Training in progress, epoch 2

logs/events.out.tfevents.1711273902.73e3a81c01ef.4225.3

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:52639697b4c0801d45c96b78b024abec657c666239e4bd37557653bccf89b094
-size 5407
+oid sha256:74e26e192d8c7f213d390797c7892456060b9a21f5cea17a996123c87d0b2be8
+size 6137

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e9b577371af93984a46904b9481062b830719a7ec03c064beb972380f48b9bc0
+oid sha256:70a388d72e55b7953b67cdfa00ddc695e26206a74815a6d36483067f798bda2f
 size 17549312

run-3/checkpoint-384/config.json

@@ -27,7 +27,7 @@
   "position_embedding_type": "absolute",
   "problem_type": "single_label_classification",
   "torch_dtype": "float32",
-  "transformers_version": "4.38.2",
+  "transformers_version": "4.39.1",
   "type_vocab_size": 2,
   "use_cache": true,
   "vocab_size": 30522

run-3/checkpoint-384/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bc097722feb26fcf6693b9dbe97ea3a76b6802da9cd21cb5bf38424a0f3cfbad
+oid sha256:70a388d72e55b7953b67cdfa00ddc695e26206a74815a6d36483067f798bda2f
 size 17549312

run-3/checkpoint-384/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:12d4c8e2b1db72df2a15dfb1bed5f04a8214b8e53b3366ebb528306bcc4f8073
+oid sha256:2f464f717f404629956b434b8f10ba552dc7072676fd35983f5993b56e5a75d8
 size 35122746

run-3/checkpoint-384/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:45b1fc07ed8b121716021ec87c686f1cd9c3b89c82ea08a6a0792d47a39077c9
+oid sha256:d2558f72cda987826e5e7caf54cc6282fe335ebecbddfed7bb83f0184d1f54cc
 size 14054

run-3/checkpoint-384/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c9ff4fe586e7eb333ddca8c1956ec7e088a7d846492f389a13a5f39b9e31e62d
+oid sha256:2ad75831118aa992f2e70ece558a78a8328cfe8f2d54db597c8aa3051d18df32
 size 1064

run-3/checkpoint-384/trainer_state.json

@@ -1,7 +1,7 @@
 {
-  "best_metric": 0.5009784735812133,
+  "best_metric": 0.5459882583170255,
   "best_model_checkpoint": "tiny-bert-sst2-distilled/run-3/checkpoint-384",
-  "epoch": 4.0,
+  "epoch": 2.0,
   "eval_steps": 500,
   "global_step": 384,
   "is_hyper_param_search": true,
@@ -10,93 +10,58 @@
   "log_history": [
     {
       "epoch": 1.0,
-      "grad_norm": 2.101290702819824,
-      "learning_rate": 1.1962059637335337e-06,
-      "loss": 0.5173,
-      "step": 96
+      "grad_norm": 0.7568773031234741,
+      "learning_rate": 0.00039540069434609415,
+      "loss": 0.3274,
+      "step": 192
     },
     {
       "epoch": 1.0,
-      "eval_accuracy": 0.5,
-      "eval_f1": 0.6666666666666666,
-      "eval_loss": 0.5046124458312988,
-      "eval_precision": 0.5,
-      "eval_recall": 1.0,
-      "eval_runtime": 28.251,
-      "eval_samples_per_second": 36.176,
-      "eval_steps_per_second": 1.133,
-      "step": 96
-    },
-    {
-      "epoch": 2.0,
-      "grad_norm": 0.6350829005241394,
-      "learning_rate": 1.0466802182668422e-06,
-      "loss": 0.4989,
+      "eval_accuracy": 0.5430528375733855,
+      "eval_f1": 0.20442930153321975,
+      "eval_loss": 0.30545827746391296,
+      "eval_mcc": 0.1640968755879385,
+      "eval_precision": 0.7894736842105263,
+      "eval_recall": 0.11741682974559686,
+      "eval_runtime": 66.7048,
+      "eval_samples_per_second": 15.321,
+      "eval_steps_per_second": 0.48,
       "step": 192
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.4843444227005871,
-      "eval_f1": 0.6216798277099785,
-      "eval_loss": 0.48798972368240356,
-      "eval_precision": 0.4909297052154195,
-      "eval_recall": 0.8473581213307241,
-      "eval_runtime": 28.492,
-      "eval_samples_per_second": 35.87,
-      "eval_steps_per_second": 1.123,
-      "step": 192
-    },
-    {
-      "epoch": 3.0,
-      "grad_norm": 1.8015046119689941,
-      "learning_rate": 8.971544728001503e-07,
-      "loss": 0.4869,
-      "step": 288
-    },
-    {
-      "epoch": 3.0,
-      "eval_accuracy": 0.48238747553816047,
-      "eval_f1": 0.09572649572649572,
-      "eval_loss": 0.4779175817966461,
-      "eval_precision": 0.3783783783783784,
-      "eval_recall": 0.0547945205479452,
-      "eval_runtime": 28.2537,
-      "eval_samples_per_second": 36.172,
-      "eval_steps_per_second": 1.133,
-      "step": 288
-    },
-    {
-      "epoch": 4.0,
-      "grad_norm": 2.0801610946655273,
-      "learning_rate": 7.476287273334586e-07,
-      "loss": 0.4795,
+      "grad_norm": 1.822717547416687,
+      "learning_rate": 0.0003295005786217451,
+      "loss": 0.3076,
       "step": 384
     },
     {
-      "epoch": 4.0,
-      "eval_accuracy": 0.5009784735812133,
-      "eval_f1": 0.00390625,
-      "eval_loss": 0.47173693776130676,
-      "eval_precision": 1.0,
-      "eval_recall": 0.0019569471624266144,
-      "eval_runtime": 28.2359,
-      "eval_samples_per_second": 36.195,
-      "eval_steps_per_second": 1.133,
+      "epoch": 2.0,
+      "eval_accuracy": 0.5459882583170255,
+      "eval_f1": 0.21355932203389827,
+      "eval_loss": 0.2995118200778961,
+      "eval_mcc": 0.17219814389174506,
+      "eval_precision": 0.7974683544303798,
+      "eval_recall": 0.1232876712328767,
+      "eval_runtime": 66.8,
+      "eval_samples_per_second": 15.299,
+      "eval_steps_per_second": 0.479,
       "step": 384
     }
   ],
   "logging_steps": 500,
-  "max_steps": 864,
+  "max_steps": 1344,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 9,
+  "num_train_epochs": 7,
   "save_steps": 500,
-  "total_flos": 942780789120.0,
-  "train_batch_size": 32,
+  "total_flos": 471390394560.0,
+  "train_batch_size": 16,
   "trial_name": null,
   "trial_params": {
-    "alpha": 0.586704262013931,
-    "learning_rate": 1.3457317092002255e-06,
-    "num_train_epochs": 9,
-    "temperature": 21
+    "alpha": 0.4369516344275761,
+    "learning_rate": 0.00046130081007044317,
+    "num_train_epochs": 7,
+    "per_device_train_batch_size": 16,
+    "temperature": 15
   }
 }

run-3/checkpoint-384/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b765cab6920f955460cc4f812305ba7f085db0f4dbb906f08e22ff912fece657
-size 4920
+oid sha256:1984bbced85533f487884fbafe7fdded01dc8a3f7df5f2a96fbc5883290bb68c
+size 4984