End of training

README.md

@@ -8,9 +8,24 @@ tags:
 - generated_from_trainer
 datasets:
 - khmer-coupus
+metrics:
+- wer
 model-index:
 - name: Whisper Large V3 Turbo Khmer
-  results: []
+  results:
+  - task:
+      name: Automatic Speech Recognition
+      type: automatic-speech-recognition
+    dataset:
+      name: Common Voice 11.0
+      type: khmer-coupus
+      config: km_kh
+      split: test
+      args: 'config: khm, split: test'
+    metrics:
+    - name: Wer
+      type: wer
+      value: 101.09561752988047
 ---
 
 <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -19,6 +34,9 @@ should probably proofread and complete it, then remove this comment. -->
 # Whisper Large V3 Turbo Khmer
 
 This model is a fine-tuned version of [openai/whisper-large-v3-turbo](https://huggingface.co/openai/whisper-large-v3-turbo) on the Common Voice 11.0 dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.6277
+- Wer: 101.0956
 
 ## Model description
 
@@ -43,15 +61,25 @@ The following hyperparameters were used during training:
 - seed: 42
 - gradient_accumulation_steps: 4
 - total_train_batch_size: 32
-- optimizer: Use OptimizerNames.ADAMW_HF with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
+- optimizer: Use adamw_hf 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: 4000
 - mixed_precision_training: Native AMP
 
+### Training results
+
+| Training Loss | Epoch   | Step | Validation Loss | Wer      |
+|:-------------:|:-------:|:----:|:---------------:|:--------:|
+| 0.3899        | 15.8835 | 1000 | 0.6277          | 101.0956 |
+| 0.0522        | 31.7550 | 2000 | 0.7151          | 102.6394 |
+| 0.0051        | 47.6265 | 3000 | 1.0525          | 104.0090 |
+| 0.0023        | 63.4980 | 4000 | 1.1451          | 104.5319 |
+
+
 ### Framework versions
 
 - Transformers 4.49.0
-- Pytorch 2.5.1+cu124
+- Pytorch 2.6.0+cu124
 - Datasets 3.3.1
 - Tokenizers 0.21.0

generation_config.json

@@ -2,27 +2,27 @@
   "alignment_heads": [
     [
       2,
-      4
+      2
     ],
     [
-      2,
-      11
+      3,
+      0
     ],
     [
       3,
-      3
+      2
     ],
     [
       3,
-      6
+      3
     ],
     [
       3,
-      11
+      4
     ],
     [
       3,
-      14
+      5
     ]
   ],
   "begin_suppress_tokens": [
@@ -132,15 +132,14 @@
     "<|vi|>": 50278,
     "<|yi|>": 50335,
     "<|yo|>": 50325,
-    "<|yue|>": 50358,
     "<|zh|>": 50260
   },
   "language": "khmer",
   "max_initial_timestamp_index": 50,
   "max_length": 448,
-  "no_timestamps_token_id": 50364,
+  "no_timestamps_token_id": 50363,
   "pad_token_id": 50257,
-  "prev_sot_token_id": 50362,
+  "prev_sot_token_id": 50361,
   "return_timestamps": false,
   "suppress_tokens": [
     1,
@@ -226,16 +225,16 @@
     49870,
     50254,
     50258,
+    50358,
     50359,
     50360,
     50361,
-    50362,
-    50363
+    50362
   ],
   "task": "transcribe",
   "task_to_id": {
-    "transcribe": 50360,
-    "translate": 50359
+    "transcribe": 50359,
+    "translate": 50358
   },
   "transformers_version": "4.49.0"
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f2b74b1a475161cd0a6ec5e9ffb1d35cefedcfd5c2ef00a564905ff32be1d248
+oid sha256:4cb02400f7c8ee414f095dfa9a72cbc5859f9ec813cffeed2862e30ac549fad3
 size 151061672

runs/Feb19_01-41-03_rathanak/events.out.tfevents.1739904067.rathanak.62045.0

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:91657d73651d52739fea6cd30fc4ca633feb113e8cd1b321ca5fcbeed8bf0e20
-size 33089
+oid sha256:e25c5f38d162034fade99719d308ec7ece71ee56014011e2dceaad47ef9217a7
+size 42201

runs/Feb19_01-41-03_rathanak/events.out.tfevents.1739917632.rathanak.62045.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bdf0906581cee9c3987d85994a6f808a251d2ca7ce660a5a9001d844857642a4
+size 406