End of training

all_results.json

@@ -1,13 +1,13 @@
 {
-    "epoch": 1.05,
-    "eval_loss": 0.974454939365387,
-    "eval_runtime": 20.3472,
+    "epoch": 1.0,
+    "eval_loss": 0.6197173595428467,
+    "eval_runtime": 18.6322,
     "eval_samples": 64,
-    "eval_samples_per_second": 3.145,
-    "eval_steps_per_second": 0.098,
-    "eval_wer": 75.27472527472527,
-    "train_loss": 0.041942973931630455,
-    "train_runtime": 33.8193,
-    "train_samples_per_second": 18.924,
-    "train_steps_per_second": 0.591
+    "eval_samples_per_second": 3.435,
+    "eval_steps_per_second": 0.107,
+    "eval_wer": 60.43956043956044,
+    "train_loss": 0.6153274965286255,
+    "train_runtime": 782.3972,
+    "train_samples_per_second": 4.09,
+    "train_steps_per_second": 0.128
 }

eval_results.json

@@ -1,9 +1,9 @@
 {
-    "epoch": 1.05,
-    "eval_loss": 0.974454939365387,
-    "eval_runtime": 20.3472,
+    "epoch": 1.0,
+    "eval_loss": 0.6197173595428467,
+    "eval_runtime": 18.6322,
     "eval_samples": 64,
-    "eval_samples_per_second": 3.145,
-    "eval_steps_per_second": 0.098,
-    "eval_wer": 75.27472527472527
+    "eval_samples_per_second": 3.435,
+    "eval_steps_per_second": 0.107,
+    "eval_wer": 60.43956043956044
 }

runs/Dec12_19-28-45_13ea4557542f/events.out.tfevents.1670874502.13ea4557542f.8478.2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9bc918c0875efeac070effa512c0f38bc47ccada1467ef3d4c037fd4f656d329
+size 352

train.log

@@ -33,3 +33,9 @@
 {'loss': 0.6678, 'learning_rate': 3.1578947368421056e-06, 'epoch': 1.0}
 {'eval_loss': 0.5758526921272278, 'eval_wer': 61.53846153846154, 'eval_runtime': 19.5593, 'eval_samples_per_second': 3.272, 'eval_steps_per_second': 0.102, 'epoch': 1.0}
 {'train_runtime': 782.3972, 'train_samples_per_second': 4.09, 'train_steps_per_second': 0.128, 'train_loss': 0.6153274965286255, 'epoch': 1.0}
+***** train metrics *****
+  epoch                    =        1.0
+  train_loss               =     0.6153
+  train_runtime            = 0:13:02.39
+  train_samples_per_second =       4.09
+  train_steps_per_second   =      0.128

train_results.json

@@ -1,7 +1,7 @@
 {
-    "epoch": 1.05,
-    "train_loss": 0.041942973931630455,
-    "train_runtime": 33.8193,
-    "train_samples_per_second": 18.924,
-    "train_steps_per_second": 0.591
+    "epoch": 1.0,
+    "train_loss": 0.6153274965286255,
+    "train_runtime": 782.3972,
+    "train_samples_per_second": 4.09,
+    "train_steps_per_second": 0.128
 }

trainer_state.json

@@ -1,8 +1,8 @@
 {
-  "best_metric": 75.27472527472527,
-  "best_model_checkpoint": "./checkpoint-20",
-  "epoch": 1.05,
-  "global_step": 21,
+  "best_metric": 60.43956043956044,
+  "best_model_checkpoint": "./checkpoint-70",
+  "epoch": 1.0,
+  "global_step": 100,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
@@ -38,18 +38,138 @@
       "step": 20
     },
     {
-      "epoch": 1.05,
-      "step": 21,
-      "total_flos": 1.654389080064e+16,
-      "train_loss": 0.041942973931630455,
-      "train_runtime": 33.8193,
-      "train_samples_per_second": 18.924,
-      "train_steps_per_second": 0.591
+      "epoch": 0.3,
+      "learning_rate": 7.68421052631579e-05,
+      "loss": 0.9934,
+      "step": 30
+    },
+    {
+      "epoch": 0.3,
+      "eval_loss": 0.811407208442688,
+      "eval_runtime": 17.9554,
+      "eval_samples_per_second": 3.564,
+      "eval_steps_per_second": 0.111,
+      "eval_wer": 72.16117216117216,
+      "step": 30
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 6.631578947368421e-05,
+      "loss": 0.9568,
+      "step": 40
+    },
+    {
+      "epoch": 0.4,
+      "eval_loss": 0.7814445495605469,
+      "eval_runtime": 44.1521,
+      "eval_samples_per_second": 1.45,
+      "eval_steps_per_second": 0.045,
+      "eval_wer": 72.7106227106227,
+      "step": 40
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 5.5789473684210526e-05,
+      "loss": 0.6856,
+      "step": 50
+    },
+    {
+      "epoch": 0.5,
+      "eval_loss": 0.7517387866973877,
+      "eval_runtime": 19.7741,
+      "eval_samples_per_second": 3.237,
+      "eval_steps_per_second": 0.101,
+      "eval_wer": 76.92307692307693,
+      "step": 50
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.5263157894736846e-05,
+      "loss": 0.7808,
+      "step": 60
+    },
+    {
+      "epoch": 0.6,
+      "eval_loss": 0.6514347195625305,
+      "eval_runtime": 19.745,
+      "eval_samples_per_second": 3.241,
+      "eval_steps_per_second": 0.101,
+      "eval_wer": 63.55311355311355,
+      "step": 60
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 3.473684210526316e-05,
+      "loss": 0.6826,
+      "step": 70
+    },
+    {
+      "epoch": 0.7,
+      "eval_loss": 0.6197173595428467,
+      "eval_runtime": 19.7014,
+      "eval_samples_per_second": 3.248,
+      "eval_steps_per_second": 0.102,
+      "eval_wer": 60.43956043956044,
+      "step": 70
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.4210526315789474e-05,
+      "loss": 0.7832,
+      "step": 80
+    },
+    {
+      "epoch": 0.8,
+      "eval_loss": 0.6129069328308105,
+      "eval_runtime": 18.9507,
+      "eval_samples_per_second": 3.377,
+      "eval_steps_per_second": 0.106,
+      "eval_wer": 65.93406593406593,
+      "step": 80
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.3684210526315791e-05,
+      "loss": 0.6031,
+      "step": 90
+    },
+    {
+      "epoch": 0.9,
+      "eval_loss": 0.5876654982566833,
+      "eval_runtime": 20.4075,
+      "eval_samples_per_second": 3.136,
+      "eval_steps_per_second": 0.098,
+      "eval_wer": 61.35531135531136,
+      "step": 90
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 3.1578947368421056e-06,
+      "loss": 0.6678,
+      "step": 100
+    },
+    {
+      "epoch": 1.0,
+      "eval_loss": 0.5758526921272278,
+      "eval_runtime": 19.5593,
+      "eval_samples_per_second": 3.272,
+      "eval_steps_per_second": 0.102,
+      "eval_wer": 61.53846153846154,
+      "step": 100
+    },
+    {
+      "epoch": 1.0,
+      "step": 100,
+      "total_flos": 7.8780432384e+16,
+      "train_loss": 0.6153274965286255,
+      "train_runtime": 782.3972,
+      "train_samples_per_second": 4.09,
+      "train_steps_per_second": 0.128
     }
   ],
-  "max_steps": 20,
+  "max_steps": 100,
   "num_train_epochs": 9223372036854775807,
-  "total_flos": 1.654389080064e+16,
+  "total_flos": 7.8780432384e+16,
   "trial_name": null,
   "trial_params": null
 }