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-male-model
-  results: []
+  results:
+  - task:
+      name: Automatic Speech Recognition
+      type: automatic-speech-recognition
+    dataset:
+      name: toigen
+      type: toigen
+    metrics:
+    - name: Wer
+      type: wer
+      value: 0.5695876288659794
 ---
 
 <!-- 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-male-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: 1.0236
-- Wer: 0.4448
+- Loss: 0.9393
+- Wer: 0.5696
 
 ## Model description
 

all_results.json

@@ -0,0 +1,15 @@
+{
+    "epoch": 9.437054631828978,
+    "eval_loss": 0.9392894506454468,
+    "eval_runtime": 129.543,
+    "eval_samples": 211,
+    "eval_samples_per_second": 1.629,
+    "eval_steps_per_second": 0.818,
+    "eval_wer": 0.5695876288659794,
+    "total_flos": 8.10972659515392e+18,
+    "train_loss": 2.490273738861084,
+    "train_runtime": 2581.6643,
+    "train_samples": 842,
+    "train_samples_per_second": 15.494,
+    "train_steps_per_second": 1.937
+}

eval_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 9.437054631828978,
+    "eval_loss": 0.9392894506454468,
+    "eval_runtime": 129.543,
+    "eval_samples": 211,
+    "eval_samples_per_second": 1.629,
+    "eval_steps_per_second": 0.818,
+    "eval_wer": 0.5695876288659794
+}

runs/Jan05_19-19-34_srvrocgpu011.uct.ac.za/events.out.tfevents.1736100544.srvrocgpu011.uct.ac.za

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d50277442a589f6597fa38388a9f790f456219adda6c18fe8a23471bc5514bd4
+size 40

train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 9.437054631828978,
+    "total_flos": 8.10972659515392e+18,
+    "train_loss": 2.490273738861084,
+    "train_runtime": 2581.6643,
+    "train_samples": 842,
+    "train_samples_per_second": 15.494,
+    "train_steps_per_second": 1.937
+}

trainer_state.json

@@ -0,0 +1,376 @@
+{
+  "best_metric": 0.9392894506454468,
+  "best_model_checkpoint": "/scratch/skscla001/speech/results/whisper-medium-toigen-male-model/checkpoint-400",
+  "epoch": 9.437054631828978,
+  "eval_steps": 200,
+  "global_step": 1000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.2375296912114014,
+      "grad_norm": 150.66383361816406,
+      "learning_rate": 4.0000000000000003e-07,
+      "loss": 14.4659,
+      "step": 25
+    },
+    {
+      "epoch": 0.4750593824228028,
+      "grad_norm": 96.092529296875,
+      "learning_rate": 9.000000000000001e-07,
+      "loss": 11.6879,
+      "step": 50
+    },
+    {
+      "epoch": 0.7125890736342043,
+      "grad_norm": 94.01177215576172,
+      "learning_rate": 1.3800000000000001e-06,
+      "loss": 9.6203,
+      "step": 75
+    },
+    {
+      "epoch": 0.9501187648456056,
+      "grad_norm": 88.87047576904297,
+      "learning_rate": 1.8800000000000002e-06,
+      "loss": 7.6311,
+      "step": 100
+    },
+    {
+      "epoch": 1.180522565320665,
+      "grad_norm": 77.90989685058594,
+      "learning_rate": 2.38e-06,
+      "loss": 6.1323,
+      "step": 125
+    },
+    {
+      "epoch": 1.4180522565320666,
+      "grad_norm": 63.12504959106445,
+      "learning_rate": 2.88e-06,
+      "loss": 5.5927,
+      "step": 150
+    },
+    {
+      "epoch": 1.655581947743468,
+      "grad_norm": 63.86695861816406,
+      "learning_rate": 3.3800000000000007e-06,
+      "loss": 5.0682,
+      "step": 175
+    },
+    {
+      "epoch": 1.8931116389548692,
+      "grad_norm": 75.34732055664062,
+      "learning_rate": 3.88e-06,
+      "loss": 4.343,
+      "step": 200
+    },
+    {
+      "epoch": 1.8931116389548692,
+      "eval_loss": 1.066441535949707,
+      "eval_runtime": 125.6362,
+      "eval_samples_per_second": 1.679,
+      "eval_steps_per_second": 0.844,
+      "eval_wer": 0.606701030927835,
+      "step": 200
+    },
+    {
+      "epoch": 2.1235154394299287,
+      "grad_norm": 74.82162475585938,
+      "learning_rate": 4.38e-06,
+      "loss": 3.9738,
+      "step": 225
+    },
+    {
+      "epoch": 2.36104513064133,
+      "grad_norm": 60.270687103271484,
+      "learning_rate": 4.880000000000001e-06,
+      "loss": 3.4967,
+      "step": 250
+    },
+    {
+      "epoch": 2.598574821852732,
+      "grad_norm": 72.7863998413086,
+      "learning_rate": 5.380000000000001e-06,
+      "loss": 3.4653,
+      "step": 275
+    },
+    {
+      "epoch": 2.836104513064133,
+      "grad_norm": 42.029544830322266,
+      "learning_rate": 5.8800000000000005e-06,
+      "loss": 3.0671,
+      "step": 300
+    },
+    {
+      "epoch": 3.0665083135391926,
+      "grad_norm": 58.11145782470703,
+      "learning_rate": 6.380000000000001e-06,
+      "loss": 2.7671,
+      "step": 325
+    },
+    {
+      "epoch": 3.304038004750594,
+      "grad_norm": 72.86565399169922,
+      "learning_rate": 6.88e-06,
+      "loss": 2.1218,
+      "step": 350
+    },
+    {
+      "epoch": 3.5415676959619953,
+      "grad_norm": 29.798490524291992,
+      "learning_rate": 7.3800000000000005e-06,
+      "loss": 1.8895,
+      "step": 375
+    },
+    {
+      "epoch": 3.7790973871733966,
+      "grad_norm": 62.0257453918457,
+      "learning_rate": 7.88e-06,
+      "loss": 1.9496,
+      "step": 400
+    },
+    {
+      "epoch": 3.7790973871733966,
+      "eval_loss": 0.9392894506454468,
+      "eval_runtime": 129.8196,
+      "eval_samples_per_second": 1.625,
+      "eval_steps_per_second": 0.817,
+      "eval_wer": 0.5695876288659794,
+      "step": 400
+    },
+    {
+      "epoch": 4.009501187648456,
+      "grad_norm": 36.066246032714844,
+      "learning_rate": 8.380000000000001e-06,
+      "loss": 1.9781,
+      "step": 425
+    },
+    {
+      "epoch": 4.247030878859857,
+      "grad_norm": 42.63113021850586,
+      "learning_rate": 8.880000000000001e-06,
+      "loss": 1.0441,
+      "step": 450
+    },
+    {
+      "epoch": 4.484560570071259,
+      "grad_norm": 33.34630584716797,
+      "learning_rate": 9.38e-06,
+      "loss": 1.1374,
+      "step": 475
+    },
+    {
+      "epoch": 4.72209026128266,
+      "grad_norm": 61.29568862915039,
+      "learning_rate": 9.88e-06,
+      "loss": 1.0169,
+      "step": 500
+    },
+    {
+      "epoch": 4.959619952494061,
+      "grad_norm": 49.31779861450195,
+      "learning_rate": 9.957777777777779e-06,
+      "loss": 1.3693,
+      "step": 525
+    },
+    {
+      "epoch": 5.190023752969121,
+      "grad_norm": 35.42589569091797,
+      "learning_rate": 9.902222222222223e-06,
+      "loss": 0.7001,
+      "step": 550
+    },
+    {
+      "epoch": 5.427553444180522,
+      "grad_norm": 22.584501266479492,
+      "learning_rate": 9.846666666666668e-06,
+      "loss": 0.5772,
+      "step": 575
+    },
+    {
+      "epoch": 5.665083135391924,
+      "grad_norm": 21.677404403686523,
+      "learning_rate": 9.791111111111112e-06,
+      "loss": 0.702,
+      "step": 600
+    },
+    {
+      "epoch": 5.665083135391924,
+      "eval_loss": 0.9410276412963867,
+      "eval_runtime": 124.8464,
+      "eval_samples_per_second": 1.69,
+      "eval_steps_per_second": 0.849,
+      "eval_wer": 0.4814432989690722,
+      "step": 600
+    },
+    {
+      "epoch": 5.902612826603326,
+      "grad_norm": 27.595264434814453,
+      "learning_rate": 9.735555555555556e-06,
+      "loss": 0.6341,
+      "step": 625
+    },
+    {
+      "epoch": 6.133016627078385,
+      "grad_norm": 15.151907920837402,
+      "learning_rate": 9.68e-06,
+      "loss": 0.39,
+      "step": 650
+    },
+    {
+      "epoch": 6.370546318289787,
+      "grad_norm": 63.603759765625,
+      "learning_rate": 9.624444444444445e-06,
+      "loss": 0.3234,
+      "step": 675
+    },
+    {
+      "epoch": 6.608076009501188,
+      "grad_norm": 33.50586700439453,
+      "learning_rate": 9.56888888888889e-06,
+      "loss": 0.2813,
+      "step": 700
+    },
+    {
+      "epoch": 6.845605700712589,
+      "grad_norm": 29.78474235534668,
+      "learning_rate": 9.513333333333334e-06,
+      "loss": 0.3366,
+      "step": 725
+    },
+    {
+      "epoch": 7.076009501187649,
+      "grad_norm": 13.300681114196777,
+      "learning_rate": 9.457777777777778e-06,
+      "loss": 0.2659,
+      "step": 750
+    },
+    {
+      "epoch": 7.31353919239905,
+      "grad_norm": 26.387012481689453,
+      "learning_rate": 9.402222222222222e-06,
+      "loss": 0.1854,
+      "step": 775
+    },
+    {
+      "epoch": 7.551068883610451,
+      "grad_norm": 20.7415771484375,
+      "learning_rate": 9.346666666666666e-06,
+      "loss": 0.2108,
+      "step": 800
+    },
+    {
+      "epoch": 7.551068883610451,
+      "eval_loss": 0.9733582735061646,
+      "eval_runtime": 125.0244,
+      "eval_samples_per_second": 1.688,
+      "eval_steps_per_second": 0.848,
+      "eval_wer": 0.4551546391752577,
+      "step": 800
+    },
+    {
+      "epoch": 7.788598574821853,
+      "grad_norm": 16.81197166442871,
+      "learning_rate": 9.291111111111112e-06,
+      "loss": 0.211,
+      "step": 825
+    },
+    {
+      "epoch": 8.019002375296912,
+      "grad_norm": 8.114081382751465,
+      "learning_rate": 9.235555555555556e-06,
+      "loss": 0.2144,
+      "step": 850
+    },
+    {
+      "epoch": 8.256532066508314,
+      "grad_norm": 7.230250835418701,
+      "learning_rate": 9.180000000000002e-06,
+      "loss": 0.1472,
+      "step": 875
+    },
+    {
+      "epoch": 8.494061757719715,
+      "grad_norm": 34.53820037841797,
+      "learning_rate": 9.124444444444444e-06,
+      "loss": 0.1506,
+      "step": 900
+    },
+    {
+      "epoch": 8.731591448931116,
+      "grad_norm": 20.564693450927734,
+      "learning_rate": 9.06888888888889e-06,
+      "loss": 0.1414,
+      "step": 925
+    },
+    {
+      "epoch": 8.969121140142517,
+      "grad_norm": 24.549924850463867,
+      "learning_rate": 9.013333333333334e-06,
+      "loss": 0.1398,
+      "step": 950
+    },
+    {
+      "epoch": 9.199524940617577,
+      "grad_norm": 9.88586139678955,
+      "learning_rate": 8.957777777777778e-06,
+      "loss": 0.0744,
+      "step": 975
+    },
+    {
+      "epoch": 9.437054631828978,
+      "grad_norm": 7.651462554931641,
+      "learning_rate": 8.902222222222224e-06,
+      "loss": 0.1073,
+      "step": 1000
+    },
+    {
+      "epoch": 9.437054631828978,
+      "eval_loss": 1.0236197710037231,
+      "eval_runtime": 125.7559,
+      "eval_samples_per_second": 1.678,
+      "eval_steps_per_second": 0.843,
+      "eval_wer": 0.44484536082474224,
+      "step": 1000
+    },
+    {
+      "epoch": 9.437054631828978,
+      "step": 1000,
+      "total_flos": 8.10972659515392e+18,
+      "train_loss": 2.490273738861084,
+      "train_runtime": 2581.6643,
+      "train_samples_per_second": 15.494,
+      "train_steps_per_second": 1.937
+    }
+  ],
+  "logging_steps": 25,
+  "max_steps": 5000,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 48,
+  "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": 8.10972659515392e+18,
+  "train_batch_size": 2,
+  "trial_name": null,
+  "trial_params": null
+}