Model save

README.md

@@ -1,27 +1,14 @@
 ---
 library_name: transformers
 license: apache-2.0
-base_model: openai/whisper-medium
+base_model: openai/whisper-small
 tags:
 - generated_from_trainer
-datasets:
-- rbcurzon/ph_dialect_asr
 metrics:
 - wer
 model-index:
 - name: whisper-medium-test
-  results:
-  - task:
-      name: Automatic Speech Recognition
-      type: automatic-speech-recognition
-    dataset:
-      name: rbcurzon/ph_dialect_asr all
-      type: rbcurzon/ph_dialect_asr
-      args: all
-    metrics:
-    - name: Wer
-      type: wer
-      value: 0.12251221112488153
+  results: []
 ---
 
 <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -29,10 +16,10 @@ should probably proofread and complete it, then remove this comment. -->
 
 # whisper-medium-test
 
-This model is a fine-tuned version of [openai/whisper-medium](https://huggingface.co/openai/whisper-medium) on the rbcurzon/ph_dialect_asr all dataset.
+This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.2824
-- Wer: 0.1225
+- Loss: 0.4044
+- Wer: 0.1675
 
 ## Model description
 
@@ -55,26 +42,31 @@ The following hyperparameters were used during training:
 - train_batch_size: 4
 - eval_batch_size: 4
 - seed: 42
+- distributed_type: multi-GPU
+- num_devices: 2
 - gradient_accumulation_steps: 4
-- total_train_batch_size: 16
-- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
+- total_train_batch_size: 32
+- total_eval_batch_size: 8
+- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
 - lr_scheduler_type: linear
 - lr_scheduler_warmup_steps: 500
-- training_steps: 3000
+- training_steps: 5000
 - mixed_precision_training: Native AMP
 
 ### Training results
 
-| Training Loss | Epoch  | Step | Validation Loss | Wer    |
-|:-------------:|:------:|:----:|:---------------:|:------:|
-| 0.1981        | 1.6453 | 1000 | 0.2715          | 0.1455 |
-| 0.0236        | 3.2897 | 2000 | 0.2709          | 0.1267 |
-| 0.0066        | 4.9350 | 3000 | 0.2824          | 0.1225 |
+| Training Loss | Epoch   | Step | Validation Loss | Wer    |
+|:-------------:|:-------:|:----:|:---------------:|:------:|
+| 0.2023        | 2.9607  | 1000 | 0.3294          | 0.1874 |
+| 0.0309        | 5.9192  | 2000 | 0.3414          | 0.1717 |
+| 0.0037        | 8.8777  | 3000 | 0.3782          | 0.1708 |
+| 0.0018        | 11.8362 | 4000 | 0.3964          | 0.1680 |
+| 0.0013        | 14.7947 | 5000 | 0.4044          | 0.1675 |
 
 
 ### Framework versions
 
-- Transformers 4.54.0.dev0
-- Pytorch 2.7.1+cu126
-- Datasets 4.0.0
+- Transformers 4.56.0.dev0
+- Pytorch 2.6.0+cu124
+- Datasets 3.6.0
 - Tokenizers 0.21.2

generation_config.json

@@ -1,28 +1,44 @@
 {
   "alignment_heads": [
     [
-      13,
-      15
+      5,
+      3
     ],
     [
-      15,
+      5,
+      9
+    ],
+    [
+      8,
+      0
+    ],
+    [
+      8,
       4
     ],
     [
-      15,
-      15
+      8,
+      7
     ],
     [
-      16,
-      1
+      8,
+      8
     ],
     [
-      20,
+      9,
       0
     ],
     [
-      23,
-      4
+      9,
+      7
+    ],
+    [
+      9,
+      9
+    ],
+    [
+      10,
+      5
     ]
   ],
   "begin_suppress_tokens": [
@@ -226,8 +242,6 @@
     49870,
     50254,
     50258,
-    50358,
-    50359,
     50360,
     50361,
     50362
@@ -237,5 +251,5 @@
     "transcribe": 50359,
     "translate": 50358
   },
-  "transformers_version": "4.54.0.dev0"
+  "transformers_version": "4.56.0.dev0"
 }