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-v3
+- name: Whisper-squeezeformer-NSQU-whisper
   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-v3
+# Whisper-squeezeformer-NSQU-whisper
 
-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: 0.1569
 - Wer: 7.1497

trainer_state.json

@@ -0,0 +1,170 @@
+{
+  "best_metric": 7.149650030432136,
+  "best_model_checkpoint": "./Whisper-squeezeformer-v3\\checkpoint-20000",
+  "epoch": 8.0,
+  "eval_steps": 2500,
+  "global_step": 20000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 1.0,
+      "grad_norm": 10.493926048278809,
+      "learning_rate": 9.972e-06,
+      "loss": 4.8718,
+      "step": 2500
+    },
+    {
+      "epoch": 1.0,
+      "eval_loss": 3.8609094619750977,
+      "eval_runtime": 650.8319,
+      "eval_samples_per_second": 4.026,
+      "eval_steps_per_second": 0.504,
+      "eval_wer": 111.85902312842362,
+      "step": 2500
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 6.234286785125732,
+      "learning_rate": 8.575428571428573e-06,
+      "loss": 2.5628,
+      "step": 5000
+    },
+    {
+      "epoch": 2.0,
+      "eval_loss": 0.2978440523147583,
+      "eval_runtime": 624.1365,
+      "eval_samples_per_second": 4.198,
+      "eval_steps_per_second": 0.526,
+      "eval_wer": 15.619293974437007,
+      "step": 5000
+    },
+    {
+      "epoch": 3.0,
+      "grad_norm": 3.2328641414642334,
+      "learning_rate": 7.146857142857143e-06,
+      "loss": 0.1698,
+      "step": 7500
+    },
+    {
+      "epoch": 3.0,
+      "eval_loss": 0.22175125777721405,
+      "eval_runtime": 585.8606,
+      "eval_samples_per_second": 4.472,
+      "eval_steps_per_second": 0.56,
+      "eval_wer": 11.090611685940353,
+      "step": 7500
+    },
+    {
+      "epoch": 4.0,
+      "grad_norm": 2.424901247024536,
+      "learning_rate": 5.718285714285715e-06,
+      "loss": 0.0867,
+      "step": 10000
+    },
+    {
+      "epoch": 4.0,
+      "eval_loss": 0.2011481672525406,
+      "eval_runtime": 614.9021,
+      "eval_samples_per_second": 4.261,
+      "eval_steps_per_second": 0.533,
+      "eval_wer": 10.189059646987218,
+      "step": 10000
+    },
+    {
+      "epoch": 5.0,
+      "grad_norm": 3.77018404006958,
+      "learning_rate": 4.290285714285714e-06,
+      "loss": 0.1697,
+      "step": 12500
+    },
+    {
+      "epoch": 5.0,
+      "eval_loss": 0.1640562266111374,
+      "eval_runtime": 597.6823,
+      "eval_samples_per_second": 4.384,
+      "eval_steps_per_second": 0.549,
+      "eval_wer": 8.985088253195373,
+      "step": 12500
+    },
+    {
+      "epoch": 6.0,
+      "grad_norm": 2.6757357120513916,
+      "learning_rate": 2.861714285714286e-06,
+      "loss": 0.0993,
+      "step": 15000
+    },
+    {
+      "epoch": 6.0,
+      "eval_loss": 0.15530936419963837,
+      "eval_runtime": 585.6998,
+      "eval_samples_per_second": 4.473,
+      "eval_steps_per_second": 0.56,
+      "eval_wer": 7.803940961655509,
+      "step": 15000
+    },
+    {
+      "epoch": 7.0,
+      "grad_norm": 2.374549388885498,
+      "learning_rate": 1.434857142857143e-06,
+      "loss": 0.0651,
+      "step": 17500
+    },
+    {
+      "epoch": 7.0,
+      "eval_loss": 0.1555396020412445,
+      "eval_runtime": 585.2066,
+      "eval_samples_per_second": 4.477,
+      "eval_steps_per_second": 0.56,
+      "eval_wer": 7.244750456482045,
+      "step": 17500
+    },
+    {
+      "epoch": 8.0,
+      "grad_norm": 1.977560043334961,
+      "learning_rate": 6.2857142857142855e-09,
+      "loss": 0.0468,
+      "step": 20000
+    },
+    {
+      "epoch": 8.0,
+      "eval_loss": 0.15691713988780975,
+      "eval_runtime": 580.4295,
+      "eval_samples_per_second": 4.514,
+      "eval_steps_per_second": 0.565,
+      "eval_wer": 7.149650030432136,
+      "step": 20000
+    },
+    {
+      "epoch": 8.0,
+      "step": 20000,
+      "total_flos": 1.38658480128e+20,
+      "train_loss": 0.047600341033935546,
+      "train_runtime": 48327.6236,
+      "train_samples_per_second": 8.277,
+      "train_steps_per_second": 0.414
+    }
+  ],
+  "logging_steps": 2500,
+  "max_steps": 20000,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 8,
+  "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": 1.38658480128e+20,
+  "train_batch_size": 20,
+  "trial_name": null,
+  "trial_params": null
+}