End of training

README.md

@@ -4,11 +4,23 @@ license: apache-2.0
 base_model: openai/whisper-medium
 tags:
 - generated_from_trainer
+datasets:
+- toigen
 metrics:
 - wer
 model-index:
 - name: whisper-medium-toigen-balanced-model
-  results: []
+  results:
+  - task:
+      name: Automatic Speech Recognition
+      type: automatic-speech-recognition
+    dataset:
+      name: toigen
+      type: toigen
+    metrics:
+    - name: Wer
+      type: wer
+      value: 0.48498635122838946
 ---
 
 <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -16,10 +28,10 @@ should probably proofread and complete it, then remove this comment. -->
 
 # whisper-medium-toigen-balanced-model
 
-This model is a fine-tuned version of [openai/whisper-medium](https://huggingface.co/openai/whisper-medium) on an unknown dataset.
+This model is a fine-tuned version of [openai/whisper-medium](https://huggingface.co/openai/whisper-medium) on the toigen dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.8748
-- Wer: 0.4604
+- Loss: 0.7784
+- Wer: 0.4850
 
 ## Model description
 

all_results.json

@@ -0,0 +1,15 @@
+{
+    "epoch": 8.766884531590414,
+    "eval_loss": 0.7783711552619934,
+    "eval_runtime": 130.1477,
+    "eval_samples": 215,
+    "eval_samples_per_second": 1.652,
+    "eval_steps_per_second": 0.83,
+    "eval_wer": 0.48498635122838946,
+    "total_flos": 8.19749861720064e+18,
+    "train_loss": 2.2032353465557097,
+    "train_runtime": 2667.1173,
+    "train_samples": 918,
+    "train_samples_per_second": 14.997,
+    "train_steps_per_second": 1.875
+}

eval_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 8.766884531590414,
+    "eval_loss": 0.7783711552619934,
+    "eval_runtime": 130.1477,
+    "eval_samples": 215,
+    "eval_samples_per_second": 1.652,
+    "eval_steps_per_second": 0.83,
+    "eval_wer": 0.48498635122838946
+}

runs/Jan05_15-12-44_srvrocgpu011.uct.ac.za/events.out.tfevents.1736085866.srvrocgpu011.uct.ac.za

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0a0042f46ac94c50db88f8cbbb16753c1ee493389162e7f7c13356409130c612
+size 40

train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 8.766884531590414,
+    "total_flos": 8.19749861720064e+18,
+    "train_loss": 2.2032353465557097,
+    "train_runtime": 2667.1173,
+    "train_samples": 918,
+    "train_samples_per_second": 14.997,
+    "train_steps_per_second": 1.875
+}

trainer_state.json

@@ -0,0 +1,376 @@
+{
+  "best_metric": 0.7783711552619934,
+  "best_model_checkpoint": "/scratch/skscla001/speech/results/whisper-medium-toigen-balanced-model/checkpoint-400",
+  "epoch": 8.766884531590414,
+  "eval_steps": 200,
+  "global_step": 1000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.2178649237472767,
+      "grad_norm": 130.9698028564453,
+      "learning_rate": 4.0000000000000003e-07,
+      "loss": 13.9102,
+      "step": 25
+    },
+    {
+      "epoch": 0.4357298474945534,
+      "grad_norm": 102.07112884521484,
+      "learning_rate": 9.000000000000001e-07,
+      "loss": 11.1467,
+      "step": 50
+    },
+    {
+      "epoch": 0.6535947712418301,
+      "grad_norm": 79.69112396240234,
+      "learning_rate": 1.4000000000000001e-06,
+      "loss": 8.6162,
+      "step": 75
+    },
+    {
+      "epoch": 0.8714596949891068,
+      "grad_norm": 67.27130126953125,
+      "learning_rate": 1.9000000000000002e-06,
+      "loss": 6.3881,
+      "step": 100
+    },
+    {
+      "epoch": 1.0958605664488017,
+      "grad_norm": 61.87217712402344,
+      "learning_rate": 2.4000000000000003e-06,
+      "loss": 5.4679,
+      "step": 125
+    },
+    {
+      "epoch": 1.3137254901960784,
+      "grad_norm": 70.62369537353516,
+      "learning_rate": 2.9e-06,
+      "loss": 4.4849,
+      "step": 150
+    },
+    {
+      "epoch": 1.5315904139433552,
+      "grad_norm": 72.37535858154297,
+      "learning_rate": 3.4000000000000005e-06,
+      "loss": 4.2698,
+      "step": 175
+    },
+    {
+      "epoch": 1.7494553376906317,
+      "grad_norm": 78.36324310302734,
+      "learning_rate": 3.900000000000001e-06,
+      "loss": 3.415,
+      "step": 200
+    },
+    {
+      "epoch": 1.7494553376906317,
+      "eval_loss": 0.8852776885032654,
+      "eval_runtime": 133.304,
+      "eval_samples_per_second": 1.613,
+      "eval_steps_per_second": 0.81,
+      "eval_wer": 0.6050955414012739,
+      "step": 200
+    },
+    {
+      "epoch": 1.9673202614379086,
+      "grad_norm": 52.00922775268555,
+      "learning_rate": 4.4e-06,
+      "loss": 3.5546,
+      "step": 225
+    },
+    {
+      "epoch": 2.1917211328976034,
+      "grad_norm": 59.29832077026367,
+      "learning_rate": 4.9000000000000005e-06,
+      "loss": 2.9609,
+      "step": 250
+    },
+    {
+      "epoch": 2.4095860566448803,
+      "grad_norm": 43.57518768310547,
+      "learning_rate": 5.400000000000001e-06,
+      "loss": 2.3833,
+      "step": 275
+    },
+    {
+      "epoch": 2.627450980392157,
+      "grad_norm": 54.65665054321289,
+      "learning_rate": 5.9e-06,
+      "loss": 2.3348,
+      "step": 300
+    },
+    {
+      "epoch": 2.845315904139434,
+      "grad_norm": 68.74827575683594,
+      "learning_rate": 6.4000000000000006e-06,
+      "loss": 2.5705,
+      "step": 325
+    },
+    {
+      "epoch": 3.0697167755991286,
+      "grad_norm": 34.880619049072266,
+      "learning_rate": 6.9e-06,
+      "loss": 2.0538,
+      "step": 350
+    },
+    {
+      "epoch": 3.287581699346405,
+      "grad_norm": 47.74470901489258,
+      "learning_rate": 7.4e-06,
+      "loss": 1.3222,
+      "step": 375
+    },
+    {
+      "epoch": 3.505446623093682,
+      "grad_norm": 53.504234313964844,
+      "learning_rate": 7.9e-06,
+      "loss": 1.4267,
+      "step": 400
+    },
+    {
+      "epoch": 3.505446623093682,
+      "eval_loss": 0.7783711552619934,
+      "eval_runtime": 133.3648,
+      "eval_samples_per_second": 1.612,
+      "eval_steps_per_second": 0.81,
+      "eval_wer": 0.48498635122838946,
+      "step": 400
+    },
+    {
+      "epoch": 3.7233115468409586,
+      "grad_norm": 36.2320671081543,
+      "learning_rate": 8.400000000000001e-06,
+      "loss": 1.561,
+      "step": 425
+    },
+    {
+      "epoch": 3.9411764705882355,
+      "grad_norm": 37.68352508544922,
+      "learning_rate": 8.900000000000001e-06,
+      "loss": 1.4199,
+      "step": 450
+    },
+    {
+      "epoch": 4.16557734204793,
+      "grad_norm": 30.85091209411621,
+      "learning_rate": 9.4e-06,
+      "loss": 1.078,
+      "step": 475
+    },
+    {
+      "epoch": 4.383442265795207,
+      "grad_norm": 31.413185119628906,
+      "learning_rate": 9.9e-06,
+      "loss": 0.7431,
+      "step": 500
+    },
+    {
+      "epoch": 4.601307189542483,
+      "grad_norm": 29.73971939086914,
+      "learning_rate": 9.955555555555556e-06,
+      "loss": 0.8453,
+      "step": 525
+    },
+    {
+      "epoch": 4.819172113289761,
+      "grad_norm": 32.07612609863281,
+      "learning_rate": 9.9e-06,
+      "loss": 0.8495,
+      "step": 550
+    },
+    {
+      "epoch": 5.0435729847494555,
+      "grad_norm": 35.00574493408203,
+      "learning_rate": 9.844444444444446e-06,
+      "loss": 0.8732,
+      "step": 575
+    },
+    {
+      "epoch": 5.261437908496732,
+      "grad_norm": 24.1536922454834,
+      "learning_rate": 9.78888888888889e-06,
+      "loss": 0.4437,
+      "step": 600
+    },
+    {
+      "epoch": 5.261437908496732,
+      "eval_loss": 0.783809244632721,
+      "eval_runtime": 142.8442,
+      "eval_samples_per_second": 1.505,
+      "eval_steps_per_second": 0.756,
+      "eval_wer": 0.5104640582347588,
+      "step": 600
+    },
+    {
+      "epoch": 5.4793028322440085,
+      "grad_norm": 24.725513458251953,
+      "learning_rate": 9.733333333333334e-06,
+      "loss": 0.4725,
+      "step": 625
+    },
+    {
+      "epoch": 5.697167755991286,
+      "grad_norm": 17.23900604248047,
+      "learning_rate": 9.677777777777778e-06,
+      "loss": 0.5129,
+      "step": 650
+    },
+    {
+      "epoch": 5.915032679738562,
+      "grad_norm": 21.22838592529297,
+      "learning_rate": 9.622222222222222e-06,
+      "loss": 0.4442,
+      "step": 675
+    },
+    {
+      "epoch": 6.139433551198257,
+      "grad_norm": 20.201953887939453,
+      "learning_rate": 9.566666666666668e-06,
+      "loss": 0.3515,
+      "step": 700
+    },
+    {
+      "epoch": 6.357298474945534,
+      "grad_norm": 28.23876190185547,
+      "learning_rate": 9.511111111111112e-06,
+      "loss": 0.2276,
+      "step": 725
+    },
+    {
+      "epoch": 6.57516339869281,
+      "grad_norm": 35.3404541015625,
+      "learning_rate": 9.455555555555557e-06,
+      "loss": 0.2374,
+      "step": 750
+    },
+    {
+      "epoch": 6.793028322440087,
+      "grad_norm": 21.424623489379883,
+      "learning_rate": 9.4e-06,
+      "loss": 0.2962,
+      "step": 775
+    },
+    {
+      "epoch": 7.017429193899782,
+      "grad_norm": 13.183309555053711,
+      "learning_rate": 9.344444444444446e-06,
+      "loss": 0.2779,
+      "step": 800
+    },
+    {
+      "epoch": 7.017429193899782,
+      "eval_loss": 0.8308559060096741,
+      "eval_runtime": 133.876,
+      "eval_samples_per_second": 1.606,
+      "eval_steps_per_second": 0.807,
+      "eval_wer": 0.4572338489535942,
+      "step": 800
+    },
+    {
+      "epoch": 7.235294117647059,
+      "grad_norm": 13.984848022460938,
+      "learning_rate": 9.28888888888889e-06,
+      "loss": 0.1554,
+      "step": 825
+    },
+    {
+      "epoch": 7.453159041394335,
+      "grad_norm": 18.76070785522461,
+      "learning_rate": 9.233333333333334e-06,
+      "loss": 0.1674,
+      "step": 850
+    },
+    {
+      "epoch": 7.671023965141612,
+      "grad_norm": 25.312210083007812,
+      "learning_rate": 9.17777777777778e-06,
+      "loss": 0.1438,
+      "step": 875
+    },
+    {
+      "epoch": 7.888888888888889,
+      "grad_norm": 19.22443962097168,
+      "learning_rate": 9.122222222222223e-06,
+      "loss": 0.1863,
+      "step": 900
+    },
+    {
+      "epoch": 8.113289760348584,
+      "grad_norm": 7.898106575012207,
+      "learning_rate": 9.066666666666667e-06,
+      "loss": 0.1604,
+      "step": 925
+    },
+    {
+      "epoch": 8.33115468409586,
+      "grad_norm": 19.490093231201172,
+      "learning_rate": 9.011111111111111e-06,
+      "loss": 0.1181,
+      "step": 950
+    },
+    {
+      "epoch": 8.549019607843137,
+      "grad_norm": 10.587133407592773,
+      "learning_rate": 8.955555555555555e-06,
+      "loss": 0.1209,
+      "step": 975
+    },
+    {
+      "epoch": 8.766884531590414,
+      "grad_norm": 22.309471130371094,
+      "learning_rate": 8.900000000000001e-06,
+      "loss": 0.1379,
+      "step": 1000
+    },
+    {
+      "epoch": 8.766884531590414,
+      "eval_loss": 0.8748428821563721,
+      "eval_runtime": 132.6338,
+      "eval_samples_per_second": 1.621,
+      "eval_steps_per_second": 0.814,
+      "eval_wer": 0.4604185623293904,
+      "step": 1000
+    },
+    {
+      "epoch": 8.766884531590414,
+      "step": 1000,
+      "total_flos": 8.19749861720064e+18,
+      "train_loss": 2.2032353465557097,
+      "train_runtime": 2667.1173,
+      "train_samples_per_second": 14.997,
+      "train_steps_per_second": 1.875
+    }
+  ],
+  "logging_steps": 25,
+  "max_steps": 5000,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 44,
+  "save_steps": 200,
+  "stateful_callbacks": {
+    "EarlyStoppingCallback": {
+      "args": {
+        "early_stopping_patience": 3,
+        "early_stopping_threshold": 0.0
+      },
+      "attributes": {
+        "early_stopping_patience_counter": 3
+      }
+    },
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 8.19749861720064e+18,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}