Training in progress, epoch 1

command.txt

@@ -1,26 +1,25 @@
-python run_image_classification.py \
-    --train_dir /trunk/shared/eebo_data/images_cropped/ \
-    --validation_dir /trunk/shared/eebo_data/images_cropped/ \
-    --output_dir ./microsoft_beit-base-patch16-224-pt22k-ft22k/ \
-    --remove_unused_columns False \
-    --label_column_name labels \
-    --do_train \
-    --do_eval \
-    --push_to_hub \
-    --push_to_hub_model_id donut-base-beans \
-    --learning_rate 3e-5 \
-    --num_train_epochs 10 \
-    --per_device_train_batch_size 64 \
-    --per_device_eval_batch_size 64 \
-    --logging_strategy steps \
-    --logging_steps 10 \
-    --eval_strategy epoch \
-    --save_strategy epoch \
-    --load_best_model_at_end True \
-    --save_total_limit 5 \
-    --seed 1337 \
-    --ignore_mismatched_sizes \
-    --image_column_name image_url \
-    --label_column_name label \
-    --model_name_or_path microsoft/beit-base-patch16-224-pt22k-ft22k \
-    --resume_from_checkpoint /trunk2/yufei/summer24/transformers/examples/pytorch/image-classification/microsoft_beit-base-patch16-224-pt22k-ft22k/checkpoint-33670
+python run_image_classification_val_test.py \
+--train_dir /trunk/shared/eebo_data/images_cropped/ \
+--validation_dir /trunk/shared/eebo_data/images_cropped/ \
+--output_dir ./test_microsoft_dit/ \
+--remove_unused_columns False \
+--label_column_name labels \
+--do_train \
+--do_eval \
+--push_to_hub \
+--push_to_hub_model_id donut-base-beans \
+--learning_rate 3e-5 \
+--num_train_epochs 5 \
+--per_device_train_batch_size 32 \
+--per_device_eval_batch_size 32 \
+--logging_strategy steps \
+--logging_steps 10 \
+--eval_strategy epoch \
+--save_strategy epoch \
+--load_best_model_at_end True \
+--save_total_limit 5 \
+--seed 1337 \
+--ignore_mismatched_sizes \
+--image_column_name image_url \
+--label_column_name label \
+--model_name_or_path microsoft/dit-base-finetuned-rvlcdip

config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "microsoft/beit-base-patch16-224-pt22k-ft22k",
+  "_name_or_path": "microsoft/dit-base-finetuned-rvlcdip",
   "add_fpn": false,
   "architectures": [
     "BeitForImageClassification"
@@ -30,10 +30,16 @@
   "num_channels": 3,
   "num_hidden_layers": 12,
   "out_features": [
-    "stage12"
+    "stage3",
+    "stage5",
+    "stage7",
+    "stage11"
   ],
   "out_indices": [
-    12
+    3,
+    5,
+    7,
+    11
   ],
   "patch_size": 16,
   "pool_scales": [
@@ -62,11 +68,11 @@
   ],
   "torch_dtype": "float32",
   "transformers_version": "4.43.0.dev0",
-  "use_absolute_position_embeddings": false,
+  "use_absolute_position_embeddings": true,
   "use_auxiliary_head": true,
   "use_mask_token": false,
   "use_mean_pooling": true,
-  "use_relative_position_bias": true,
+  "use_relative_position_bias": false,
   "use_shared_relative_position_bias": false,
   "vocab_size": 8192
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f0ba8bd13150b149523e0607dde1876a3f30f4d9a2c833f6d2b9d41c7564539f
-size 343077252
+oid sha256:b5ab9b8b1e14e6152ce977b120fdb25da84fd41b3cc0f9e81bfe849771b48f31
+size 343258940

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:310b373eea79944bc0b943e2e1bbed1ba76fe58dc0da1f1998cca06c428e175f
-size 5432
+oid sha256:28e008fa9d7fba0a7a650bf2907bd71f3c36c3f0d4b0f597443c926a867bb792
+size 5240