Training in progress, step 100

all_results.json

@@ -1,20 +1,20 @@
 {
     "epoch": 20.0,
-    "eval_accuracy": 85.71428571428571,
-    "eval_average_metrics": 73.0,
-    "eval_f1_multiclass": 60.285714285714285,
-    "eval_loss": 0.18336069583892822,
-    "eval_runtime": 0.35,
-    "eval_samples_per_second": 80.007,
-    "test_accuracy": 78.57142857142857,
-    "test_average_metrics": 66.48351648351648,
-    "test_f1_multiclass": 54.395604395604394,
-    "test_loss": 0.2799363434314728,
-    "test_runtime": 0.3233,
-    "test_samples_per_second": 86.617,
-    "train_loss": 0.5726573467254639,
-    "train_runtime": 63.2805,
+    "eval_accuracy": 100.0,
+    "eval_average_metrics": 100.0,
+    "eval_f1_multiclass": 100.0,
+    "eval_loss": 0.007105956319719553,
+    "eval_runtime": 0.2511,
+    "eval_samples_per_second": 111.49,
+    "test_accuracy": 82.14285714285714,
+    "test_average_metrics": 76.00732600732601,
+    "test_f1_multiclass": 69.87179487179488,
+    "test_loss": 0.1506887525320053,
+    "test_runtime": 0.2093,
+    "test_samples_per_second": 133.792,
+    "train_loss": 0.19091113805770873,
+    "train_runtime": 58.4291,
     "train_samples": 250,
-    "train_samples_per_second": 79.013,
-    "train_steps_per_second": 2.528
+    "train_samples_per_second": 85.574,
+    "train_steps_per_second": 2.738
 }

eval_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 20.0,
-    "eval_accuracy": 85.71428571428571,
-    "eval_average_metrics": 73.0,
-    "eval_f1_multiclass": 60.285714285714285,
-    "eval_loss": 0.18336069583892822,
-    "eval_runtime": 0.35,
-    "eval_samples_per_second": 80.007
+    "eval_accuracy": 100.0,
+    "eval_average_metrics": 100.0,
+    "eval_f1_multiclass": 100.0,
+    "eval_loss": 0.007105956319719553,
+    "eval_runtime": 0.2511,
+    "eval_samples_per_second": 111.49
 }

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0573c78cbc10c56dd04e51ad3381b15ae1bb52835f7a90be1fd2daa19a59947e
-size 7551621
+oid sha256:ea14cf46ad1619ee1ef08e0ad3fd81cec57e290129f14d4585b6a468f1f0bd36
+size 2631685

runs/Feb01_02-59-20_node2/events.out.tfevents.1643655675.node2

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:96fde66482d741bcfe90036e27c50a26b7a2302a2d15fbe9abfb70f670a40b1a
-size 4324
+oid sha256:b63714958fb6afa0b43c78c085d6bc30bc7f8d6ce3c2cef0174e13548e430044
+size 5050

runs/Feb01_02-59-20_node2/events.out.tfevents.1643655734.node2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d412c355b5d78bcd746aa140c413f9f9c3408d1f12f4ec20b3c48319366a6036
+size 798

runs/Feb02_16-19-40_node1/1643790011.7594028/events.out.tfevents.1643790011.node1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d2eec56d84bc60fca26bf5ca6fc935bf39a6628c8eb25fb7889c3db7cc9bb069
+size 5035

runs/Feb02_16-19-40_node1/events.out.tfevents.1643790011.node1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1d9109a7c681983ea1d1823d23c916baf9d2d0778ae35e13ff774d9516b83166
+size 4324

test_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 20.0,
-    "test_accuracy": 78.57142857142857,
-    "test_average_metrics": 66.48351648351648,
-    "test_f1_multiclass": 54.395604395604394,
-    "test_loss": 0.2799363434314728,
-    "test_runtime": 0.3233,
-    "test_samples_per_second": 86.617
+    "test_accuracy": 82.14285714285714,
+    "test_average_metrics": 76.00732600732601,
+    "test_f1_multiclass": 69.87179487179488,
+    "test_loss": 0.1506887525320053,
+    "test_runtime": 0.2093,
+    "test_samples_per_second": 133.792
 }

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 20.0,
-    "train_loss": 0.5726573467254639,
-    "train_runtime": 63.2805,
+    "train_loss": 0.19091113805770873,
+    "train_runtime": 58.4291,
     "train_samples": 250,
-    "train_samples_per_second": 79.013,
-    "train_steps_per_second": 2.528
+    "train_samples_per_second": 85.574,
+    "train_steps_per_second": 2.738
 }

trainer_state.json

@@ -1,25 +1,35 @@
 {
-  "best_metric": null,
-  "best_model_checkpoint": null,
+  "best_metric": 100.0,
+  "best_model_checkpoint": "outputs/bitfit/t5-base/superglue-cb/checkpoint-100",
   "epoch": 20.0,
   "global_step": 160,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
+    {
+      "epoch": 12.5,
+      "eval_accuracy": 100.0,
+      "eval_average_metrics": 100.0,
+      "eval_f1_multiclass": 100.0,
+      "eval_loss": 0.007105956319719553,
+      "eval_runtime": 0.2475,
+      "eval_samples_per_second": 113.147,
+      "step": 100
+    },
     {
       "epoch": 20.0,
       "step": 160,
-      "total_flos": 1311789053515248.0,
-      "train_loss": 0.5726573467254639,
-      "train_runtime": 63.2805,
-      "train_samples_per_second": 79.013,
-      "train_steps_per_second": 2.528
+      "total_flos": 1322112125892096.0,
+      "train_loss": 0.19091113805770873,
+      "train_runtime": 58.4291,
+      "train_samples_per_second": 85.574,
+      "train_steps_per_second": 2.738
     }
   ],
   "max_steps": 160,
   "num_train_epochs": 20,
-  "total_flos": 1311789053515248.0,
+  "total_flos": 1322112125892096.0,
   "trial_name": null,
   "trial_params": null
 }

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:77ed515ec8421abcec58cda3fc06acdc764638bd9430acc9cf100024288e02aa
+oid sha256:4011332cc5f8c28977c05caffbccdc5053455cedcf074b3a5f9c7bae07772739
 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": "superglue-cb", "eval_steps": 100, "evaluation_strategy": "steps", "greater_is_better": true, "learning_rate": 0.0003, "load_best_model_at_end": true, "max_source_length": 256, "metric_for_best_model": "average_metrics", "model_name_or_path": "../../../../plm_cache/t5-base", "num_train_epochs": 20, "output_dir": "outputs/bitfit/t5-base/superglue-cb", "overwrite_output_dir": true, "per_device_eval_batch_size": 32, "per_device_train_batch_size": 32, "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": "superglue-cb", "test_dataset_config_name": ["en"], "test_dataset_name": "superglue-cb", "tokenizer_name": "../../../../plm_cache/t5-base", "unfrozen_modules": ["deltas", "layer_norm", "final_layer_norm"], "warmup_steps": 0}
+{"dataset_config_name": ["en"], "delta_type": "lora", "do_eval": true, "do_test": true, "do_train": true, "eval_dataset_config_name": ["en"], "eval_dataset_name": "superglue-cb", "eval_steps": 100, "evaluation_strategy": "steps", "greater_is_better": true, "learning_rate": 0.0003, "load_best_model_at_end": true, "lora_r": 8, "max_source_length": 256, "metric_for_best_model": "average_metrics", "model_name_or_path": "../../../../plm_cache/t5-base", "num_train_epochs": 20, "output_dir": "outputs/bitfit/t5-base/superglue-cb", "overwrite_output_dir": true, "per_device_eval_batch_size": 32, "per_device_train_batch_size": 32, "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": "superglue-cb", "test_dataset_config_name": ["en"], "test_dataset_name": "superglue-cb", "tokenizer_name": "../../../../plm_cache/t5-base", "unfrozen_modules": ["deltas", "layer_norm", "final_layer_norm"], "warmup_steps": 0}