Training in progress, step 4000, checkpoint

last-checkpoint/config.json

@@ -19,7 +19,7 @@
   "num_hidden_layers": 12,
   "pad_token_id": 1,
   "position_embedding_type": "absolute",
-  "transformers_version": "4.57.1",
+  "transformers_version": "4.57.3",
   "type_vocab_size": 1,
   "use_cache": true,
   "vocab_size": 50265

last-checkpoint/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c24cf57beff4ec0159701d6664a39b941c462c3470600d0acfe1d10ec6593643
+oid sha256:023c80d2320b0d4178dc5c59c3265b0a63fed5e07a3b985d349aa28e5278b6fb
 size 498612824

last-checkpoint/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:eefbc16dabf619a659109e3c6599c009df54b378af21a01a970925b45b28b803
+oid sha256:c6e92ab7b77afc3bf58700ce341f06081ea7d6857dacf39ae87546b2aa24b776
 size 997345931

last-checkpoint/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:abf2b0fc64c4be716a75868bdb69c6ff3620ba2a21085b2c861288da2529d5e7
+oid sha256:dd8861c76c95a138fb74b2279a575c494edce4804040a54509842ed6fd8b90fe
 size 14645

last-checkpoint/scaler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5ac1c46a2776d12775d23d0f587efc112188137ce2140da35bc15d301c9f620e
+oid sha256:c0033c7745b46bdca3ecab5787678834ca68f7f7e1288869dceeb38812abc253
 size 1383

last-checkpoint/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4d3f655ea9c5a4a9f86c577483cce348f86f4499f93b7851c6926bdee4b92c1f
+oid sha256:012264319a2f0e83b703857fe0c90509f0ba126782a5ba9dccbac93f5b664216
 size 1465

last-checkpoint/trainer_state.json

@@ -1,86 +1,109 @@
 {
-  "best_global_step": 1000,
-  "best_metric": 0.08424179255962372,
-  "best_model_checkpoint": "./training_output/checkpoint-1000",
-  "epoch": 0.375,
+  "best_global_step": 4000,
+  "best_metric": 0.07567641884088516,
+  "best_model_checkpoint": "./training_output/checkpoint-4000",
+  "epoch": 0.2,
   "eval_steps": 1000,
-  "global_step": 3000,
+  "global_step": 4000,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.0625,
-      "grad_norm": 33.47822570800781,
-      "learning_rate": 1.8752500000000004e-05,
-      "loss": 0.1408,
+      "epoch": 0.025,
+      "grad_norm": 0.8834348917007446,
+      "learning_rate": 1.9501000000000002e-05,
+      "loss": 0.1437,
       "step": 500
     },
     {
-      "epoch": 0.125,
-      "grad_norm": 66.1669692993164,
-      "learning_rate": 1.7502500000000004e-05,
-      "loss": 0.1244,
+      "epoch": 0.05,
+      "grad_norm": 9.513919830322266,
+      "learning_rate": 1.9001e-05,
+      "loss": 0.1085,
       "step": 1000
     },
     {
-      "epoch": 0.125,
-      "eval_accuracy": 0.931,
-      "eval_loss": 0.08424179255962372,
-      "eval_runtime": 148.5886,
-      "eval_samples_per_second": 53.84,
-      "eval_steps_per_second": 3.365,
+      "epoch": 0.05,
+      "eval_accuracy": 0.91835,
+      "eval_loss": 0.10802757740020752,
+      "eval_runtime": 381.1922,
+      "eval_samples_per_second": 52.467,
+      "eval_steps_per_second": 3.279,
       "step": 1000
     },
     {
-      "epoch": 0.1875,
-      "grad_norm": 40.86865234375,
-      "learning_rate": 1.62525e-05,
-      "loss": 0.083,
+      "epoch": 0.075,
+      "grad_norm": 0.013394818641245365,
+      "learning_rate": 1.8501e-05,
+      "loss": 0.0965,
       "step": 1500
     },
     {
-      "epoch": 0.25,
-      "grad_norm": 0.046407029032707214,
-      "learning_rate": 1.5002500000000002e-05,
-      "loss": 0.0851,
+      "epoch": 0.1,
+      "grad_norm": 0.017180055379867554,
+      "learning_rate": 1.8001000000000003e-05,
+      "loss": 0.0716,
       "step": 2000
     },
     {
-      "epoch": 0.25,
-      "eval_accuracy": 0.7715,
-      "eval_loss": 0.20412930846214294,
-      "eval_runtime": 149.0575,
-      "eval_samples_per_second": 53.671,
-      "eval_steps_per_second": 3.354,
+      "epoch": 0.1,
+      "eval_accuracy": 0.8473,
+      "eval_loss": 0.25237375497817993,
+      "eval_runtime": 381.1616,
+      "eval_samples_per_second": 52.471,
+      "eval_steps_per_second": 3.279,
       "step": 2000
     },
     {
-      "epoch": 0.3125,
-      "grad_norm": 0.12031727284193039,
-      "learning_rate": 1.3752500000000003e-05,
-      "loss": 0.0689,
+      "epoch": 0.125,
+      "grad_norm": 0.017619747668504715,
+      "learning_rate": 1.7501e-05,
+      "loss": 0.0658,
       "step": 2500
     },
     {
-      "epoch": 0.375,
-      "grad_norm": 0.48724281787872314,
-      "learning_rate": 1.2502500000000003e-05,
-      "loss": 0.0632,
+      "epoch": 0.15,
+      "grad_norm": 0.527113676071167,
+      "learning_rate": 1.7001000000000002e-05,
+      "loss": 0.0615,
       "step": 3000
     },
     {
-      "epoch": 0.375,
-      "eval_accuracy": 0.91225,
-      "eval_loss": 0.16098278760910034,
-      "eval_runtime": 148.8996,
-      "eval_samples_per_second": 53.727,
-      "eval_steps_per_second": 3.358,
+      "epoch": 0.15,
+      "eval_accuracy": 0.9299,
+      "eval_loss": 0.11818733811378479,
+      "eval_runtime": 381.4889,
+      "eval_samples_per_second": 52.426,
+      "eval_steps_per_second": 3.277,
       "step": 3000
+    },
+    {
+      "epoch": 0.175,
+      "grad_norm": 140.78282165527344,
+      "learning_rate": 1.6501e-05,
+      "loss": 0.056,
+      "step": 3500
+    },
+    {
+      "epoch": 0.2,
+      "grad_norm": 0.9989501237869263,
+      "learning_rate": 1.6001e-05,
+      "loss": 0.0648,
+      "step": 4000
+    },
+    {
+      "epoch": 0.2,
+      "eval_accuracy": 0.9498,
+      "eval_loss": 0.07567641884088516,
+      "eval_runtime": 380.8034,
+      "eval_samples_per_second": 52.521,
+      "eval_steps_per_second": 3.283,
+      "step": 4000
     }
   ],
   "logging_steps": 500,
-  "max_steps": 8000,
+  "max_steps": 20000,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 1,
   "save_steps": 1000,
@@ -96,7 +119,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 3157332664320000.0,
+  "total_flos": 4209776885760000.0,
   "train_batch_size": 2,
   "trial_name": null,
   "trial_params": null

last-checkpoint/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5d1dcf0aa731896a45dc92ce645a0147a079fc2db258f2967bafe27b497eba6b
+oid sha256:4672bef08597eaac4166af409003080f023f24990f8f58c185b2b990119dc93e
 size 5841