Training checkpoint

README.md

@@ -1,19 +1,23 @@
 ---
+language:
+- zh
 license: apache-2.0
-base_model: jun-han/whisper-small-zh_chinese
+base_model: openai/whisper-small
 tags:
 - generated_from_trainer
+datasets:
+- Voice_Data_Collection
 model-index:
-- name: whisper-small-zh_chinese
+- name: Whisper-Keep-train
   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-small-zh_chinese
+# Whisper-Keep-train
 
-This model is a fine-tuned version of [jun-han/whisper-small-zh_chinese](https://huggingface.co/jun-han/whisper-small-zh_chinese) on the None dataset.
+This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the Voice_Data_Collection dataset.
 It achieves the following results on the evaluation set:
 - Loss: 4.0346
 - Cer: 736.0397

trainer_state.json

@@ -1,73 +1,119 @@
 {
-  "best_metric": 96.97789806044203,
-  "best_model_checkpoint": "/kaggle/working/whisper-small-zh_chinese/checkpoint-50",
-  "epoch": 1.6129032258064515,
+  "best_metric": 265.17816869643667,
+  "best_model_checkpoint": "/kaggle/working/whisper-small-zh_chinese/checkpoint-100",
+  "epoch": 3.225806451612903,
   "eval_steps": 50,
-  "global_step": 100,
+  "global_step": 200,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 0.4032258064516129,
-      "grad_norm": 43.397666931152344,
-      "learning_rate": 4.000000000000001e-06,
-      "loss": 11.8023,
+      "grad_norm": 21.932130813598633,
+      "learning_rate": 4.600000000000001e-06,
+      "loss": 3.5703,
       "step": 25
     },
     {
       "epoch": 0.8064516129032258,
-      "grad_norm": 18.771446228027344,
-      "learning_rate": 9e-06,
-      "loss": 6.4546,
+      "grad_norm": 21.00741958618164,
+      "learning_rate": 9.600000000000001e-06,
+      "loss": 3.4342,
       "step": 50
     },
     {
       "epoch": 0.8064516129032258,
-      "eval_cer": 96.97789806044203,
-      "eval_loss": 5.9649481773376465,
-      "eval_runtime": 49.0271,
-      "eval_samples_per_second": 2.04,
-      "eval_steps_per_second": 0.265,
+      "eval_cer": 275.1465944970681,
+      "eval_loss": 4.133521556854248,
+      "eval_runtime": 108.8146,
+      "eval_samples_per_second": 0.919,
+      "eval_steps_per_second": 0.119,
       "step": 50
     },
     {
       "epoch": 1.2096774193548387,
-      "grad_norm": 19.21599006652832,
-      "learning_rate": 6e-06,
-      "loss": 5.0872,
+      "grad_norm": 21.135292053222656,
+      "learning_rate": 8.533333333333335e-06,
+      "loss": 3.3497,
       "step": 75
     },
     {
       "epoch": 1.6129032258064515,
-      "grad_norm": 18.573108673095703,
-      "learning_rate": 1.0000000000000002e-06,
-      "loss": 4.3235,
+      "grad_norm": 21.045244216918945,
+      "learning_rate": 6.866666666666667e-06,
+      "loss": 3.0912,
       "step": 100
     },
     {
       "epoch": 1.6129032258064515,
-      "eval_cer": 283.2656743346865,
-      "eval_loss": 4.595037460327148,
-      "eval_runtime": 105.134,
-      "eval_samples_per_second": 0.951,
-      "eval_steps_per_second": 0.124,
+      "eval_cer": 265.17816869643667,
+      "eval_loss": 4.106583595275879,
+      "eval_runtime": 97.0216,
+      "eval_samples_per_second": 1.031,
+      "eval_steps_per_second": 0.134,
       "step": 100
     },
     {
-      "epoch": 1.6129032258064515,
-      "step": 100,
-      "total_flos": 6.436739948544e+17,
-      "train_loss": 6.916923751831055,
-      "train_runtime": 1054.0975,
-      "train_samples_per_second": 1.518,
-      "train_steps_per_second": 0.095
+      "epoch": 2.0161290322580645,
+      "grad_norm": 20.702259063720703,
+      "learning_rate": 5.2e-06,
+      "loss": 3.1659,
+      "step": 125
+    },
+    {
+      "epoch": 2.4193548387096775,
+      "grad_norm": 26.67952537536621,
+      "learning_rate": 3.5333333333333335e-06,
+      "loss": 2.7991,
+      "step": 150
+    },
+    {
+      "epoch": 2.4193548387096775,
+      "eval_cer": 318.1326116373478,
+      "eval_loss": 4.060230255126953,
+      "eval_runtime": 102.4275,
+      "eval_samples_per_second": 0.976,
+      "eval_steps_per_second": 0.127,
+      "step": 150
+    },
+    {
+      "epoch": 2.8225806451612905,
+      "grad_norm": 24.625123977661133,
+      "learning_rate": 1.8666666666666669e-06,
+      "loss": 2.8228,
+      "step": 175
+    },
+    {
+      "epoch": 3.225806451612903,
+      "grad_norm": 17.782535552978516,
+      "learning_rate": 2.0000000000000002e-07,
+      "loss": 2.6758,
+      "step": 200
+    },
+    {
+      "epoch": 3.225806451612903,
+      "eval_cer": 736.0396932792061,
+      "eval_loss": 4.034571170806885,
+      "eval_runtime": 123.6186,
+      "eval_samples_per_second": 0.809,
+      "eval_steps_per_second": 0.105,
+      "step": 200
+    },
+    {
+      "epoch": 3.225806451612903,
+      "step": 200,
+      "total_flos": 1.2824839544832e+18,
+      "train_loss": 3.1136483764648437,
+      "train_runtime": 2250.2712,
+      "train_samples_per_second": 1.422,
+      "train_steps_per_second": 0.089
     }
   ],
   "logging_steps": 25,
-  "max_steps": 100,
+  "max_steps": 200,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 2,
+  "num_train_epochs": 4,
   "save_steps": 50,
   "stateful_callbacks": {
     "TrainerControl": {
@@ -81,7 +127,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 6.436739948544e+17,
+  "total_flos": 1.2824839544832e+18,
   "train_batch_size": 16,
   "trial_name": null,
   "trial_params": null