Training in progress, step 100

all_results.json

@@ -1,20 +1,20 @@
 {
     "epoch": 20.0,
-    "eval_accuracy": 75.36231884057972,
-    "eval_average_metrics": 75.36231884057972,
-    "eval_loss": 0.2184353768825531,
-    "eval_runtime": 0.5022,
+    "eval_accuracy": 81.15942028985508,
+    "eval_average_metrics": 81.15942028985508,
+    "eval_loss": 0.18789426982402802,
+    "eval_runtime": 0.4156,
     "eval_samples": 277,
-    "eval_samples_per_second": 274.794,
+    "eval_samples_per_second": 332.064,
     "eval_steps_per_second": 6.498,
-    "test_accuracy": 69.06474820143885,
-    "test_average_metrics": 69.06474820143885,
-    "test_loss": 0.26203805208206177,
-    "test_runtime": 0.504,
-    "test_samples_per_second": 275.803,
-    "train_loss": 1.2798988037109376,
-    "train_runtime": 213.577,
+    "test_accuracy": 69.7841726618705,
+    "test_average_metrics": 69.7841726618705,
+    "test_loss": 0.23357094824314117,
+    "test_runtime": 0.5204,
+    "test_samples_per_second": 267.078,
+    "train_loss": 0.7619044189453125,
+    "train_runtime": 190.0328,
     "train_samples": 2490,
-    "train_samples_per_second": 233.171,
-    "train_steps_per_second": 2.341
+    "train_samples_per_second": 262.06,
+    "train_steps_per_second": 2.631
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 20.0,
-    "eval_accuracy": 75.36231884057972,
-    "eval_average_metrics": 75.36231884057972,
-    "eval_loss": 0.2184353768825531,
-    "eval_runtime": 0.5022,
-    "eval_samples_per_second": 274.794
+    "eval_accuracy": 81.15942028985508,
+    "eval_average_metrics": 81.15942028985508,
+    "eval_loss": 0.18789426982402802,
+    "eval_runtime": 0.4156,
+    "eval_samples_per_second": 332.064
 }

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0998262c867de39b60696374da2d82164d84613e829b153bb4b586337380a849
+oid sha256:e7f427431b2843b9e85a385dab5e4d575b37a04309d009def964f8f88bbcbb7f
 size 7551621

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

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3ce183e9a1573a06c4ae0d589f6c022a20d5590f822d6bb3879014e68d936bfb
-size 4623
+oid sha256:d7b6d00bcdfe6cfe20060c7c0bbcd8063fbe525d341fb9555416ecc90a6187f8
+size 5456

runs/Feb01_00-37-34_node2/events.out.tfevents.1643647383.node2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f4025b8b8795a95c5caad4f0ec6024cbf2a5bf5b7df216b90ba3eb8f60fe4206
+size 684

runs/Feb01_00-55-50_node2/1643648239.5129175/events.out.tfevents.1643648239.node2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:07d2b62ef04f521dfc1008957e257f0ad2d4a58e91044bd0b4d6dd675211e501
+size 5008

runs/Feb01_00-55-50_node2/events.out.tfevents.1643648239.node2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8a509ece67e5e28fcb8b79847fd91f9ae37465d0c3c3ddeb9883288332b46d9a
+size 4301

test_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 20.0,
-    "test_accuracy": 69.06474820143885,
-    "test_average_metrics": 69.06474820143885,
-    "test_loss": 0.26203805208206177,
-    "test_runtime": 0.504,
-    "test_samples_per_second": 275.803
+    "test_accuracy": 69.7841726618705,
+    "test_average_metrics": 69.7841726618705,
+    "test_loss": 0.23357094824314117,
+    "test_runtime": 0.5204,
+    "test_samples_per_second": 267.078
 }

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 20.0,
-    "train_loss": 1.2798988037109376,
-    "train_runtime": 213.577,
+    "train_loss": 0.7619044189453125,
+    "train_runtime": 190.0328,
     "train_samples": 2490,
-    "train_samples_per_second": 233.171,
-    "train_steps_per_second": 2.341
+    "train_samples_per_second": 262.06,
+    "train_steps_per_second": 2.631
 }

trainer_state.json

@@ -1,6 +1,6 @@
 {
-  "best_metric": 75.36231884057972,
-  "best_model_checkpoint": "outputs/bitfit/t5-base/rte/checkpoint-400",
+  "best_metric": 81.15942028985508,
+  "best_model_checkpoint": "outputs/bitfit/t5-base/rte/checkpoint-200",
   "epoch": 20.0,
   "global_step": 500,
   "is_hyper_param_search": false,
@@ -9,41 +9,41 @@
   "log_history": [
     {
       "epoch": 8.0,
-      "eval_accuracy": 54.347826086956516,
-      "eval_average_metrics": 54.347826086956516,
-      "eval_loss": 0.3104463517665863,
-      "eval_runtime": 0.512,
-      "eval_samples_per_second": 269.506,
+      "eval_accuracy": 81.15942028985508,
+      "eval_average_metrics": 81.15942028985508,
+      "eval_loss": 0.18789426982402802,
+      "eval_runtime": 0.5472,
+      "eval_samples_per_second": 252.185,
       "step": 200
     },
     {
       "epoch": 16.0,
-      "eval_accuracy": 75.36231884057972,
-      "eval_average_metrics": 75.36231884057972,
-      "eval_loss": 0.2184353768825531,
-      "eval_runtime": 0.5093,
-      "eval_samples_per_second": 270.959,
+      "eval_accuracy": 78.26086956521739,
+      "eval_average_metrics": 78.26086956521739,
+      "eval_loss": 0.24881036579608917,
+      "eval_runtime": 0.3946,
+      "eval_samples_per_second": 349.683,
       "step": 400
     },
     {
       "epoch": 20.0,
       "learning_rate": 0.0,
-      "loss": 1.2799,
+      "loss": 0.7619,
       "step": 500
     },
     {
       "epoch": 20.0,
       "step": 500,
-      "total_flos": 7590919206297600.0,
-      "train_loss": 1.2798988037109376,
-      "train_runtime": 213.577,
-      "train_samples_per_second": 233.171,
-      "train_steps_per_second": 2.341
+      "total_flos": 7650655646515200.0,
+      "train_loss": 0.7619044189453125,
+      "train_runtime": 190.0328,
+      "train_samples_per_second": 262.06,
+      "train_steps_per_second": 2.631
     }
   ],
   "max_steps": 500,
   "num_train_epochs": 20,
-  "total_flos": 7590919206297600.0,
+  "total_flos": 7650655646515200.0,
   "trial_name": null,
   "trial_params": null
 }

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:56413463a8bc5502959735438cbd17111e8f0d9e4f25bc1cad3aa8a0e688b298
+oid sha256:e5cd1b6dc14d58aac88259d963a6c2496413cfecbb9ff19605486a66162852e9
 size 3183

training_config.json

@@ -1 +1 @@
-{"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}
+{"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": 100, "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": 100, "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}