Training in progress, step 200

all_results.json

@@ -3,18 +3,18 @@
     "eval_accuracy": 75.36231884057972,
     "eval_average_metrics": 75.36231884057972,
     "eval_loss": 0.2184353768825531,
-    "eval_runtime": 0.505,
+    "eval_runtime": 0.5022,
     "eval_samples": 277,
-    "eval_samples_per_second": 273.265,
+    "eval_samples_per_second": 274.794,
     "eval_steps_per_second": 6.498,
     "test_accuracy": 69.06474820143885,
     "test_average_metrics": 69.06474820143885,
     "test_loss": 0.26203805208206177,
-    "test_runtime": 0.5052,
-    "test_samples_per_second": 275.13,
+    "test_runtime": 0.504,
+    "test_samples_per_second": 275.803,
     "train_loss": 1.2798988037109376,
-    "train_runtime": 273.704,
+    "train_runtime": 213.577,
     "train_samples": 2490,
-    "train_samples_per_second": 181.948,
-    "train_steps_per_second": 1.827
+    "train_samples_per_second": 233.171,
+    "train_steps_per_second": 2.341
 }

eval_results.json

@@ -3,6 +3,6 @@
     "eval_accuracy": 75.36231884057972,
     "eval_average_metrics": 75.36231884057972,
     "eval_loss": 0.2184353768825531,
-    "eval_runtime": 0.505,
-    "eval_samples_per_second": 273.265
+    "eval_runtime": 0.5022,
+    "eval_samples_per_second": 274.794
 }

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5a181fa83a25220bf33cf44d9adf3957ff9e907d54405e714d32d8f0226f5855
-size 1084131
+oid sha256:8f4f85cf5857e083786598351fbba0df4be7de00d0f9037b03b101501f1f34f1
+size 7551621

runs/Feb01_00-07-44_node1/events.out.tfevents.1643645355.node1

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f33c18073972edb7b97ec256ece9b5326854ef806609e135110a24cfd7c8c3c8
-size 4623
+oid sha256:f2328c976844d6b59391ba04300fce90d4b144155a1f17c7133475d446bedc32
+size 5456

runs/Feb01_00-07-44_node1/events.out.tfevents.1643645569.node1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d557d9f6bf3b17ee5b080861b0fdb00e555379951f749baece459ee49edcd39b
+size 684

runs/Feb01_00-37-34_node2/1643647192.4988313/events.out.tfevents.1643647192.node2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c133aa36617b736015b2041c86207c3b559d0797e2d7a06d174ae8c1cdc5228c
+size 5008

runs/Feb01_00-37-34_node2/events.out.tfevents.1643647192.node2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9cc2de13871b4bd3f1268b8f47f7225916a91f97b8f7bf90fb905d4ec01f1ccb
+size 4301

test_results.json

@@ -3,6 +3,6 @@
     "test_accuracy": 69.06474820143885,
     "test_average_metrics": 69.06474820143885,
     "test_loss": 0.26203805208206177,
-    "test_runtime": 0.5052,
-    "test_samples_per_second": 275.13
+    "test_runtime": 0.504,
+    "test_samples_per_second": 275.803
 }

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 20.0,
     "train_loss": 1.2798988037109376,
-    "train_runtime": 273.704,
+    "train_runtime": 213.577,
     "train_samples": 2490,
-    "train_samples_per_second": 181.948,
-    "train_steps_per_second": 1.827
+    "train_samples_per_second": 233.171,
+    "train_steps_per_second": 2.341
 }

trainer_state.json

@@ -12,8 +12,8 @@
       "eval_accuracy": 54.347826086956516,
       "eval_average_metrics": 54.347826086956516,
       "eval_loss": 0.3104463517665863,
-      "eval_runtime": 0.6899,
-      "eval_samples_per_second": 200.036,
+      "eval_runtime": 0.512,
+      "eval_samples_per_second": 269.506,
       "step": 200
     },
     {
@@ -21,8 +21,8 @@
       "eval_accuracy": 75.36231884057972,
       "eval_average_metrics": 75.36231884057972,
       "eval_loss": 0.2184353768825531,
-      "eval_runtime": 0.6861,
-      "eval_samples_per_second": 201.146,
+      "eval_runtime": 0.5093,
+      "eval_samples_per_second": 270.959,
       "step": 400
     },
     {
@@ -36,9 +36,9 @@
       "step": 500,
       "total_flos": 7590919206297600.0,
       "train_loss": 1.2798988037109376,
-      "train_runtime": 273.704,
-      "train_samples_per_second": 181.948,
-      "train_steps_per_second": 1.827
+      "train_runtime": 213.577,
+      "train_samples_per_second": 233.171,
+      "train_steps_per_second": 2.341
     }
   ],
   "max_steps": 500,

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:cb2732fc0f205ecb2ff6a4dc01793c5bad05e44f957c3092d344cc75969c345d
+oid sha256:56413463a8bc5502959735438cbd17111e8f0d9e4f25bc1cad3aa8a0e688b298
 size 3183

training_config.json

@@ -1 +1 @@
-{"dataset_config_name": ["en"], "delta_type": "bitfit", "do_eval": true, "do_test": true, "do_train": true, "eval_dataset_config_name": ["en"], "eval_dataset_name": "rte", "eval_steps": 200, "evaluation_strategy": "steps", "greater_is_better": true, "learning_rate": 0.0003, "load_best_model_at_end": true, "max_source_length": 128, "metric_for_best_model": "average_metrics", "model_name_or_path": "../../../../plm_cache/t5-base", "num_train_epochs": 20, "output_dir": "outputs/bitfit/t5-base/rte", "overwrite_output_dir": true, "per_device_eval_batch_size": 100, "per_device_train_batch_size": 100, "predict_with_generate": true, "push_to_hub": true, "save_steps": 200, "save_strategy": "steps", "save_total_limit": 1, "seed": 42, "split_validation_test": true, "task_name": "rte", "test_dataset_config_name": ["en"], "test_dataset_name": "rte", "tokenizer_name": "../../../../plm_cache/t5-base", "warmup_steps": 500}
+{"bottleneck_dim": 24, "dataset_config_name": ["en"], "delta_type": "adapter", "do_eval": true, "do_test": true, "do_train": true, "eval_dataset_config_name": ["en"], "eval_dataset_name": "rte", "eval_steps": 200, "evaluation_strategy": "steps", "greater_is_better": true, "learning_rate": 0.0003, "load_best_model_at_end": true, "max_source_length": 128, "metric_for_best_model": "average_metrics", "model_name_or_path": "../../../../plm_cache/t5-base", "num_train_epochs": 20, "output_dir": "outputs/bitfit/t5-base/rte", "overwrite_output_dir": true, "per_device_eval_batch_size": 100, "per_device_train_batch_size": 100, "predict_with_generate": true, "push_to_hub": true, "save_steps": 200, "save_strategy": "steps", "save_total_limit": 1, "seed": 42, "split_validation_test": true, "task_name": "rte", "test_dataset_config_name": ["en"], "test_dataset_name": "rte", "tokenizer_name": "../../../../plm_cache/t5-base", "unfrozen_modules": ["deltas", "layer_norm", "final_layer_norm"], "warmup_steps": 500}