End of training

trainer_state.json

@@ -11261,24 +11261,24 @@
       "epoch": 49.07,
       "step": 45000,
       "total_flos": 5.83519441453056e+18,
-      "train_loss": 0.05880288976563348,
-      "train_runtime": 37826.6777,
-      "train_samples_per_second": 4.759,
-      "train_steps_per_second": 1.19
+      "train_loss": 0.0,
+      "train_runtime": 0.3863,
+      "train_samples_per_second": 90611.862,
+      "train_steps_per_second": 90611.862
     },
     {
       "epoch": 49.07,
-      "eval_cer": 9.956974355224656,
-      "eval_loss": 0.7910374999046326,
-      "eval_runtime": 3295.3661,
-      "eval_samples_per_second": 0.657,
-      "eval_steps_per_second": 0.164,
-      "eval_wer": 31.400474471700036,
+      "eval_cer": 8.242936323479084,
+      "eval_loss": 0.5834927558898926,
+      "eval_runtime": 6238.1277,
+      "eval_samples_per_second": 0.745,
+      "eval_steps_per_second": 0.373,
+      "eval_wer": 24.854767937704718,
       "step": 45000
     }
   ],
-  "max_steps": 45000,
-  "num_train_epochs": 50,
+  "max_steps": 35000,
+  "num_train_epochs": 2,
   "total_flos": 5.83519441453056e+18,
   "trial_name": null,
   "trial_params": null