Training checkpoint

README.md

@@ -1,22 +1,24 @@
 ---
-base_model: openai/whisper-small
 library_name: transformers
+language:
+- en
 license: apache-2.0
-metrics:
-- wer
+base_model: openai/whisper-small
 tags:
 - generated_from_trainer
+metrics:
+- wer
 model-index:
-- name: Whisper-squeezeformer-v7
+- name: Whisper-squeezeformer-N6SQU-
   results: []
 ---
 
 <!-- This model card has been generated automatically according to the information the Trainer had access to. You
 should probably proofread and complete it, then remove this comment. -->
 
-# Whisper-squeezeformer-v7
+# Whisper-squeezeformer-N6SQU-
 
-This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the None dataset.
+This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the LibriSpeech dataset.
 It achieves the following results on the evaluation set:
 - Loss: 0.1579
 - Wer: 5.4340

trainer_state.json

@@ -0,0 +1,330 @@
+{
+  "best_metric": 5.066950699939135,
+  "best_model_checkpoint": "../Whisper-squeezeformer-v7\\checkpoint-40000",
+  "epoch": 18.0,
+  "eval_steps": 2500,
+  "global_step": 45000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 1.0,
+      "grad_norm": 10.708171844482422,
+      "learning_rate": 9.976e-06,
+      "loss": 4.7861,
+      "step": 2500
+    },
+    {
+      "epoch": 1.0,
+      "eval_loss": 3.87461519241333,
+      "eval_runtime": 644.5089,
+      "eval_samples_per_second": 4.065,
+      "eval_steps_per_second": 0.509,
+      "eval_wer": 131.59996956786367,
+      "step": 2500
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 5.609691143035889,
+      "learning_rate": 9.413176470588235e-06,
+      "loss": 2.7936,
+      "step": 5000
+    },
+    {
+      "epoch": 2.0,
+      "eval_loss": 0.2787948250770569,
+      "eval_runtime": 654.0833,
+      "eval_samples_per_second": 4.006,
+      "eval_steps_per_second": 0.501,
+      "eval_wer": 14.139531345100426,
+      "step": 5000
+    },
+    {
+      "epoch": 3.0,
+      "grad_norm": 3.4609644412994385,
+      "learning_rate": 8.82494117647059e-06,
+      "loss": 0.1896,
+      "step": 7500
+    },
+    {
+      "epoch": 3.0,
+      "eval_loss": 0.20549724996089935,
+      "eval_runtime": 1218.2705,
+      "eval_samples_per_second": 2.151,
+      "eval_steps_per_second": 0.269,
+      "eval_wer": 10.453438831405965,
+      "step": 7500
+    },
+    {
+      "epoch": 4.0,
+      "grad_norm": 3.3690125942230225,
+      "learning_rate": 8.236705882352943e-06,
+      "loss": 0.1024,
+      "step": 10000
+    },
+    {
+      "epoch": 4.0,
+      "eval_loss": 0.19734220206737518,
+      "eval_runtime": 1187.8954,
+      "eval_samples_per_second": 2.206,
+      "eval_steps_per_second": 0.276,
+      "eval_wer": 8.690276932440657,
+      "step": 10000
+    },
+    {
+      "epoch": 5.0,
+      "grad_norm": 2.1822853088378906,
+      "learning_rate": 7.648470588235296e-06,
+      "loss": 0.0602,
+      "step": 12500
+    },
+    {
+      "epoch": 5.0,
+      "eval_loss": 0.19494721293449402,
+      "eval_runtime": 1272.2089,
+      "eval_samples_per_second": 2.059,
+      "eval_steps_per_second": 0.258,
+      "eval_wer": 8.94704808277541,
+      "step": 12500
+    },
+    {
+      "epoch": 6.0,
+      "grad_norm": 3.247063636779785,
+      "learning_rate": 7.060470588235294e-06,
+      "loss": 0.1756,
+      "step": 15000
+    },
+    {
+      "epoch": 6.0,
+      "eval_loss": 0.15839330852031708,
+      "eval_runtime": 1226.4236,
+      "eval_samples_per_second": 2.136,
+      "eval_steps_per_second": 0.267,
+      "eval_wer": 7.5034236153377964,
+      "step": 15000
+    },
+    {
+      "epoch": 7.0,
+      "grad_norm": 4.580276966094971,
+      "learning_rate": 6.472235294117648e-06,
+      "loss": 0.1005,
+      "step": 17500
+    },
+    {
+      "epoch": 7.0,
+      "eval_loss": 0.1525421440601349,
+      "eval_runtime": 1332.1465,
+      "eval_samples_per_second": 1.967,
+      "eval_steps_per_second": 0.246,
+      "eval_wer": 6.704580036518563,
+      "step": 17500
+    },
+    {
+      "epoch": 8.0,
+      "grad_norm": 2.37420916557312,
+      "learning_rate": 5.884235294117648e-06,
+      "loss": 0.0619,
+      "step": 20000
+    },
+    {
+      "epoch": 8.0,
+      "eval_loss": 0.15489771962165833,
+      "eval_runtime": 1391.991,
+      "eval_samples_per_second": 1.882,
+      "eval_steps_per_second": 0.236,
+      "eval_wer": 6.771150334753499,
+      "step": 20000
+    },
+    {
+      "epoch": 9.0,
+      "grad_norm": 4.263836860656738,
+      "learning_rate": 5.2962352941176475e-06,
+      "loss": 0.2214,
+      "step": 22500
+    },
+    {
+      "epoch": 9.0,
+      "eval_loss": 0.14554446935653687,
+      "eval_runtime": 1191.0225,
+      "eval_samples_per_second": 2.2,
+      "eval_steps_per_second": 0.275,
+      "eval_wer": 6.318472306755934,
+      "step": 22500
+    },
+    {
+      "epoch": 10.0,
+      "grad_norm": 3.3093366622924805,
+      "learning_rate": 4.708470588235294e-06,
+      "loss": 0.1398,
+      "step": 25000
+    },
+    {
+      "epoch": 10.0,
+      "eval_loss": 0.14445646107196808,
+      "eval_runtime": 1180.1189,
+      "eval_samples_per_second": 2.22,
+      "eval_steps_per_second": 0.278,
+      "eval_wer": 6.162507608034084,
+      "step": 25000
+    },
+    {
+      "epoch": 11.0,
+      "grad_norm": 3.834306001663208,
+      "learning_rate": 4.120705882352942e-06,
+      "loss": 0.1967,
+      "step": 27500
+    },
+    {
+      "epoch": 11.0,
+      "eval_loss": 0.13017532229423523,
+      "eval_runtime": 1349.0935,
+      "eval_samples_per_second": 1.942,
+      "eval_steps_per_second": 0.243,
+      "eval_wer": 5.517726719415704,
+      "step": 27500
+    },
+    {
+      "epoch": 12.0,
+      "grad_norm": 3.707526206970215,
+      "learning_rate": 3.5327058823529413e-06,
+      "loss": 0.1329,
+      "step": 30000
+    },
+    {
+      "epoch": 12.0,
+      "eval_loss": 0.12983843684196472,
+      "eval_runtime": 1194.9126,
+      "eval_samples_per_second": 2.193,
+      "eval_steps_per_second": 0.274,
+      "eval_wer": 5.548158855751674,
+      "step": 30000
+    },
+    {
+      "epoch": 13.0,
+      "grad_norm": 3.3835012912750244,
+      "learning_rate": 2.944705882352941e-06,
+      "loss": 0.1778,
+      "step": 32500
+    },
+    {
+      "epoch": 13.0,
+      "eval_loss": 0.12274244427680969,
+      "eval_runtime": 1372.9808,
+      "eval_samples_per_second": 1.908,
+      "eval_steps_per_second": 0.239,
+      "eval_wer": 5.3237218502738894,
+      "step": 32500
+    },
+    {
+      "epoch": 14.0,
+      "grad_norm": 3.3048934936523438,
+      "learning_rate": 2.356705882352941e-06,
+      "loss": 0.1281,
+      "step": 35000
+    },
+    {
+      "epoch": 14.0,
+      "eval_loss": 0.12345948815345764,
+      "eval_runtime": 1191.4459,
+      "eval_samples_per_second": 2.199,
+      "eval_steps_per_second": 0.275,
+      "eval_wer": 5.179169202678028,
+      "step": 35000
+    },
+    {
+      "epoch": 15.0,
+      "grad_norm": 7.545591831207275,
+      "learning_rate": 1.7689411764705882e-06,
+      "loss": 0.3553,
+      "step": 37500
+    },
+    {
+      "epoch": 15.0,
+      "eval_loss": 0.12376432865858078,
+      "eval_runtime": 1257.7405,
+      "eval_samples_per_second": 2.083,
+      "eval_steps_per_second": 0.261,
+      "eval_wer": 5.236229458307974,
+      "step": 37500
+    },
+    {
+      "epoch": 16.0,
+      "grad_norm": 6.769348621368408,
+      "learning_rate": 1.1814117647058825e-06,
+      "loss": 0.2678,
+      "step": 40000
+    },
+    {
+      "epoch": 16.0,
+      "eval_loss": 0.121117502450943,
+      "eval_runtime": 1262.3643,
+      "eval_samples_per_second": 2.075,
+      "eval_steps_per_second": 0.26,
+      "eval_wer": 5.066950699939135,
+      "step": 40000
+    },
+    {
+      "epoch": 17.0,
+      "grad_norm": 15.616162300109863,
+      "learning_rate": 5.945882352941177e-07,
+      "loss": 0.8916,
+      "step": 42500
+    },
+    {
+      "epoch": 17.0,
+      "eval_loss": 0.14307822287082672,
+      "eval_runtime": 1570.3407,
+      "eval_samples_per_second": 1.668,
+      "eval_steps_per_second": 0.209,
+      "eval_wer": 5.3617620206938525,
+      "step": 42500
+    },
+    {
+      "epoch": 18.0,
+      "grad_norm": 9.218775749206543,
+      "learning_rate": 6.3529411764705896e-09,
+      "loss": 0.8058,
+      "step": 45000
+    },
+    {
+      "epoch": 18.0,
+      "eval_loss": 0.15793535113334656,
+      "eval_runtime": 1362.4738,
+      "eval_samples_per_second": 1.923,
+      "eval_steps_per_second": 0.241,
+      "eval_wer": 5.434038344491783,
+      "step": 45000
+    },
+    {
+      "epoch": 18.0,
+      "step": 45000,
+      "total_flos": 3.1197915376459776e+20,
+      "train_loss": 0.09429476996527777,
+      "train_runtime": 46691.7188,
+      "train_samples_per_second": 19.275,
+      "train_steps_per_second": 0.964
+    }
+  ],
+  "logging_steps": 2500,
+  "max_steps": 45000,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 18,
+  "save_steps": 2500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 3.1197915376459776e+20,
+  "train_batch_size": 20,
+  "trial_name": null,
+  "trial_params": null
+}