End of training

all_results.json

@@ -1,17 +1,17 @@
 {
     "epoch": 1.6,
-    "eval_cer": 1.0,
-    "eval_loss": NaN,
+    "eval_cer": 0.3090127193119299,
+    "eval_loss": 1.500978946685791,
     "eval_model_preparation_time": 0.0045,
-    "eval_runtime": 30.9217,
-    "eval_samples": 572,
-    "eval_samples_per_second": 18.498,
-    "eval_steps_per_second": 1.164,
-    "eval_wer": 1.0,
+    "eval_runtime": 157.1756,
+    "eval_samples": 3136,
+    "eval_samples_per_second": 19.952,
+    "eval_steps_per_second": 1.247,
+    "eval_wer": 0.4351051665913261,
     "total_flos": 6.212261523683712e+18,
-    "train_loss": 3.21392811447382,
-    "train_runtime": 2133.1271,
+    "train_loss": 3.1903547011613846,
+    "train_runtime": 2112.9078,
     "train_samples": 20000,
-    "train_samples_per_second": 15.001,
-    "train_steps_per_second": 0.469
+    "train_samples_per_second": 15.145,
+    "train_steps_per_second": 0.473
 }

eval_results.json

@@ -1,10 +1,10 @@
 {
-    "eval_cer": 1.0,
-    "eval_loss": NaN,
-    "eval_model_preparation_time": 0.0045,
-    "eval_runtime": 30.9217,
-    "eval_samples": 572,
-    "eval_samples_per_second": 18.498,
-    "eval_steps_per_second": 1.164,
-    "eval_wer": 1.0
+    "epoch": 1.6,
+    "eval_cer": 0.3090127193119299,
+    "eval_loss": 1.500978946685791,
+    "eval_runtime": 157.1756,
+    "eval_samples": 3136,
+    "eval_samples_per_second": 19.952,
+    "eval_steps_per_second": 1.247,
+    "eval_wer": 0.4351051665913261
 }

indicwav2vec_MUCS_warmup500_s300shuff100_2143522.out

@@ -1304,3 +1304,67 @@ CHECK MODEL PARAMS Wav2Vec2ForCTC(
 
                                                 
 
 
                                                 
                                                   
+Printing predictions for a few samples:
+Sample 1:
+  Reference: लिबर ऑफिस impress में एक प्रस्तुति document बनाना और बुनियादी formatting के इस spoken tutorial में आपका स्वागत है
+######
+
+
+  Prediction: liber ऑfis impes में एक प्रस्तुति document बनाना और बुनियादी formating के इस spoken tutorial में आपका स्वागै
+
+
+
+Sample 2:
+  Reference: इस tutorial में हम impress window के भागों के बारे में सीखेंगे और कैसे स्लाइड इन्सर्ट करें और कॉपी करें फॉन्ट तथा फॉन्ट को फॉर्मेट करना सीखेंगे
+######
+
+
+  Prediction: इस tutorial में हम impres window के भागों के बारे में सीखेंगे और कैसे slide insert करें और copyfornt तथा font को format करना सीखेंगे
+
+
+
+Sample 3:
+  Reference: यहाँ हम अपने ऑपरेटिंग सिस्टम के रूप में gnu/linux और लिबरऑफिस वर्जन 334 का उपयोग कर रहे हैं
+######
+
+
+  Prediction: यहाँ हम अपने operating system के रूप में gnu lिnuक और libr ofic version 334 का उपयोग कर रहे हं
+
+
+
+Sample 4:
+  Reference: चलिए अपनी प्रस्तुति प्रेजैटेशन sample impress open करते हैं जिसे पिछले tutorial में बनाया था
+######
+
+
+  Prediction: चलिए अपनी प्रस्तुति sampl impres open करते हैं जिे
+
+
+
+Sample 5:
+  Reference: चलिए देखते हैं कि screen पर क्या क्या है
+######
+
+
+  Prediction: चलिए देखते हैं कि scren पर क्या क्या है
+
+
+
+last Reference string यह स्क्रिप्ट लता द्वारा अनुवादित है आईआईटी मुंबई की ओर से मैं रवि कुमार अब आपसे विदा लेता हूँहमसे जुड़ने के लिए धन्यवाद
+
+
+last prediction string लता द्वारा अनुवादित है आई आई टी मुmबई की ओर से मैं रवि कुमार अब आपसे विदा लेता हूँ हमसे जुड़ने के लिए धन्यवाद
+{'eval_loss': 1.500978946685791, 'eval_cer': 0.3090127193119299, 'eval_wer': 0.4351051665913261, 'eval_runtime': 162.132, 'eval_samples_per_second': 19.342, 'eval_steps_per_second': 1.209, 'epoch': 1.6}
+{'train_runtime': 2112.9078, 'train_samples_per_second': 15.145, 'train_steps_per_second': 0.473, 'train_loss': 3.1903547011613846, 'epoch': 1.6}
+***** train metrics *****
+  epoch                    =          1.6
+  total_flos               = 5785619396GF
+  train_loss               =       3.1904
+  train_runtime            =   0:35:12.90
+  train_samples            =        20000
+  train_samples_per_second =       15.145
+  train_steps_per_second   =        0.473
+08/22/2024 16:49:25 - INFO - __main__ - *** Evaluate ***
+/scratch/work/palp3/myenv/lib/python3.11/site-packages/transformers/models/wav2vec2/processing_wav2vec2.py:157: UserWarning: `as_target_processor` is deprecated and will be removed in v5 of Transformers. You can process your labels by using the argument `text` of the regular `__call__` method (either in the same call as your audio inputs, or in a separate call.
+  warnings.warn(
+
  0%|          | 0/196 [00:00<?, ?it/s]
  1%|          | 2/196 [00:00<01:00,  3.19it/s]
  2%|▏         | 3/196 [00:01<01:28,  2.18it/s]
  2%|▏         | 4/196 [00:02<01:49,  1.75it/s]
  3%|▎         | 5/196 [00:02<02:03,  1.55it/s]
  3%|▎         | 6/196 [00:03<02:22,  1.33it/s]
  4%|▎         | 7/196 [00:04<02:25,  1.30it/s]
  4%|▍         | 8/196 [00:05<02:35,  1.21it/s]
  5%|▍         | 9/196 [00:07<03:27,  1.11s/it]
  5%|▌         | 10/196 [00:08<03:58,  1.28s/it]
  6%|▌         | 11/196 [00:10<04:31,  1.47s/it]
  6%|▌         | 12/196 [00:12<04:35,  1.50s/it]
  7%|▋         | 13/196 [00:13<03:59,  1.31s/it]
  7%|▋         | 14/196 [00:13<03:22,  1.11s/it]
  8%|▊         | 15/196 [00:14<02:52,  1.05it/s]
  8%|▊         | 16/196 [00:15<02:40,  1.12it/s]
  9%|▊         | 17/196 [00:16<02:39,  1.12it/s]
  9%|▉         | 18/196 [00:17<03:05,  1.04s/it]
 10%|▉         | 19/196 [00:19<03:52,  1.31s/it]
 10%|█         | 20/196 [00:21<04:03,  1.38s/it]
 11%|█         | 21/196 [00:22<04:09,  1.43s/it]
 11%|█         | 22/196 [00:23<03:53,  1.34s/it]
 12%|█▏        | 23/196 [00:24<03:26,  1.19s/it]
 12%|█▏        | 24/196 [00:25<02:46,  1.03it/s]
 13%|█▎        | 25/196 [00:25<02:25,  1.18it/s]
 13%|█▎        | 26/196 [00:26<02:08,  1.32it/s]
 14%|█▍        | 27/196 [00:26<01:59,  1.42it/s]
 14%|█▍        | 28/196 [00:27<01:55,  1.46it/s]
 15%|█▍        | 29/196 [00:28<01:56,  1.44it/s]
 15%|█▌        | 30/196 [00:28<01:53,  1.46it/s]
 16%|█▌        | 31/196 [00:29<01:42,  1.60it/s]
 16%|█▋        | 32/196 [00:29<01:45,  1.55it/s]
 17%|█▋        | 33/196 [00:30<02:01,  1.34it/s]
 17%|█▋        | 34/196 [00:32<02:25,  1.11it/s]
 18%|█▊        | 35/196 [00:33<02:35,  1.03it/s]
 18%|█▊        | 36/196 [00:34<02:51,  1.07s/it]
 19%|█▉        | 37/196 [00:35<02:44,  1.03s/it]
 19%|█▉        | 38/196 [00:36<02:34,  1.02it/s]
 20%|█▉        | 39/196 [00:37<02:22,  1.10it/s]
 20%|██        | 40/196 [00:37<02:14,  1.16it/s]
 21%|██        | 41/196 [00:38<02:02,  1.27it/s]
 21%|██▏       | 42/196 [00:39<01:57,  1.31it/s]
 22%|██▏       | 43/196 [00:39<01:53,  1.34it/s]
 22%|██▏       | 44/196 [00:40<01:48,  1.40it/s]
 23%|██▎       | 45/196 [00:41<01:40,  1.50it/s]
 23%|██▎       | 46/196 [00:41<01:36,  1.56it/s]
 24%|██▍       | 47/196 [00:42<01:35,  1.56it/s]
 24%|██▍       | 48/196 [00:42<01:31,  1.61it/s]
 25%|██▌       | 49/196 [00:43<01:31,  1.61it/s]
 26%|██▌       | 50/196 [00:44<01:29,  1.63it/s]
 26%|██▌       | 51/196 [00:44<01:27,  1.67it/s]
 27%|██▋       | 52/196 [00:45<01:28,  1.63it/s]
 27%|██▋       | 53/196 [00:45<01:28,  1.61it/s]
 28%|██▊       | 54/196 [00:46<01:29,  1.59it/s]
 28%|██▊       | 55/196 [00:47<01:37,  1.44it/s]
 29%|██▊       | 56/196 [00:48<01:45,  1.32it/s]
 29%|██▉       | 57/196 [00:49<01:52,  1.24it/s]
 30%|██▉       | 58/196 [00:50<01:52,  1.22it/s]
 30%|███       | 59/196 [00:50<01:49,  1.25it/s]
 31%|███       | 60/196 [00:51<01:37,  1.39it/s]
 31%|███       | 61/196 [00:52<01:31,  1.47it/s]
 32%|███▏      | 62/196 [00:52<01:33,  1.44it/s]
 32%|███▏      | 63/196 [00:53<01:33,  1.42it/s]
 33%|███▎      | 64/196 [00:54<01:33,  1.42it/s]
 33%|███▎      | 65/196 [00:54<01:31,  1.43it/s]
 34%|███▎      | 66/196 [00:55<01:36,  1.35it/s]
 34%|███▍      | 67/196 [00:56<01:38,  1.31it/s]
 35%|███▍      | 68/196 [00:57<01:50,  1.16it/s]
 35%|███▌      | 69/196 [00:58<01:48,  1.17it/s]
 36%|███▌      | 70/196 [00:59<01:40,  1.25it/s]
 36%|███▌      | 71/196 [00:59<01:36,  1.29it/s]
 37%|███▋      | 72/196 [01:00<01:30,  1.38it/s]
 37%|███▋      | 73/196 [01:00<01:22,  1.49it/s]
 38%|███▊      | 74/196 [01:01<01:17,  1.57it/s]
 38%|███▊      | 75/196 [01:02<01:15,  1.60it/s]
 39%|███▉      | 76/196 [01:02<01:13,  1.64it/s]
 39%|███▉      | 77/196 [01:03<01:16,  1.56it/s]
 40%|███▉      | 78/196 [01:04<01:17,  1.52it/s]
 40%|████      | 79/196 [01:04<01:16,  1.53it/s]
 41%|████      | 80/196 [01:05<01:20,  1.44it/s]
 41%|████▏     | 81/196 [01:06<01:20,  1.42it/s]
 42%|████▏     | 82/196 [01:06<01:19,  1.44it/s]
 42%|████▏     | 83/196 [01:07<01:21,  1.39it/s]
 43%|████▎     | 84/196 [01:08<01:20,  1.38it/s]
 43%|████▎     | 85/196 [01:09<01:20,  1.37it/s]
 44%|████▍     | 86/196 [01:10<01:22,  1.33it/s]
 44%|████▍     | 87/196 [01:10<01:20,  1.35it/s]
 45%|████▍     | 88/196 [01:11<01:22,  1.31it/s]
 45%|████▌     | 89/196 [01:12<01:23,  1.28it/s]
 46%|████▌     | 90/196 [01:13<01:21,  1.30it/s]
 46%|████▋     | 91/196 [01:13<01:16,  1.37it/s]
 47%|████▋     | 92/196 [01:14<01:13,  1.42it/s]
 47%|████▋     | 93/196 [01:15<01:16,  1.34it/s]
 48%|████▊     | 94/196 [01:15<01:15,  1.35it/s]
 48%|████▊     | 95/196 [01:16<01:14,  1.36it/s]
 49%|████▉     | 96/196 [01:17<01:16,  1.30it/s]
 49%|████▉     | 97/196 [01:18<01:13,  1.35it/s]
 50%|█████     | 98/196 [01:19<01:14,  1.32it/s]
 51%|█████     | 99/196 [01:19<01:08,  1.42it/s]
 51%|█████     | 100/196 [01:20<01:00,  1.59it/s]
 52%|█████▏    | 101/196 [01:20<00:58,  1.62it/s]
 52%|█████▏    | 102/196 [01:21<01:02,  1.51it/s]
 53%|█████▎    | 103/196 [01:22<01:07,  1.37it/s]
 53%|█████▎    | 104/196 [01:23<01:15,  1.22it/s]
 54%|█████▎    | 105/196 [01:24<01:15,  1.20it/s]
 54%|█████▍    | 106/196 [01:24<01:13,  1.22it/s]
 55%|█████▍    | 107/196 [01:25<01:08,  1.31it/s]
 55%|█████▌    | 108/196 [01:26<01:00,  1.46it/s]
 56%|█████▌    | 109/196 [01:26<00:56,  1.53it/s]
 56%|█████▌    | 110/196 [01:27<00:55,  1.54it/s]
 57%|█████▋    | 111/196 [01:28<00:55,  1.52it/s]
 57%|█████▋    | 112/196 [01:28<00:57,  1.47it/s]
 58%|█████▊    | 113/196 [01:29<00:56,  1.47it/s]
 58%|█████▊    | 114/196 [01:30<00:53,  1.53it/s]
 59%|█████▊    | 115/196 [01:30<00:51,  1.57it/s]
 59%|█████▉    | 116/196 [01:31<00:50,  1.59it/s]
 60%|█████▉    | 117/196 [01:31<00:46,  1.70it/s]
 60%|██████    | 118/196 [01:32<00:41,  1.87it/s]
 61%|██████    | 119/196 [01:32<00:44,  1.73it/s]
 61%|██████    | 120/196 [01:33<00:45,  1.67it/s]
 62%|██████▏   | 121/196 [01:34<00:45,  1.64it/s]
 62%|██████▏   | 122/196 [01:34<00:46,  1.60it/s]
 63%|██████▎   | 123/196 [01:35<00:44,  1.64it/s]
 63%|██████▎   | 124/196 [01:35<00:45,  1.58it/s]
 64%|██████▍   | 125/196 [01:36<00:45,  1.57it/s]
 64%|██████▍   | 126/196 [01:37<00:50,  1.38it/s]
 65%|██████▍   | 127/196 [01:38<00:48,  1.42it/s]
 65%|██████▌   | 128/196 [01:38<00:46,  1.48it/s]
 66%|██████▌   | 129/196 [01:39<00:44,  1.49it/s]
 66%|██████▋   | 130/196 [01:40<00:45,  1.47it/s]
 67%|██████▋   | 131/196 [01:40<00:43,  1.50it/s]
 67%|██████▋   | 132/196 [01:41<00:40,  1.59it/s]
 68%|██████▊   | 133/196 [01:42<00:39,  1.59it/s]
 68%|██████▊   | 134/196 [01:42<00:40,  1.55it/s]
 69%|██████▉   | 135/196 [01:43<00:38,  1.58it/s]
 69%|██████▉   | 136/196 [01:43<00:38,  1.58it/s]
 70%|██████▉   | 137/196 [01:44<00:37,  1.58it/s]
 70%|███████   | 138/196 [01:45<00:36,  1.59it/s]
 71%|███████   | 139/196 [01:45<00:37,  1.54it/s]
 71%|███████▏  | 140/196 [01:46<00:35,  1.59it/s]
 72%|███████▏  | 141/196 [01:47<00:33,  1.62it/s]
 72%|███████▏  | 142/196 [01:47<00:34,  1.57it/s]
 73%|███████▎  | 143/196 [01:48<00:35,  1.50it/s]
 73%|███████▎  | 144/196 [01:49<00:33,  1.56it/s]
 74%|███████▍  | 145/196 [01:49<00:30,  1.68it/s]
 74%|███████▍  | 146/196 [01:50<00:28,  1.74it/s]
 75%|███████▌  | 147/196 [01:50<00:27,  1.75it/s]
 76%|███████▌  | 148/196 [01:51<00:27,  1.74it/s]
 76%|███████▌  | 149/196 [01:51<00:25,  1.83it/s]
 77%|███████▋  | 150/196 [01:52<00:27,  1.68it/s]
 77%|███████▋  | 151/196 [01:53<00:27,  1.62it/s]
 78%|███████▊  | 152/196 [01:53<00:27,  1.61it/s]
 78%|███████▊  | 153/196 [01:54<00:26,  1.61it/s]
 79%|███████▊  | 154/196 [01:54<00:26,  1.61it/s]
 79%|███████▉  | 155/196 [01:55<00:27,  1.50it/s]
 80%|███████▉  | 156/196 [01:56<00:29,  1.36it/s]
 80%|████████  | 157/196 [01:57<00:29,  1.30it/s]
 81%|████████  | 158/196 [01:57<00:26,  1.44it/s]
 81%|████████  | 159/196 [01:58<00:23,  1.54it/s]
 82%|████████▏ | 160/196 [01:59<00:23,  1.56it/s]
 82%|████████▏ | 161/196 [01:59<00:22,  1.54it/s]
 83%|████████▎ | 162/196 [02:00<00:21,  1.55it/s]
 83%|████████▎ | 163/196 [02:01<00:20,  1.58it/s]
 84%|████████▎ | 164/196 [02:01<00:20,  1.59it/s]
 84%|████████▍ | 165/196 [02:02<00:20,  1.54it/s]
 85%|████████▍ | 166/196 [02:02<00:18,  1.58it/s]
 85%|████████▌ | 167/196 [02:03<00:17,  1.62it/s]
 86%|████████▌ | 168/196 [02:04<00:16,  1.70it/s]
 86%|████████▌ | 169/196 [02:04<00:16,  1.62it/s]
 87%|████████▋ | 170/196 [02:05<00:17,  1.52it/s]
 87%|████████▋ | 171/196 [02:06<00:16,  1.53it/s]
 88%|████████▊ | 172/196 [02:06<00:15,  1.50it/s]
 88%|████████▊ | 173/196 [02:07<00:15,  1.51it/s]
 89%|████████▉ | 174/196 [02:08<00:15,  1.42it/s]
 89%|████████▉ | 175/196 [02:09<00:18,  1.13it/s]
 90%|████████▉ | 176/196 [02:11<00:26,  1.30s/it]
 90%|█████████ | 177/196 [02:13<00:28,  1.48s/it]
 91%|█████████ | 178/196 [02:15<00:30,  1.67s/it]
 91%|█████████▏| 179/196 [02:17<00:28,  1.66s/it]
 92%|█████████▏| 180/196 [02:18<00:21,  1.37s/it]
 92%|█████████▏| 181/196 [02:18<00:17,  1.15s/it]
 93%|█████████▎| 182/196 [02:19<00:14,  1.00s/it]
 93%|█████████▎| 183/196 [02:20<00:12,  1.01it/s]
 94%|█████████▍| 184/196 [02:21<00:10,  1.14it/s]
 94%|█████████▍| 185/196 [02:21<00:09,  1.18it/s]
 95%|█████████▍| 186/196 [02:22<00:08,  1.17it/s]
 95%|█████████▌| 187/196 [02:23<00:07,  1.27it/s]
 96%|█████████▌| 188/196 [02:24<00:05,  1.35it/s]
 96%|█████████▋| 189/196 [02:24<00:05,  1.38it/s]
 97%|█████████▋| 190/196 [02:25<00:04,  1.46it/s]
 97%|█████████▋| 191/196 [02:25<00:03,  1.54it/s]
 98%|█████████▊| 192/196 [02:26<00:02,  1.50it/s]
 98%|█████████▊| 193/196 [02:27<00:02,  1.49it/s]
 99%|█████████▉| 194/196 [02:27<00:01,  1.52it/s]
 99%|█████████▉| 195/196 [02:28<00:00,  1.57it/s]

train_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 1.6,
     "total_flos": 6.212261523683712e+18,
-    "train_loss": 3.21392811447382,
-    "train_runtime": 2133.1271,
+    "train_loss": 3.1903547011613846,
+    "train_runtime": 2112.9078,
     "train_samples": 20000,
-    "train_samples_per_second": 15.001,
-    "train_steps_per_second": 0.469
+    "train_samples_per_second": 15.145,
+    "train_steps_per_second": 0.473
 }