Training in progress, epoch 2

logs/events.out.tfevents.1711287195.73e3a81c01ef.4225.15

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e6fb25125124b9b9d9a0341a47b0387f83df93bc31ae4225a44024ee3b73ad5d
-size 5406
+oid sha256:403f1a518bcc463138e66a72fadd414921820dfc01b7fa5fbd1395f6c258df54
+size 6136

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:50615611d943de584f269e9a5fa4620b06173132a27e296d08c759db90b52dd1
+oid sha256:95c2c941e35cf59694efe03c8007ed371ccb0bdd18075010faf556d3621e9eb1
 size 17549312

run-15/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-15/checkpoint-384/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a98259bcc6c7b8ce9dc742272195ffc7502605916932ada5796b99ee363d2dc6
+oid sha256:95c2c941e35cf59694efe03c8007ed371ccb0bdd18075010faf556d3621e9eb1
 size 17549312

run-15/checkpoint-384/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8eea14a21a183c87812de3bb42f4343bd0b1a3449c4c678df6f3dba2207a2c24
+oid sha256:06eefa92114924ff70ff50185c0bce8156966dfa739607831153a16301821504
 size 35122746

run-15/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-15/checkpoint-384/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f8475ccba5300e3a468967022cfeb45e592c79a539db09f05fe6b49f028b6c33
+oid sha256:92a5ac9a84fcd94b2c073c17746994dbc8a540506908fc8e09de1d60c0414208
 size 1064

run-15/checkpoint-384/trainer_state.json

@@ -1,7 +1,7 @@
 {
-  "best_metric": 0.815068493150685,
+  "best_metric": 0.8082191780821918,
   "best_model_checkpoint": "tiny-bert-sst2-distilled/run-15/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": 1.150837779045105,
-      "learning_rate": 0.00014120530022151875,
-      "loss": 0.5007,
-      "step": 96
+      "grad_norm": 3.1954071521759033,
+      "learning_rate": 4.7129202867976326e-05,
+      "loss": 0.5731,
+      "step": 192
     },
     {
       "epoch": 1.0,
-      "eval_accuracy": 0.7681017612524462,
-      "eval_f1": 0.7889581478183438,
-      "eval_loss": 0.4356862008571625,
-      "eval_precision": 0.7238562091503268,
-      "eval_recall": 0.8669275929549902,
-      "eval_runtime": 28.2732,
-      "eval_samples_per_second": 36.147,
-      "eval_steps_per_second": 1.132,
-      "step": 96
-    },
-    {
-      "epoch": 2.0,
-      "grad_norm": 2.0905497074127197,
-      "learning_rate": 9.413686681434584e-05,
-      "loss": 0.4302,
+      "eval_accuracy": 0.7896281800391389,
+      "eval_f1": 0.8088888888888889,
+      "eval_loss": 0.47491252422332764,
+      "eval_mcc": 0.5913947215491926,
+      "eval_precision": 0.741042345276873,
+      "eval_recall": 0.8904109589041096,
+      "eval_runtime": 66.6851,
+      "eval_samples_per_second": 15.326,
+      "eval_steps_per_second": 0.48,
       "step": 192
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.7984344422700587,
-      "eval_f1": 0.81993006993007,
-      "eval_loss": 0.4125403165817261,
-      "eval_precision": 0.740916271721959,
-      "eval_recall": 0.9178082191780822,
-      "eval_runtime": 29.852,
-      "eval_samples_per_second": 34.236,
-      "eval_steps_per_second": 1.072,
-      "step": 192
-    },
-    {
-      "epoch": 3.0,
-      "grad_norm": 2.322948455810547,
-      "learning_rate": 4.706843340717292e-05,
-      "loss": 0.4123,
-      "step": 288
-    },
-    {
-      "epoch": 3.0,
-      "eval_accuracy": 0.812133072407045,
-      "eval_f1": 0.8303886925795053,
-      "eval_loss": 0.40187525749206543,
-      "eval_precision": 0.7568438003220612,
-      "eval_recall": 0.9197651663405088,
-      "eval_runtime": 28.0204,
-      "eval_samples_per_second": 36.473,
-      "eval_steps_per_second": 1.142,
-      "step": 288
-    },
-    {
-      "epoch": 4.0,
-      "grad_norm": 1.9369499683380127,
-      "learning_rate": 0.0,
-      "loss": 0.4038,
+      "grad_norm": 7.988830089569092,
+      "learning_rate": 2.3564601433988163e-05,
+      "loss": 0.4767,
       "step": 384
     },
     {
-      "epoch": 4.0,
-      "eval_accuracy": 0.815068493150685,
-      "eval_f1": 0.8307967770814683,
-      "eval_loss": 0.39612340927124023,
-      "eval_precision": 0.7656765676567657,
-      "eval_recall": 0.9080234833659491,
-      "eval_runtime": 28.9135,
-      "eval_samples_per_second": 35.347,
-      "eval_steps_per_second": 1.107,
+      "epoch": 2.0,
+      "eval_accuracy": 0.8082191780821918,
+      "eval_f1": 0.8208409506398537,
+      "eval_loss": 0.4437917172908783,
+      "eval_mcc": 0.6226500342007837,
+      "eval_precision": 0.7701543739279588,
+      "eval_recall": 0.8786692759295499,
+      "eval_runtime": 66.9186,
+      "eval_samples_per_second": 15.272,
+      "eval_steps_per_second": 0.478,
       "step": 384
     }
   ],
   "logging_steps": 500,
-  "max_steps": 384,
+  "max_steps": 576,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 4,
+  "num_train_epochs": 3,
   "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": 0.7235742113106859,
-    "learning_rate": 0.00018827373362869168,
-    "num_train_epochs": 4,
-    "temperature": 25
+    "alpha": 0.8612888560371268,
+    "learning_rate": 7.069380430196449e-05,
+    "num_train_epochs": 3,
+    "per_device_train_batch_size": 16,
+    "temperature": 23
   }
 }

run-15/checkpoint-384/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b4455ec233a06250ca78d744f669f3caf07ff1e989cb0faa606b8567c0ee6a96
-size 4920
+oid sha256:d957ce0a97207c7175ba451db3f66fa371ed707404ff50f15e7363524a12dff1
+size 4984