End of training

README.md

@@ -4,11 +4,24 @@ license: apache-2.0
 base_model: openai/whisper-base
 tags:
 - generated_from_trainer
+datasets:
+- razhan/DOLMA-speech
 metrics:
 - wer
 model-index:
 - name: whisper-base-sdh
-  results: []
+  results:
+  - task:
+      name: Automatic Speech Recognition
+      type: automatic-speech-recognition
+    dataset:
+      name: razhan/DOLMA-speech southern_kurdish
+      type: razhan/DOLMA-speech
+      args: southern_kurdish
+    metrics:
+    - name: Wer
+      type: wer
+      value: 0.6882181444006173
 ---
 
 <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -16,7 +29,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # whisper-base-sdh
 
-This model is a fine-tuned version of [openai/whisper-base](https://huggingface.co/openai/whisper-base) on an unknown dataset.
+This model is a fine-tuned version of [openai/whisper-base](https://huggingface.co/openai/whisper-base) on the razhan/DOLMA-speech southern_kurdish dataset.
 It achieves the following results on the evaluation set:
 - Loss: 0.6611
 - Wer: 0.6882

all_results.json

@@ -0,0 +1,16 @@
+{
+    "epoch": 5.0,
+    "eval_cer": 0.21645465736561345,
+    "eval_loss": 0.6611320376396179,
+    "eval_runtime": 51.5526,
+    "eval_samples": 757,
+    "eval_samples_per_second": 14.684,
+    "eval_steps_per_second": 0.058,
+    "eval_wer": 0.6882181444006173,
+    "total_flos": 1.9924990745190072e+18,
+    "train_loss": 1.3853384613990785,
+    "train_runtime": 2046.8169,
+    "train_samples": 5912,
+    "train_samples_per_second": 14.442,
+    "train_steps_per_second": 0.039
+}

eval_results.json

@@ -0,0 +1,10 @@
+{
+    "epoch": 5.0,
+    "eval_cer": 0.21645465736561345,
+    "eval_loss": 0.6611320376396179,
+    "eval_runtime": 51.5526,
+    "eval_samples": 757,
+    "eval_samples_per_second": 14.684,
+    "eval_steps_per_second": 0.058,
+    "eval_wer": 0.6882181444006173
+}

train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 5.0,
+    "total_flos": 1.9924990745190072e+18,
+    "train_loss": 1.3853384613990785,
+    "train_runtime": 2046.8169,
+    "train_samples": 5912,
+    "train_samples_per_second": 14.442,
+    "train_steps_per_second": 0.039
+}

trainer_state.json

@@ -0,0 +1,148 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 5.0,
+  "eval_steps": 500,
+  "global_step": 80,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.625,
+      "grad_norm": 35.632633209228516,
+      "learning_rate": 9.620253164556963e-06,
+      "loss": 3.8394,
+      "step": 10
+    },
+    {
+      "epoch": 1.0,
+      "eval_cer": 0.6162631650936944,
+      "eval_loss": 1.9695916175842285,
+      "eval_runtime": 53.8754,
+      "eval_samples_per_second": 14.051,
+      "eval_steps_per_second": 0.056,
+      "eval_wer": 0.9849082490138913,
+      "step": 16
+    },
+    {
+      "epoch": 1.25,
+      "grad_norm": 14.219266891479492,
+      "learning_rate": 8.354430379746837e-06,
+      "loss": 2.1751,
+      "step": 20
+    },
+    {
+      "epoch": 1.875,
+      "grad_norm": 9.090901374816895,
+      "learning_rate": 7.08860759493671e-06,
+      "loss": 1.3507,
+      "step": 30
+    },
+    {
+      "epoch": 2.0,
+      "eval_cer": 0.41071672821775407,
+      "eval_loss": 1.0905992984771729,
+      "eval_runtime": 54.5857,
+      "eval_samples_per_second": 13.868,
+      "eval_steps_per_second": 0.055,
+      "eval_wer": 0.8955582232893158,
+      "step": 32
+    },
+    {
+      "epoch": 2.5,
+      "grad_norm": 4.849871635437012,
+      "learning_rate": 5.8227848101265824e-06,
+      "loss": 0.9898,
+      "step": 40
+    },
+    {
+      "epoch": 3.0,
+      "eval_cer": 0.2880590890439064,
+      "eval_loss": 0.7982303500175476,
+      "eval_runtime": 54.2612,
+      "eval_samples_per_second": 13.951,
+      "eval_steps_per_second": 0.055,
+      "eval_wer": 0.7871720116618076,
+      "step": 48
+    },
+    {
+      "epoch": 3.125,
+      "grad_norm": 3.600881576538086,
+      "learning_rate": 4.556962025316456e-06,
+      "loss": 0.8102,
+      "step": 50
+    },
+    {
+      "epoch": 3.75,
+      "grad_norm": 3.1939327716827393,
+      "learning_rate": 3.2911392405063294e-06,
+      "loss": 0.6834,
+      "step": 60
+    },
+    {
+      "epoch": 4.0,
+      "eval_cer": 0.22661058678703325,
+      "eval_loss": 0.6941161155700684,
+      "eval_runtime": 51.0421,
+      "eval_samples_per_second": 14.831,
+      "eval_steps_per_second": 0.059,
+      "eval_wer": 0.7153146973074944,
+      "step": 64
+    },
+    {
+      "epoch": 4.375,
+      "grad_norm": 2.299682378768921,
+      "learning_rate": 2.0253164556962026e-06,
+      "loss": 0.636,
+      "step": 70
+    },
+    {
+      "epoch": 5.0,
+      "grad_norm": 2.070575714111328,
+      "learning_rate": 7.59493670886076e-07,
+      "loss": 0.5981,
+      "step": 80
+    },
+    {
+      "epoch": 5.0,
+      "eval_cer": 0.21645465736561345,
+      "eval_loss": 0.6611320376396179,
+      "eval_runtime": 51.3922,
+      "eval_samples_per_second": 14.73,
+      "eval_steps_per_second": 0.058,
+      "eval_wer": 0.6882181444006173,
+      "step": 80
+    },
+    {
+      "epoch": 5.0,
+      "step": 80,
+      "total_flos": 1.9924990745190072e+18,
+      "train_loss": 1.3853384613990785,
+      "train_runtime": 2046.8169,
+      "train_samples_per_second": 14.442,
+      "train_steps_per_second": 0.039
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 80,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 5,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 1.9924990745190072e+18,
+  "train_batch_size": 192,
+  "trial_name": null,
+  "trial_params": null
+}