End of training

all_results.json

@@ -1,15 +1,15 @@
 {
-    "epoch": 50.0,
-    "eval_cer": 0.19657237249638654,
-    "eval_loss": 0.9153681993484497,
-    "eval_runtime": 45.7458,
+    "epoch": 50.63,
+    "eval_cer": 0.1969102547256584,
+    "eval_loss": 0.9162458181381226,
+    "eval_runtime": 46.1646,
     "eval_samples": 481,
-    "eval_samples_per_second": 10.515,
-    "eval_steps_per_second": 1.333,
-    "eval_wer": 0.5137278308321964,
-    "train_loss": 2.643854008807412,
-    "train_runtime": 15714.4593,
+    "eval_samples_per_second": 10.419,
+    "eval_steps_per_second": 1.321,
+    "eval_wer": 0.5156036834924966,
+    "train_loss": 0.07298430502414703,
+    "train_runtime": 2125.6437,
     "train_samples": 2528,
-    "train_samples_per_second": 8.044,
-    "train_steps_per_second": 0.251
+    "train_samples_per_second": 60.217,
+    "train_steps_per_second": 1.882
 }

eval_results.json

@@ -1,10 +1,10 @@
 {
-    "epoch": 50.0,
-    "eval_cer": 0.19657237249638654,
-    "eval_loss": 0.9153681993484497,
-    "eval_runtime": 45.7458,
+    "epoch": 50.63,
+    "eval_cer": 0.1969102547256584,
+    "eval_loss": 0.9162458181381226,
+    "eval_runtime": 46.1646,
     "eval_samples": 481,
-    "eval_samples_per_second": 10.515,
-    "eval_steps_per_second": 1.333,
-    "eval_wer": 0.5137278308321964
+    "eval_samples_per_second": 10.419,
+    "eval_steps_per_second": 1.321,
+    "eval_wer": 0.5156036834924966
 }

runs/Dec16_13-55-02_129-146-104-29/events.out.tfevents.1671201437.129-146-104-29.128095.2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:08206f99f4f7b3fbfa384e19ec6ae571dfe46052f6556baacc5e7735d060a873
+size 405

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 50.0,
-    "train_loss": 2.643854008807412,
-    "train_runtime": 15714.4593,
+    "epoch": 50.63,
+    "train_loss": 0.07298430502414703,
+    "train_runtime": 2125.6437,
     "train_samples": 2528,
-    "train_samples_per_second": 8.044,
-    "train_steps_per_second": 0.251
+    "train_samples_per_second": 60.217,
+    "train_steps_per_second": 1.882
 }

trainer_state.json

@@ -1,8 +1,8 @@
 {
-  "best_metric": null,
-  "best_model_checkpoint": null,
-  "epoch": 50.0,
-  "global_step": 3950,
+  "best_metric": 0.9162458181381226,
+  "best_model_checkpoint": "./checkpoint-4000",
+  "epoch": 50.63291139240506,
+  "global_step": 4000,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
@@ -2179,287 +2179,327 @@
     },
     {
       "epoch": 44.43,
-      "learning_rate": 1.696153846153846e-05,
-      "loss": 0.6001,
+      "learning_rate": 1.8412499999999997e-05,
+      "loss": 0.6002,
       "step": 3510
     },
     {
       "epoch": 44.56,
-      "learning_rate": 1.6576923076923074e-05,
-      "loss": 0.6187,
+      "learning_rate": 1.8037499999999998e-05,
+      "loss": 0.6191,
       "step": 3520
     },
     {
       "epoch": 44.68,
-      "learning_rate": 1.6192307692307692e-05,
-      "loss": 0.5808,
+      "learning_rate": 1.76625e-05,
+      "loss": 0.5811,
       "step": 3530
     },
     {
       "epoch": 44.81,
-      "learning_rate": 1.5807692307692307e-05,
-      "loss": 0.6285,
+      "learning_rate": 1.72875e-05,
+      "loss": 0.6299,
       "step": 3540
     },
     {
       "epoch": 44.94,
-      "learning_rate": 1.542307692307692e-05,
-      "loss": 0.5612,
+      "learning_rate": 1.6912499999999998e-05,
+      "loss": 0.5605,
       "step": 3550
     },
     {
       "epoch": 45.06,
-      "learning_rate": 1.5038461538461538e-05,
-      "loss": 0.617,
+      "learning_rate": 1.65375e-05,
+      "loss": 0.6183,
       "step": 3560
     },
     {
       "epoch": 45.19,
-      "learning_rate": 1.4653846153846151e-05,
-      "loss": 0.5858,
+      "learning_rate": 1.61625e-05,
+      "loss": 0.5852,
       "step": 3570
     },
     {
       "epoch": 45.32,
-      "learning_rate": 1.426923076923077e-05,
-      "loss": 0.5933,
+      "learning_rate": 1.5787499999999997e-05,
+      "loss": 0.594,
       "step": 3580
     },
     {
       "epoch": 45.44,
-      "learning_rate": 1.3884615384615383e-05,
-      "loss": 0.5971,
+      "learning_rate": 1.54125e-05,
+      "loss": 0.5965,
       "step": 3590
     },
     {
       "epoch": 45.57,
-      "learning_rate": 1.3499999999999998e-05,
-      "loss": 0.5997,
+      "learning_rate": 1.50375e-05,
+      "loss": 0.6005,
       "step": 3600
     },
     {
       "epoch": 45.7,
-      "learning_rate": 1.3115384615384614e-05,
-      "loss": 0.5872,
+      "learning_rate": 1.4662499999999999e-05,
+      "loss": 0.5884,
       "step": 3610
     },
     {
       "epoch": 45.82,
-      "learning_rate": 1.273076923076923e-05,
-      "loss": 0.5875,
+      "learning_rate": 1.4287499999999998e-05,
+      "loss": 0.5884,
       "step": 3620
     },
     {
       "epoch": 45.95,
-      "learning_rate": 1.2346153846153844e-05,
-      "loss": 0.5623,
+      "learning_rate": 1.39125e-05,
+      "loss": 0.5628,
       "step": 3630
     },
     {
       "epoch": 46.08,
-      "learning_rate": 1.196153846153846e-05,
-      "loss": 0.6336,
+      "learning_rate": 1.3537499999999999e-05,
+      "loss": 0.6339,
       "step": 3640
     },
     {
       "epoch": 46.2,
-      "learning_rate": 1.1576923076923076e-05,
-      "loss": 0.5559,
+      "learning_rate": 1.3162499999999998e-05,
+      "loss": 0.5578,
       "step": 3650
     },
     {
       "epoch": 46.33,
-      "learning_rate": 1.119230769230769e-05,
-      "loss": 0.6232,
+      "learning_rate": 1.2787499999999999e-05,
+      "loss": 0.6239,
       "step": 3660
     },
     {
       "epoch": 46.46,
-      "learning_rate": 1.0807692307692307e-05,
-      "loss": 0.5881,
+      "learning_rate": 1.24125e-05,
+      "loss": 0.5872,
       "step": 3670
     },
     {
       "epoch": 46.58,
-      "learning_rate": 1.0423076923076922e-05,
-      "loss": 0.5716,
+      "learning_rate": 1.20375e-05,
+      "loss": 0.5697,
       "step": 3680
     },
     {
       "epoch": 46.71,
-      "learning_rate": 1.0038461538461537e-05,
-      "loss": 0.549,
+      "learning_rate": 1.1662499999999999e-05,
+      "loss": 0.5475,
       "step": 3690
     },
     {
       "epoch": 46.84,
-      "learning_rate": 9.653846153846154e-06,
-      "loss": 0.5987,
+      "learning_rate": 1.1287499999999998e-05,
+      "loss": 0.5979,
       "step": 3700
     },
     {
       "epoch": 46.96,
-      "learning_rate": 9.269230769230768e-06,
-      "loss": 0.5751,
+      "learning_rate": 1.0912499999999998e-05,
+      "loss": 0.5742,
       "step": 3710
     },
     {
       "epoch": 47.09,
-      "learning_rate": 8.884615384615383e-06,
-      "loss": 0.607,
+      "learning_rate": 1.05375e-05,
+      "loss": 0.6054,
       "step": 3720
     },
     {
       "epoch": 47.22,
-      "learning_rate": 8.499999999999998e-06,
-      "loss": 0.5768,
+      "learning_rate": 1.01625e-05,
+      "loss": 0.5777,
       "step": 3730
     },
     {
       "epoch": 47.34,
-      "learning_rate": 8.115384615384615e-06,
-      "loss": 0.5738,
+      "learning_rate": 9.787499999999999e-06,
+      "loss": 0.5734,
       "step": 3740
     },
     {
       "epoch": 47.47,
-      "learning_rate": 7.73076923076923e-06,
-      "loss": 0.5304,
+      "learning_rate": 9.412499999999998e-06,
+      "loss": 0.5322,
       "step": 3750
     },
     {
       "epoch": 47.59,
-      "learning_rate": 7.346153846153845e-06,
-      "loss": 0.628,
+      "learning_rate": 9.0375e-06,
+      "loss": 0.6287,
       "step": 3760
     },
     {
       "epoch": 47.72,
-      "learning_rate": 6.96153846153846e-06,
-      "loss": 0.5455,
+      "learning_rate": 8.6625e-06,
+      "loss": 0.547,
       "step": 3770
     },
     {
       "epoch": 47.85,
-      "learning_rate": 6.576923076923076e-06,
-      "loss": 0.6422,
+      "learning_rate": 8.2875e-06,
+      "loss": 0.6414,
       "step": 3780
     },
     {
       "epoch": 47.97,
-      "learning_rate": 6.192307692307692e-06,
-      "loss": 0.5657,
+      "learning_rate": 7.9125e-06,
+      "loss": 0.5661,
       "step": 3790
     },
     {
       "epoch": 48.1,
-      "learning_rate": 5.807692307692307e-06,
-      "loss": 0.5906,
+      "learning_rate": 7.537499999999999e-06,
+      "loss": 0.5893,
       "step": 3800
     },
     {
       "epoch": 48.23,
-      "learning_rate": 5.4230769230769225e-06,
-      "loss": 0.5559,
+      "learning_rate": 7.1625e-06,
+      "loss": 0.556,
       "step": 3810
     },
     {
       "epoch": 48.35,
-      "learning_rate": 5.038461538461538e-06,
-      "loss": 0.6286,
+      "learning_rate": 6.787499999999999e-06,
+      "loss": 0.6265,
       "step": 3820
     },
     {
       "epoch": 48.48,
-      "learning_rate": 4.653846153846154e-06,
-      "loss": 0.5632,
+      "learning_rate": 6.4125e-06,
+      "loss": 0.5644,
       "step": 3830
     },
     {
       "epoch": 48.61,
-      "learning_rate": 4.269230769230769e-06,
-      "loss": 0.6231,
+      "learning_rate": 6.037499999999999e-06,
+      "loss": 0.6202,
       "step": 3840
     },
     {
       "epoch": 48.73,
-      "learning_rate": 3.884615384615384e-06,
-      "loss": 0.559,
+      "learning_rate": 5.6624999999999996e-06,
+      "loss": 0.5581,
       "step": 3850
     },
     {
       "epoch": 48.86,
-      "learning_rate": 3.5e-06,
-      "loss": 0.5731,
+      "learning_rate": 5.287499999999999e-06,
+      "loss": 0.572,
       "step": 3860
     },
     {
       "epoch": 48.99,
-      "learning_rate": 3.1153846153846153e-06,
-      "loss": 0.5573,
+      "learning_rate": 4.9125e-06,
+      "loss": 0.5559,
       "step": 3870
     },
     {
       "epoch": 49.11,
-      "learning_rate": 2.73076923076923e-06,
-      "loss": 0.6044,
+      "learning_rate": 4.537499999999999e-06,
+      "loss": 0.6013,
       "step": 3880
     },
     {
       "epoch": 49.24,
-      "learning_rate": 2.346153846153846e-06,
-      "loss": 0.5496,
+      "learning_rate": 4.1624999999999995e-06,
+      "loss": 0.5498,
       "step": 3890
     },
     {
       "epoch": 49.37,
-      "learning_rate": 1.9615384615384612e-06,
-      "loss": 0.5895,
+      "learning_rate": 3.7874999999999997e-06,
+      "loss": 0.5883,
       "step": 3900
     },
     {
       "epoch": 49.49,
-      "learning_rate": 1.5769230769230768e-06,
-      "loss": 0.5815,
+      "learning_rate": 3.4124999999999995e-06,
+      "loss": 0.5777,
       "step": 3910
     },
     {
       "epoch": 49.62,
-      "learning_rate": 1.1923076923076923e-06,
-      "loss": 0.5782,
+      "learning_rate": 3.0374999999999997e-06,
+      "loss": 0.5768,
       "step": 3920
     },
     {
       "epoch": 49.75,
-      "learning_rate": 8.076923076923076e-07,
-      "loss": 0.5605,
+      "learning_rate": 2.6624999999999995e-06,
+      "loss": 0.5603,
       "step": 3930
     },
     {
       "epoch": 49.87,
-      "learning_rate": 4.230769230769231e-07,
-      "loss": 0.5849,
+      "learning_rate": 2.2874999999999997e-06,
+      "loss": 0.5814,
       "step": 3940
     },
     {
       "epoch": 50.0,
-      "learning_rate": 3.846153846153846e-08,
-      "loss": 0.5558,
+      "learning_rate": 1.9125e-06,
+      "loss": 0.5562,
       "step": 3950
     },
     {
-      "epoch": 50.0,
-      "step": 3950,
-      "total_flos": 4.545011460667043e+19,
-      "train_loss": 2.643854008807412,
-      "train_runtime": 15714.4593,
-      "train_samples_per_second": 8.044,
-      "train_steps_per_second": 0.251
+      "epoch": 50.13,
+      "learning_rate": 1.5374999999999999e-06,
+      "loss": 0.5858,
+      "step": 3960
+    },
+    {
+      "epoch": 50.25,
+      "learning_rate": 1.1624999999999999e-06,
+      "loss": 0.5279,
+      "step": 3970
+    },
+    {
+      "epoch": 50.38,
+      "learning_rate": 7.875e-07,
+      "loss": 0.5734,
+      "step": 3980
+    },
+    {
+      "epoch": 50.51,
+      "learning_rate": 4.124999999999999e-07,
+      "loss": 0.5895,
+      "step": 3990
+    },
+    {
+      "epoch": 50.63,
+      "learning_rate": 3.75e-08,
+      "loss": 0.5935,
+      "step": 4000
+    },
+    {
+      "epoch": 50.63,
+      "eval_cer": 0.1969102547256584,
+      "eval_loss": 0.9162458181381226,
+      "eval_runtime": 44.8405,
+      "eval_samples_per_second": 10.727,
+      "eval_steps_per_second": 1.36,
+      "eval_wer": 0.5156036834924966,
+      "step": 4000
+    },
+    {
+      "epoch": 50.63,
+      "step": 4000,
+      "total_flos": 4.6049855774374625e+19,
+      "train_loss": 0.07298430502414703,
+      "train_runtime": 2125.6437,
+      "train_samples_per_second": 60.217,
+      "train_steps_per_second": 1.882
     }
   ],
-  "max_steps": 3950,
-  "num_train_epochs": 50,
-  "total_flos": 4.545011460667043e+19,
+  "max_steps": 4000,
+  "num_train_epochs": 51,
+  "total_flos": 4.6049855774374625e+19,
   "trial_name": null,
   "trial_params": null
 }