset $1 \
     --backbone resnet18 \
     --data_dir ./datasets \
@@ -30,6 +30,7 @@ python3 ../../../main_pretrain.py \
     --entity unitn-mhug \
     --wandb \
     --save_checkpoint \
+    --auto_resume \
     --method nnclr \
     --temperature 0.2 \
     --proj_hidden_dim 2048 \

bash_files/pretrain/cifar/nnsiam.sh

@@ -1,4 +1,4 @@
-python3 ../../../main_pretrain.py \
+python3 main_pretrain.py \
     --dataset $1 \
     --backbone resnet18 \
     --data_dir ./datasets \
@@ -26,6 +26,7 @@ python3 ../../../main_pretrain.py \
     --entity unitn-mhug \
     --wandb \
     --save_checkpoint \
+    --auto_resume \
     --method nnsiam \
     --proj_hidden_dim 2048 \
     --pred_hidden_dim 4096 \

bash_files/pretrain/cifar/ressl.sh

@@ -1,4 +1,4 @@
-python3 ../../../main_pretrain.py \
+python3 main_pretrain.py \
     --dataset $1 \
     --backbone resnet18 \
     --data_dir ./datasets \
@@ -26,6 +26,7 @@ python3 ../../../main_pretrain.py \
     --entity unitn-mhug \
     --wandb \
     --save_checkpoint \
+    --auto_resume \
     --method ressl \
     --proj_output_dim 256 \
     --proj_hidden_dim 4096 \

bash_files/pretrain/cifar/simclr.sh

@@ -1,4 +1,4 @@
-python3 ../../../main_pretrain.py \
+python3 main_pretrain.py \
     --dataset $1 \
     --backbone resnet18 \
     --data_dir ./datasets \
@@ -30,6 +30,7 @@ python3 ../../../main_pretrain.py \
     --entity unitn-mhug \
     --wandb \
     --save_checkpoint \
+    --auto_resume \
     --method simclr \
     --temperature 0.2 \
     --proj_hidden_dim 2048 \

bash_files/pretrain/cifar/simsiam.sh

@@ -1,4 +1,4 @@
-python3 ../../../main_pretrain.py \
+python3 main_pretrain.py \
     --dataset $1 \
     --backbone resnet18 \
     --data_dir ./datasets \
@@ -27,6 +27,7 @@ python3 ../../../main_pretrain.py \
     --entity unitn-mhug \
     --wandb \
     --save_checkpoint \
+    --auto_resume \
     --method simsiam \
     --proj_hidden_dim 2048 \
     --pred_hidden_dim 512 \