End of training

all_results.json

@@ -1,14 +1,14 @@
 {
     "epoch": 100.0,
-    "eval_loss": 0.5885377526283264,
-    "eval_runtime": 12.7017,
+    "eval_loss": 0.5620014667510986,
+    "eval_runtime": 12.4265,
     "eval_samples": 301,
-    "eval_samples_per_second": 23.698,
-    "eval_steps_per_second": 2.992,
-    "eval_wer": 0.5898617511520737,
-    "train_loss": 2.0729129652543503,
-    "train_runtime": 5588.1144,
+    "eval_samples_per_second": 24.223,
+    "eval_steps_per_second": 3.058,
+    "eval_wer": 0.5651445328864684,
+    "train_loss": 2.969521954276345,
+    "train_runtime": 5494.3742,
     "train_samples": 704,
-    "train_samples_per_second": 12.598,
-    "train_steps_per_second": 0.394
+    "train_samples_per_second": 12.813,
+    "train_steps_per_second": 0.4
 }

eval_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 100.0,
-    "eval_loss": 0.5885377526283264,
-    "eval_runtime": 12.7017,
+    "eval_loss": 0.5620014667510986,
+    "eval_runtime": 12.4265,
     "eval_samples": 301,
-    "eval_samples_per_second": 23.698,
-    "eval_steps_per_second": 2.992,
-    "eval_wer": 0.5898617511520737
+    "eval_samples_per_second": 24.223,
+    "eval_steps_per_second": 3.058,
+    "eval_wer": 0.5651445328864684
 }

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8e7a4394b288e2177da5753bbb6429436ebf67bd1aa05d5ad56cd386dad6d64a
+oid sha256:5f0eb7adda32a4da8134eb8bc761eb80fae0ded02c211509c87f3ab8a96a2f79
 size 1262108145

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 100.0,
-    "train_loss": 2.0729129652543503,
-    "train_runtime": 5588.1144,
+    "train_loss": 2.969521954276345,
+    "train_runtime": 5494.3742,
     "train_samples": 704,
-    "train_samples_per_second": 12.598,
-    "train_steps_per_second": 0.394
+    "train_samples_per_second": 12.813,
+    "train_steps_per_second": 0.4
 }

trainer_state.json

@@ -9,117 +9,117 @@
   "log_history": [
     {
       "epoch": 13.64,
-      "learning_rate": 2.99e-05,
-      "loss": 7.5798,
+      "learning_rate": 1.1212499999999998e-05,
+      "loss": 9.6445,
       "step": 300
     },
     {
       "epoch": 13.64,
-      "eval_loss": 3.4349324703216553,
-      "eval_runtime": 13.2159,
-      "eval_samples_per_second": 22.776,
-      "eval_steps_per_second": 2.875,
+      "eval_loss": 4.396285057067871,
+      "eval_runtime": 12.5672,
+      "eval_samples_per_second": 23.951,
+      "eval_steps_per_second": 3.024,
       "eval_wer": 1.0,
       "step": 300
     },
     {
       "epoch": 27.27,
-      "learning_rate": 5.989999999999999e-05,
-      "loss": 3.1252,
+      "learning_rate": 2.2462499999999997e-05,
+      "loss": 3.6459,
       "step": 600
     },
     {
       "epoch": 27.27,
-      "eval_loss": 3.0706284046173096,
-      "eval_runtime": 13.3085,
-      "eval_samples_per_second": 22.617,
-      "eval_steps_per_second": 2.855,
+      "eval_loss": 3.2267072200775146,
+      "eval_runtime": 12.4767,
+      "eval_samples_per_second": 24.125,
+      "eval_steps_per_second": 3.046,
       "eval_wer": 1.0,
       "step": 600
     },
     {
       "epoch": 40.91,
-      "learning_rate": 6.729310344827586e-05,
-      "loss": 2.2546,
+      "learning_rate": 3.37125e-05,
+      "loss": 3.0978,
       "step": 900
     },
     {
       "epoch": 40.91,
-      "eval_loss": 0.8426555395126343,
-      "eval_runtime": 12.9127,
-      "eval_samples_per_second": 23.31,
-      "eval_steps_per_second": 2.943,
-      "eval_wer": 0.7762882279011312,
+      "eval_loss": 3.0927422046661377,
+      "eval_runtime": 12.5507,
+      "eval_samples_per_second": 23.983,
+      "eval_steps_per_second": 3.028,
+      "eval_wer": 1.0,
       "step": 900
     },
     {
       "epoch": 54.55,
-      "learning_rate": 5.177586206896551e-05,
-      "loss": 0.7564,
+      "learning_rate": 4.4962499999999995e-05,
+      "loss": 2.8357,
       "step": 1200
     },
     {
       "epoch": 54.55,
-      "eval_loss": 0.6129118800163269,
-      "eval_runtime": 13.2137,
-      "eval_samples_per_second": 22.779,
-      "eval_steps_per_second": 2.876,
-      "eval_wer": 0.637620444072057,
+      "eval_loss": 2.146217107772827,
+      "eval_runtime": 12.4871,
+      "eval_samples_per_second": 24.105,
+      "eval_steps_per_second": 3.043,
+      "eval_wer": 1.002932551319648,
       "step": 1200
     },
     {
       "epoch": 68.18,
-      "learning_rate": 3.625862068965517e-05,
-      "loss": 0.5239,
+      "learning_rate": 5.62125e-05,
+      "loss": 1.2723,
       "step": 1500
     },
     {
       "epoch": 68.18,
-      "eval_loss": 0.5769144892692566,
-      "eval_runtime": 12.822,
-      "eval_samples_per_second": 23.475,
-      "eval_steps_per_second": 2.964,
-      "eval_wer": 0.6036866359447005,
+      "eval_loss": 0.6747255325317383,
+      "eval_runtime": 12.3256,
+      "eval_samples_per_second": 24.421,
+      "eval_steps_per_second": 3.083,
+      "eval_wer": 0.6996229576874738,
       "step": 1500
     },
     {
       "epoch": 81.82,
-      "learning_rate": 2.074137931034483e-05,
-      "loss": 0.438,
+      "learning_rate": 6.746249999999999e-05,
+      "loss": 0.6528,
       "step": 1800
     },
     {
       "epoch": 81.82,
-      "eval_loss": 0.5937696099281311,
-      "eval_runtime": 13.4551,
-      "eval_samples_per_second": 22.371,
-      "eval_steps_per_second": 2.824,
-      "eval_wer": 0.5915374947633012,
+      "eval_loss": 0.5928319096565247,
+      "eval_runtime": 12.4668,
+      "eval_samples_per_second": 24.144,
+      "eval_steps_per_second": 3.048,
+      "eval_wer": 0.6422287390029325,
       "step": 1800
     },
     {
       "epoch": 95.45,
-      "learning_rate": 5.224137931034482e-06,
-      "loss": 0.3945,
+      "learning_rate": 3.7875e-05,
+      "loss": 0.4905,
       "step": 2100
     },
     {
       "epoch": 95.45,
-      "eval_loss": 0.5868746042251587,
-      "eval_runtime": 13.1379,
-      "eval_samples_per_second": 22.911,
-      "eval_steps_per_second": 2.892,
-      "eval_wer": 0.5860913280268119,
+      "eval_loss": 0.5586517453193665,
+      "eval_runtime": 12.2287,
+      "eval_samples_per_second": 24.614,
+      "eval_steps_per_second": 3.107,
+      "eval_wer": 0.5680770842061165,
       "step": 2100
     },
     {
       "epoch": 100.0,
       "step": 2200,
       "total_flos": 1.3980125790314312e+19,
-      "train_loss": 2.0729129652543503,
-      "train_runtime": 5588.1144,
-      "train_samples_per_second": 12.598,
-      "train_steps_per_second": 0.394
+      "train_loss": 2.969521954276345,
+      "train_runtime": 5494.3742,
+      "train_samples_per_second": 12.813,
+      "train_steps_per_second": 0.4
     }
   ],
   "max_steps": 2200,