Upload trainer_state.json

trainer_state.json

@@ -1,211 +1,151 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 20.8,
-  "global_step": 5200,
+  "epoch": 7.72454448017149,
+  "global_step": 3600,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 1.6,
+      "epoch": 0.86,
       "learning_rate": 0.0002394,
-      "loss": 3.4203,
+      "loss": 3.2993,
       "step": 400
     },
     {
-      "epoch": 1.6,
+      "epoch": 0.86,
       "eval_loss": Infinity,
-      "eval_runtime": 1205.6526,
-      "eval_samples_per_second": 3.822,
-      "eval_steps_per_second": 0.478,
-      "eval_wer": 0.5028872209091932,
+      "eval_runtime": 828.7984,
+      "eval_samples_per_second": 9.006,
+      "eval_steps_per_second": 1.126,
+      "eval_wer": 0.36457931084030065,
       "step": 400
     },
     {
-      "epoch": 3.2,
-      "learning_rate": 0.0002873142857142857,
-      "loss": 0.4156,
+      "epoch": 1.72,
+      "learning_rate": 0.0002723048327137546,
+      "loss": 0.3837,
       "step": 800
     },
     {
-      "epoch": 3.2,
+      "epoch": 1.72,
       "eval_loss": Infinity,
-      "eval_runtime": 1205.8762,
-      "eval_samples_per_second": 3.821,
-      "eval_steps_per_second": 0.478,
-      "eval_wer": 0.21501354872780476,
+      "eval_runtime": 820.745,
+      "eval_samples_per_second": 9.094,
+      "eval_steps_per_second": 1.137,
+      "eval_wer": 0.25638285452569753,
       "step": 800
     },
     {
-      "epoch": 4.8,
-      "learning_rate": 0.0002702142857142857,
-      "loss": 0.248,
+      "epoch": 2.57,
+      "learning_rate": 0.00023522304832713752,
+      "loss": 0.2576,
       "step": 1200
     },
     {
-      "epoch": 4.8,
+      "epoch": 2.57,
       "eval_loss": Infinity,
-      "eval_runtime": 1206.2112,
-      "eval_samples_per_second": 3.82,
-      "eval_steps_per_second": 0.478,
-      "eval_wer": 0.19620116181423092,
+      "eval_runtime": 816.7249,
+      "eval_samples_per_second": 9.139,
+      "eval_steps_per_second": 1.142,
+      "eval_wer": 0.17959103032697818,
       "step": 1200
     },
     {
-      "epoch": 6.4,
-      "learning_rate": 0.00025315714285714284,
-      "loss": 0.1884,
+      "epoch": 3.43,
+      "learning_rate": 0.0001981412639405204,
+      "loss": 0.1963,
       "step": 1600
     },
     {
-      "epoch": 6.4,
+      "epoch": 3.43,
       "eval_loss": Infinity,
-      "eval_runtime": 1198.1731,
-      "eval_samples_per_second": 3.846,
-      "eval_steps_per_second": 0.481,
-      "eval_wer": 0.18049675782839433,
+      "eval_runtime": 819.1322,
+      "eval_samples_per_second": 9.112,
+      "eval_steps_per_second": 1.139,
+      "eval_wer": 0.1695059854696326,
       "step": 1600
     },
     {
-      "epoch": 8.0,
-      "learning_rate": 0.00023605714285714283,
-      "loss": 0.1488,
+      "epoch": 4.29,
+      "learning_rate": 0.00016096654275092934,
+      "loss": 0.158,
       "step": 2000
     },
     {
-      "epoch": 8.0,
+      "epoch": 4.29,
       "eval_loss": Infinity,
-      "eval_runtime": 1195.3799,
-      "eval_samples_per_second": 3.855,
-      "eval_steps_per_second": 0.482,
-      "eval_wer": 0.18608073829744867,
+      "eval_runtime": 815.3058,
+      "eval_samples_per_second": 9.155,
+      "eval_steps_per_second": 1.144,
+      "eval_wer": 0.16685675284904763,
       "step": 2000
     },
     {
-      "epoch": 9.6,
-      "learning_rate": 0.00021899999999999998,
-      "loss": 0.1232,
+      "epoch": 5.15,
+      "learning_rate": 0.00012388475836431226,
+      "loss": 0.125,
       "step": 2400
     },
     {
-      "epoch": 9.6,
+      "epoch": 5.15,
       "eval_loss": Infinity,
-      "eval_runtime": 1196.1097,
-      "eval_samples_per_second": 3.852,
-      "eval_steps_per_second": 0.482,
-      "eval_wer": 0.17618540546624073,
+      "eval_runtime": 815.712,
+      "eval_samples_per_second": 9.15,
+      "eval_steps_per_second": 1.144,
+      "eval_wer": 0.15157202769570643,
       "step": 2400
     },
     {
-      "epoch": 11.2,
-      "learning_rate": 0.0002019,
-      "loss": 0.107,
+      "epoch": 6.01,
+      "learning_rate": 8.680297397769517e-05,
+      "loss": 0.101,
       "step": 2800
     },
     {
-      "epoch": 11.2,
+      "epoch": 6.01,
       "eval_loss": Infinity,
-      "eval_runtime": 1192.8335,
-      "eval_samples_per_second": 3.863,
-      "eval_steps_per_second": 0.483,
-      "eval_wer": 0.17763118025435248,
+      "eval_runtime": 815.5191,
+      "eval_samples_per_second": 9.152,
+      "eval_steps_per_second": 1.144,
+      "eval_wer": 0.1440015087155263,
       "step": 2800
     },
     {
-      "epoch": 12.8,
-      "learning_rate": 0.0001848428571428571,
-      "loss": 0.2435,
+      "epoch": 6.87,
+      "learning_rate": 4.972118959107806e-05,
+      "loss": 0.0793,
       "step": 3200
     },
     {
-      "epoch": 12.8,
+      "epoch": 6.87,
       "eval_loss": Infinity,
-      "eval_runtime": 1189.2109,
-      "eval_samples_per_second": 3.875,
-      "eval_steps_per_second": 0.484,
-      "eval_wer": 0.5853483278359262,
+      "eval_runtime": 816.2609,
+      "eval_samples_per_second": 9.144,
+      "eval_steps_per_second": 1.143,
+      "eval_wer": 0.139223909548912,
       "step": 3200
     },
     {
-      "epoch": 14.4,
-      "learning_rate": 0.00016774285714285715,
-      "loss": 0.7331,
+      "epoch": 7.72,
+      "learning_rate": 1.2639405204460965e-05,
+      "loss": 0.0642,
       "step": 3600
     },
     {
-      "epoch": 14.4,
+      "epoch": 7.72,
       "eval_loss": Infinity,
-      "eval_runtime": 1196.7777,
-      "eval_samples_per_second": 3.85,
-      "eval_steps_per_second": 0.481,
-      "eval_wer": 0.5020128301690777,
+      "eval_runtime": 815.052,
+      "eval_samples_per_second": 9.158,
+      "eval_steps_per_second": 1.145,
+      "eval_wer": 0.1354611011827252,
       "step": 3600
-    },
-    {
-      "epoch": 16.0,
-      "learning_rate": 0.0001506857142857143,
-      "loss": 0.5402,
-      "step": 4000
-    },
-    {
-      "epoch": 16.0,
-      "eval_loss": Infinity,
-      "eval_runtime": 1194.9864,
-      "eval_samples_per_second": 3.856,
-      "eval_steps_per_second": 0.482,
-      "eval_wer": 0.29286895393432544,
-      "step": 4000
-    },
-    {
-      "epoch": 17.6,
-      "learning_rate": 0.00013358571428571426,
-      "loss": 1.3361,
-      "step": 4400
-    },
-    {
-      "epoch": 17.6,
-      "eval_loss": Infinity,
-      "eval_runtime": 1207.0654,
-      "eval_samples_per_second": 3.818,
-      "eval_steps_per_second": 0.477,
-      "eval_wer": 0.9734566137703555,
-      "step": 4400
-    },
-    {
-      "epoch": 19.2,
-      "learning_rate": 0.0001165285714285714,
-      "loss": 0.9035,
-      "step": 4800
-    },
-    {
-      "epoch": 19.2,
-      "eval_loss": Infinity,
-      "eval_runtime": 1210.7223,
-      "eval_samples_per_second": 3.806,
-      "eval_steps_per_second": 0.476,
-      "eval_wer": 0.4126085413257841,
-      "step": 4800
-    },
-    {
-      "epoch": 20.8,
-      "learning_rate": 9.947142857142856e-05,
-      "loss": 0.5824,
-      "step": 5200
-    },
-    {
-      "epoch": 20.8,
-      "eval_loss": Infinity,
-      "eval_runtime": 1209.2109,
-      "eval_samples_per_second": 3.811,
-      "eval_steps_per_second": 0.476,
-      "eval_wer": 0.2804110502211949,
-      "step": 5200
     }
   ],
-  "max_steps": 7500,
-  "num_train_epochs": 30,
-  "total_flos": 8.725540278590487e+19,
+  "max_steps": 3728,
+  "num_train_epochs": 8,
+  "total_flos": 4.913815144448316e+19,
   "trial_name": null,
   "trial_params": null
 }