End of training

README.md

@@ -2,38 +2,39 @@
 license: apache-2.0
 base_model: openai/whisper-base
 tags:
+- whisper-event
 - generated_from_trainer
 datasets:
-- fleurs
+- google/fleurs
 metrics:
 - wer
 model-index:
-- name: breeze-listen-dsw-base-kn
+- name: Breeze DSW Kannada - base
   results:
   - task:
       name: Automatic Speech Recognition
       type: automatic-speech-recognition
     dataset:
-      name: fleurs
-      type: fleurs
+      name: google/fleurs kn_in
+      type: google/fleurs
       config: kn_in
       split: test
       args: kn_in
     metrics:
     - name: Wer
       type: wer
-      value: 31.10585305105853
+      value: 30.612702366127024
 ---
 
 <!-- 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. -->
 
-# breeze-listen-dsw-base-kn
+# Breeze DSW Kannada - base
 
-This model is a fine-tuned version of [openai/whisper-base](https://huggingface.co/openai/whisper-base) on the fleurs dataset.
+This model is a fine-tuned version of [openai/whisper-base](https://huggingface.co/openai/whisper-base) on the google/fleurs kn_in dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.2549
-- Wer: 31.1059
+- Loss: 0.2258
+- Wer: 30.6127
 
 ## Model description
 

all_results.json

@@ -0,0 +1,12 @@
+{
+    "epoch": 14.01,
+    "eval_loss": 0.225830078125,
+    "eval_runtime": 2101.4644,
+    "eval_samples_per_second": 0.39,
+    "eval_steps_per_second": 0.025,
+    "eval_wer": 30.612702366127024,
+    "train_loss": 0.24649163818359374,
+    "train_runtime": 54303.9792,
+    "train_samples_per_second": 0.589,
+    "train_steps_per_second": 0.018
+}

eval_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 14.01,
+    "eval_loss": 0.225830078125,
+    "eval_runtime": 2101.4644,
+    "eval_samples_per_second": 0.39,
+    "eval_steps_per_second": 0.025,
+    "eval_wer": 30.612702366127024
+}

runs/Jan22_17-12-12_knight/events.out.tfevents.1705996372.knight.93645.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9f70601db6ad7ddd268137b31c70da9386d8dbd3dffec60706cabc467ba9cf50
+size 406

train_results.json

@@ -0,0 +1,7 @@
+{
+    "epoch": 14.01,
+    "train_loss": 0.24649163818359374,
+    "train_runtime": 54303.9792,
+    "train_samples_per_second": 0.589,
+    "train_steps_per_second": 0.018
+}

trainer_state.json

@@ -0,0 +1,360 @@
+{
+  "best_metric": 30.612702366127024,
+  "best_model_checkpoint": "/cosmos/home/sp-operator/ai/training/models/huggingface/scripts/../breeze-listen-dsw-base-kn/checkpoint-700",
+  "epoch": 14.006,
+  "eval_steps": 100,
+  "global_step": 1000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.03,
+      "learning_rate": 5.0453611334320685e-06,
+      "loss": 1.8464,
+      "step": 25
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 6.229195710491767e-06,
+      "loss": 1.642,
+      "step": 50
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 6.903829450223392e-06,
+      "loss": 1.2861,
+      "step": 75
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 7.377725845391017e-06,
+      "loss": 0.7196,
+      "step": 100
+    },
+    {
+      "epoch": 1.03,
+      "eval_loss": 0.5166015625,
+      "eval_runtime": 2113.7829,
+      "eval_samples_per_second": 0.388,
+      "eval_steps_per_second": 0.025,
+      "eval_wer": 55.21295143212951,
+      "step": 100
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 7.743343231239583e-06,
+      "loss": 0.4782,
+      "step": 125
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 8.041073861170494e-06,
+      "loss": 0.3642,
+      "step": 150
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 8.292222957399574e-06,
+      "loss": 0.3108,
+      "step": 175
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 8.509413541357755e-06,
+      "loss": 0.2769,
+      "step": 200
+    },
+    {
+      "epoch": 2.06,
+      "eval_loss": 0.253173828125,
+      "eval_runtime": 2105.4925,
+      "eval_samples_per_second": 0.389,
+      "eval_steps_per_second": 0.025,
+      "eval_wer": 36.15940224159402,
+      "step": 200
+    },
+    {
+      "epoch": 3.01,
+      "learning_rate": 8.700744577655557e-06,
+      "loss": 0.247,
+      "step": 225
+    },
+    {
+      "epoch": 3.04,
+      "learning_rate": 8.871723942761204e-06,
+      "loss": 0.2221,
+      "step": 250
+    },
+    {
+      "epoch": 3.06,
+      "learning_rate": 9.026267958246849e-06,
+      "loss": 0.2084,
+      "step": 275
+    },
+    {
+      "epoch": 4.02,
+      "learning_rate": 9.16726106663399e-06,
+      "loss": 0.1896,
+      "step": 300
+    },
+    {
+      "epoch": 4.02,
+      "eval_loss": 0.2166748046875,
+      "eval_runtime": 2098.7697,
+      "eval_samples_per_second": 0.391,
+      "eval_steps_per_second": 0.025,
+      "eval_wer": 32.72976338729763,
+      "step": 300
+    },
+    {
+      "epoch": 4.04,
+      "learning_rate": 9.296889251455016e-06,
+      "loss": 0.1747,
+      "step": 325
+    },
+    {
+      "epoch": 4.07,
+      "learning_rate": 9.416848797368692e-06,
+      "loss": 0.1644,
+      "step": 350
+    },
+    {
+      "epoch": 5.02,
+      "learning_rate": 9.528482449516371e-06,
+      "loss": 0.153,
+      "step": 375
+    },
+    {
+      "epoch": 5.04,
+      "learning_rate": 9.632871309784314e-06,
+      "loss": 0.1384,
+      "step": 400
+    },
+    {
+      "epoch": 5.04,
+      "eval_loss": 0.2037353515625,
+      "eval_runtime": 2094.8651,
+      "eval_samples_per_second": 0.391,
+      "eval_steps_per_second": 0.025,
+      "eval_wer": 31.835616438356162,
+      "step": 400
+    },
+    {
+      "epoch": 5.07,
+      "learning_rate": 9.73089868785391e-06,
+      "loss": 0.1342,
+      "step": 425
+    },
+    {
+      "epoch": 6.02,
+      "learning_rate": 9.823295589572114e-06,
+      "loss": 0.1195,
+      "step": 450
+    },
+    {
+      "epoch": 6.05,
+      "learning_rate": 9.910673836465484e-06,
+      "loss": 0.109,
+      "step": 475
+    },
+    {
+      "epoch": 7.0,
+      "learning_rate": 9.993550644973805e-06,
+      "loss": 0.1099,
+      "step": 500
+    },
+    {
+      "epoch": 7.0,
+      "eval_loss": 0.2030029296875,
+      "eval_runtime": 2095.5793,
+      "eval_samples_per_second": 0.391,
+      "eval_steps_per_second": 0.025,
+      "eval_wer": 31.056039850560396,
+      "step": 500
+    },
+    {
+      "epoch": 7.03,
+      "learning_rate": 9.56e-06,
+      "loss": 0.0928,
+      "step": 525
+    },
+    {
+      "epoch": 7.05,
+      "learning_rate": 9.060000000000001e-06,
+      "loss": 0.0871,
+      "step": 550
+    },
+    {
+      "epoch": 8.01,
+      "learning_rate": 8.560000000000001e-06,
+      "loss": 0.0842,
+      "step": 575
+    },
+    {
+      "epoch": 8.03,
+      "learning_rate": 8.06e-06,
+      "loss": 0.0707,
+      "step": 600
+    },
+    {
+      "epoch": 8.03,
+      "eval_loss": 0.21533203125,
+      "eval_runtime": 2096.294,
+      "eval_samples_per_second": 0.391,
+      "eval_steps_per_second": 0.025,
+      "eval_wer": 31.245330012453298,
+      "step": 600
+    },
+    {
+      "epoch": 8.06,
+      "learning_rate": 7.5600000000000005e-06,
+      "loss": 0.0659,
+      "step": 625
+    },
+    {
+      "epoch": 9.01,
+      "learning_rate": 7.06e-06,
+      "loss": 0.0643,
+      "step": 650
+    },
+    {
+      "epoch": 9.04,
+      "learning_rate": 6.560000000000001e-06,
+      "loss": 0.056,
+      "step": 675
+    },
+    {
+      "epoch": 9.06,
+      "learning_rate": 6.0600000000000004e-06,
+      "loss": 0.052,
+      "step": 700
+    },
+    {
+      "epoch": 9.06,
+      "eval_loss": 0.225830078125,
+      "eval_runtime": 2096.1998,
+      "eval_samples_per_second": 0.391,
+      "eval_steps_per_second": 0.025,
+      "eval_wer": 30.612702366127024,
+      "step": 700
+    },
+    {
+      "epoch": 10.02,
+      "learning_rate": 5.560000000000001e-06,
+      "loss": 0.0483,
+      "step": 725
+    },
+    {
+      "epoch": 10.04,
+      "learning_rate": 5.060000000000001e-06,
+      "loss": 0.0424,
+      "step": 750
+    },
+    {
+      "epoch": 10.06,
+      "learning_rate": 4.56e-06,
+      "loss": 0.0422,
+      "step": 775
+    },
+    {
+      "epoch": 11.02,
+      "learning_rate": 4.060000000000001e-06,
+      "loss": 0.0375,
+      "step": 800
+    },
+    {
+      "epoch": 11.02,
+      "eval_loss": 0.2413330078125,
+      "eval_runtime": 2094.1326,
+      "eval_samples_per_second": 0.392,
+      "eval_steps_per_second": 0.025,
+      "eval_wer": 31.220423412204234,
+      "step": 800
+    },
+    {
+      "epoch": 11.04,
+      "learning_rate": 3.5600000000000002e-06,
+      "loss": 0.0329,
+      "step": 825
+    },
+    {
+      "epoch": 11.07,
+      "learning_rate": 3.0600000000000003e-06,
+      "loss": 0.0329,
+      "step": 850
+    },
+    {
+      "epoch": 12.02,
+      "learning_rate": 2.56e-06,
+      "loss": 0.0303,
+      "step": 875
+    },
+    {
+      "epoch": 12.05,
+      "learning_rate": 2.06e-06,
+      "loss": 0.0256,
+      "step": 900
+    },
+    {
+      "epoch": 12.05,
+      "eval_loss": 0.250732421875,
+      "eval_runtime": 2095.7098,
+      "eval_samples_per_second": 0.391,
+      "eval_steps_per_second": 0.025,
+      "eval_wer": 31.06351183063512,
+      "step": 900
+    },
+    {
+      "epoch": 13.0,
+      "learning_rate": 1.56e-06,
+      "loss": 0.0281,
+      "step": 925
+    },
+    {
+      "epoch": 13.03,
+      "learning_rate": 1.06e-06,
+      "loss": 0.0247,
+      "step": 950
+    },
+    {
+      "epoch": 13.05,
+      "learning_rate": 5.6e-07,
+      "loss": 0.0229,
+      "step": 975
+    },
+    {
+      "epoch": 14.01,
+      "learning_rate": 6.000000000000001e-08,
+      "loss": 0.0245,
+      "step": 1000
+    },
+    {
+      "epoch": 14.01,
+      "eval_loss": 0.2548828125,
+      "eval_runtime": 2097.376,
+      "eval_samples_per_second": 0.391,
+      "eval_steps_per_second": 0.025,
+      "eval_wer": 31.10585305105853,
+      "step": 1000
+    },
+    {
+      "epoch": 14.01,
+      "step": 1000,
+      "total_flos": 2.067347538305876e+18,
+      "train_loss": 0.24649163818359374,
+      "train_runtime": 54303.9792,
+      "train_samples_per_second": 0.589,
+      "train_steps_per_second": 0.018
+    }
+  ],
+  "logging_steps": 25,
+  "max_steps": 1000,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 9223372036854775807,
+  "save_steps": 100,
+  "total_flos": 2.067347538305876e+18,
+  "train_batch_size": 32,
+  "trial_name": null,
+  "trial_params": null
+}