Training in progress, step 7000, checkpoint

last-checkpoint/config.json

@@ -19,7 +19,7 @@
   "num_hidden_layers": 12,
   "pad_token_id": 1,
   "position_embedding_type": "absolute",
-  "transformers_version": "4.57.3",
+  "transformers_version": "4.57.1",
   "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 sha256:7839750122f45b034760dbba82058b3dbda08cb6abecf42d205af5e913b05f35
+oid sha256:41da70138a1773e4f5cd2778674af186f8e5ad4d5bfbd72c39a71b1b96bc5185
 size 498612824

last-checkpoint/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9e2657af39b3a797b07074a3b10528e18436b00c2a7b3e28c2f4ef26ab0e4474
+oid sha256:fcb38ea326cd284005454648e073cc75673ef6d3edbb0d25347b119555de5f08
 size 997345931

last-checkpoint/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:903ae64856702f56aca34b3a83e463f26492f960398b5b01c03e210465cf26b7
+oid sha256:db4e3de49d1806c7044e452123303f3ec21f0d1186d1b23be8ec33eb1e34488d
 size 14645

last-checkpoint/scaler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7bb2b51b22f222dc6df63edb791d76de5de25b8bb724e885f8753e34cfb8ae10
+oid sha256:be40a03f70bb2a35bb567114fb5ceb05ce1222c8df24b1e1100746c617c5be10
 size 1383

last-checkpoint/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d6ac56f1033e9e261dd3b41721e9cff6132800378e9f573cebb3fd9f04365c7d
+oid sha256:c7bbc1b39311f89c58ff1ea8b0baf3cad68be72482b939daed1a03b8c1807179
 size 1465

last-checkpoint/trainer_state.json

@@ -1,155 +1,178 @@
 {
-  "best_global_step": 4000,
-  "best_metric": 0.07567641884088516,
-  "best_model_checkpoint": "./training_output/checkpoint-4000",
-  "epoch": 0.3,
+  "best_global_step": 6000,
+  "best_metric": 0.06669048964977264,
+  "best_model_checkpoint": "./training_output/checkpoint-6000",
+  "epoch": 0.875,
   "eval_steps": 1000,
-  "global_step": 6000,
+  "global_step": 7000,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.025,
-      "grad_norm": 0.8834348917007446,
-      "learning_rate": 1.9501000000000002e-05,
-      "loss": 0.1437,
+      "epoch": 0.0625,
+      "grad_norm": 33.47822570800781,
+      "learning_rate": 1.8752500000000004e-05,
+      "loss": 0.1408,
       "step": 500
     },
     {
-      "epoch": 0.05,
-      "grad_norm": 9.513919830322266,
-      "learning_rate": 1.9001e-05,
-      "loss": 0.1085,
+      "epoch": 0.125,
+      "grad_norm": 66.1669692993164,
+      "learning_rate": 1.7502500000000004e-05,
+      "loss": 0.1244,
       "step": 1000
     },
     {
-      "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,
+      "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,
       "step": 1000
     },
     {
-      "epoch": 0.075,
-      "grad_norm": 0.013394818641245365,
-      "learning_rate": 1.8501e-05,
-      "loss": 0.0965,
+      "epoch": 0.1875,
+      "grad_norm": 40.86865234375,
+      "learning_rate": 1.62525e-05,
+      "loss": 0.083,
       "step": 1500
     },
     {
-      "epoch": 0.1,
-      "grad_norm": 0.017180055379867554,
-      "learning_rate": 1.8001000000000003e-05,
-      "loss": 0.0716,
+      "epoch": 0.25,
+      "grad_norm": 0.046407029032707214,
+      "learning_rate": 1.5002500000000002e-05,
+      "loss": 0.0851,
       "step": 2000
     },
     {
-      "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,
+      "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,
       "step": 2000
     },
     {
-      "epoch": 0.125,
-      "grad_norm": 0.017619747668504715,
-      "learning_rate": 1.7501e-05,
-      "loss": 0.0658,
+      "epoch": 0.3125,
+      "grad_norm": 0.12031727284193039,
+      "learning_rate": 1.3752500000000003e-05,
+      "loss": 0.0689,
       "step": 2500
     },
     {
-      "epoch": 0.15,
-      "grad_norm": 0.527113676071167,
-      "learning_rate": 1.7001000000000002e-05,
-      "loss": 0.0615,
+      "epoch": 0.375,
+      "grad_norm": 0.48724281787872314,
+      "learning_rate": 1.2502500000000003e-05,
+      "loss": 0.0632,
       "step": 3000
     },
     {
-      "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,
+      "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,
       "step": 3000
     },
     {
-      "epoch": 0.175,
-      "grad_norm": 140.78282165527344,
-      "learning_rate": 1.6501e-05,
-      "loss": 0.056,
+      "epoch": 0.4375,
+      "grad_norm": 0.004755391739308834,
+      "learning_rate": 1.1252500000000001e-05,
+      "loss": 0.0525,
       "step": 3500
     },
     {
-      "epoch": 0.2,
-      "grad_norm": 0.9989501237869263,
-      "learning_rate": 1.6001e-05,
-      "loss": 0.0648,
+      "epoch": 0.5,
+      "grad_norm": 0.2557525336742401,
+      "learning_rate": 1.0002500000000001e-05,
+      "loss": 0.0521,
       "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,
+      "epoch": 0.5,
+      "eval_accuracy": 0.922375,
+      "eval_loss": 0.09066133201122284,
+      "eval_runtime": 148.5612,
+      "eval_samples_per_second": 53.85,
+      "eval_steps_per_second": 3.366,
       "step": 4000
     },
     {
-      "epoch": 0.225,
-      "grad_norm": 0.01812303625047207,
-      "learning_rate": 1.5501000000000003e-05,
-      "loss": 0.0487,
+      "epoch": 0.5625,
+      "grad_norm": 0.641652524471283,
+      "learning_rate": 8.7525e-06,
+      "loss": 0.0463,
       "step": 4500
     },
     {
-      "epoch": 0.25,
-      "grad_norm": 0.05552659556269646,
-      "learning_rate": 1.5001000000000001e-05,
-      "loss": 0.0522,
+      "epoch": 0.625,
+      "grad_norm": 0.1168695017695427,
+      "learning_rate": 7.502500000000001e-06,
+      "loss": 0.0426,
       "step": 5000
     },
     {
-      "epoch": 0.25,
-      "eval_accuracy": 0.92725,
-      "eval_loss": 0.12006673216819763,
-      "eval_runtime": 380.6188,
-      "eval_samples_per_second": 52.546,
-      "eval_steps_per_second": 3.284,
+      "epoch": 0.625,
+      "eval_accuracy": 0.878375,
+      "eval_loss": 0.24468091130256653,
+      "eval_runtime": 148.7322,
+      "eval_samples_per_second": 53.788,
+      "eval_steps_per_second": 3.362,
       "step": 5000
     },
     {
-      "epoch": 0.275,
-      "grad_norm": 0.14319103956222534,
-      "learning_rate": 1.4501e-05,
-      "loss": 0.0554,
+      "epoch": 0.6875,
+      "grad_norm": 0.001182532636448741,
+      "learning_rate": 6.2525e-06,
+      "loss": 0.0286,
       "step": 5500
     },
     {
-      "epoch": 0.3,
-      "grad_norm": 0.012562028132379055,
-      "learning_rate": 1.4001e-05,
-      "loss": 0.0377,
+      "epoch": 0.75,
+      "grad_norm": 0.0016105002723634243,
+      "learning_rate": 5.0025e-06,
+      "loss": 0.0341,
       "step": 6000
     },
     {
-      "epoch": 0.3,
-      "eval_accuracy": 0.95545,
-      "eval_loss": 0.08464282751083374,
-      "eval_runtime": 380.6212,
-      "eval_samples_per_second": 52.546,
-      "eval_steps_per_second": 3.284,
+      "epoch": 0.75,
+      "eval_accuracy": 0.964125,
+      "eval_loss": 0.06669048964977264,
+      "eval_runtime": 148.8829,
+      "eval_samples_per_second": 53.733,
+      "eval_steps_per_second": 3.358,
       "step": 6000
+    },
+    {
+      "epoch": 0.8125,
+      "grad_norm": 0.00258028507232666,
+      "learning_rate": 3.7525e-06,
+      "loss": 0.0337,
+      "step": 6500
+    },
+    {
+      "epoch": 0.875,
+      "grad_norm": 0.005781837739050388,
+      "learning_rate": 2.5024999999999998e-06,
+      "loss": 0.0252,
+      "step": 7000
+    },
+    {
+      "epoch": 0.875,
+      "eval_accuracy": 0.957,
+      "eval_loss": 0.0872097983956337,
+      "eval_runtime": 148.3454,
+      "eval_samples_per_second": 53.928,
+      "eval_steps_per_second": 3.371,
+      "step": 7000
     }
   ],
   "logging_steps": 500,
-  "max_steps": 20000,
+  "max_steps": 8000,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 1,
   "save_steps": 1000,
@@ -165,7 +188,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 6314665328640000.0,
+  "total_flos": 7367109550080000.0,
   "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 sha256:4672bef08597eaac4166af409003080f023f24990f8f58c185b2b990119dc93e
+oid sha256:5d1dcf0aa731896a45dc92ce645a0147a079fc2db258f2967bafe27b497eba6b
 size 5841