End of training

README.md

@@ -3,6 +3,7 @@ library_name: transformers
 license: apache-2.0
 base_model: openai/whisper-tiny
 tags:
+- audio-classification
 - generated_from_trainer
 metrics:
 - accuracy
@@ -16,7 +17,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # whisper_tiny_fleurs
 
-This model is a fine-tuned version of [openai/whisper-tiny](https://huggingface.co/openai/whisper-tiny) on an unknown dataset.
+This model is a fine-tuned version of [openai/whisper-tiny](https://huggingface.co/openai/whisper-tiny) on the /home/investigacion/disco4TB/workspace_pablo/firvox_whisper_research/finetunnig/dataset/dataset_parquet/dataset_1000x6_noFirVox_correctedpaths.parquet dataset.
 It achieves the following results on the evaluation set:
 - Loss: 0.4373
 - Accuracy: 0.87

all_results.json

@@ -1,13 +1,13 @@
 {
-    "epoch": 1.0,
-    "eval_accuracy": 0.7744444444444445,
-    "eval_loss": 0.9658203125,
-    "eval_runtime": 31.5059,
-    "eval_samples_per_second": 28.566,
-    "eval_steps_per_second": 0.476,
-    "total_flos": 5.701007303717683e+16,
-    "train_loss": 1.3089996337890626,
-    "train_runtime": 250.7779,
-    "train_samples_per_second": 20.337,
-    "train_steps_per_second": 0.319
+    "epoch": 3.0,
+    "eval_accuracy": 0.87,
+    "eval_loss": 0.437255859375,
+    "eval_runtime": 41.2311,
+    "eval_samples_per_second": 21.828,
+    "eval_steps_per_second": 0.364,
+    "total_flos": 1.710302191115305e+17,
+    "train_loss": 0.698149045308431,
+    "train_runtime": 744.8875,
+    "train_samples_per_second": 20.54,
+    "train_steps_per_second": 0.322
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 1.0,
-    "eval_accuracy": 0.7744444444444445,
-    "eval_loss": 0.9658203125,
-    "eval_runtime": 31.5059,
-    "eval_samples_per_second": 28.566,
-    "eval_steps_per_second": 0.476
+    "epoch": 3.0,
+    "eval_accuracy": 0.87,
+    "eval_loss": 0.437255859375,
+    "eval_runtime": 41.2311,
+    "eval_samples_per_second": 21.828,
+    "eval_steps_per_second": 0.364
 }

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 1.0,
-    "total_flos": 5.701007303717683e+16,
-    "train_loss": 1.3089996337890626,
-    "train_runtime": 250.7779,
-    "train_samples_per_second": 20.337,
-    "train_steps_per_second": 0.319
+    "epoch": 3.0,
+    "total_flos": 1.710302191115305e+17,
+    "train_loss": 0.698149045308431,
+    "train_runtime": 744.8875,
+    "train_samples_per_second": 20.54,
+    "train_steps_per_second": 0.322
 }

trainer_state.json

@@ -1,57 +1,117 @@
 {
-  "best_metric": 0.7744444444444445,
-  "best_model_checkpoint": "/home/investigacion/disco4TB/workspace_pablo/finetuning_whisper_workbench/lang_detection_inference_tests/models/whisper_tiny_fleurs/checkpoint-80",
-  "epoch": 1.0,
+  "best_metric": 0.87,
+  "best_model_checkpoint": "/home/investigacion/disco4TB/workspace_pablo/finetuning_whisper_workbench/lang_detection_inference_tests/models/whisper_tiny_fleurs/checkpoint-240",
+  "epoch": 3.0,
   "eval_steps": 500,
-  "global_step": 80,
+  "global_step": 240,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 0.3125,
-      "grad_norm": 3.242062568664551,
-      "learning_rate": 2.3333333333333336e-05,
-      "loss": 1.6888,
+      "grad_norm": 3.2010996341705322,
+      "learning_rate": 3e-05,
+      "loss": 1.7132,
       "step": 25
     },
     {
       "epoch": 0.625,
-      "grad_norm": 11.466547012329102,
-      "learning_rate": 1.2916666666666668e-05,
-      "loss": 1.2639,
+      "grad_norm": 8.872227668762207,
+      "learning_rate": 2.652777777777778e-05,
+      "loss": 1.2437,
       "step": 50
     },
     {
       "epoch": 0.9375,
-      "grad_norm": 15.550214767456055,
-      "learning_rate": 2.916666666666667e-06,
-      "loss": 1.0287,
+      "grad_norm": 19.28681755065918,
+      "learning_rate": 2.3055555555555554e-05,
+      "loss": 0.9061,
       "step": 75
     },
     {
       "epoch": 1.0,
-      "eval_accuracy": 0.7744444444444445,
-      "eval_loss": 0.9658203125,
-      "eval_runtime": 48.9568,
-      "eval_samples_per_second": 18.384,
-      "eval_steps_per_second": 0.306,
+      "eval_accuracy": 0.7855555555555556,
+      "eval_loss": 0.7685546875,
+      "eval_runtime": 47.7585,
+      "eval_samples_per_second": 18.845,
+      "eval_steps_per_second": 0.314,
       "step": 80
     },
     {
-      "epoch": 1.0,
-      "step": 80,
-      "total_flos": 5.701007303717683e+16,
-      "train_loss": 1.3089996337890626,
-      "train_runtime": 250.7779,
-      "train_samples_per_second": 20.337,
-      "train_steps_per_second": 0.319
+      "epoch": 1.25,
+      "grad_norm": 14.38220500946045,
+      "learning_rate": 1.9861111111111114e-05,
+      "loss": 0.6677,
+      "step": 100
+    },
+    {
+      "epoch": 1.5625,
+      "grad_norm": 10.145868301391602,
+      "learning_rate": 1.638888888888889e-05,
+      "loss": 0.5584,
+      "step": 125
+    },
+    {
+      "epoch": 1.875,
+      "grad_norm": 14.750308990478516,
+      "learning_rate": 1.2916666666666668e-05,
+      "loss": 0.4682,
+      "step": 150
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.8388888888888889,
+      "eval_loss": 0.5185546875,
+      "eval_runtime": 47.9759,
+      "eval_samples_per_second": 18.759,
+      "eval_steps_per_second": 0.313,
+      "step": 160
+    },
+    {
+      "epoch": 2.1875,
+      "grad_norm": 12.981527328491211,
+      "learning_rate": 9.444444444444445e-06,
+      "loss": 0.3672,
+      "step": 175
+    },
+    {
+      "epoch": 2.5,
+      "grad_norm": 19.92900848388672,
+      "learning_rate": 6.111111111111111e-06,
+      "loss": 0.3378,
+      "step": 200
+    },
+    {
+      "epoch": 2.8125,
+      "grad_norm": 12.515904426574707,
+      "learning_rate": 2.638888888888889e-06,
+      "loss": 0.286,
+      "step": 225
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.87,
+      "eval_loss": 0.437255859375,
+      "eval_runtime": 47.9886,
+      "eval_samples_per_second": 18.754,
+      "eval_steps_per_second": 0.313,
+      "step": 240
+    },
+    {
+      "epoch": 3.0,
+      "step": 240,
+      "total_flos": 1.710302191115305e+17,
+      "train_loss": 0.698149045308431,
+      "train_runtime": 744.8875,
+      "train_samples_per_second": 20.54,
+      "train_steps_per_second": 0.322
     }
   ],
   "logging_steps": 25,
-  "max_steps": 80,
+  "max_steps": 240,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 1,
+  "num_train_epochs": 3,
   "save_steps": 500,
   "stateful_callbacks": {
     "TrainerControl": {
@@ -65,7 +125,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 5.701007303717683e+16,
+  "total_flos": 1.710302191115305e+17,
   "train_batch_size": 16,
   "trial_name": null,
   "trial_params": null