End of training

README.md

@@ -6,8 +6,9 @@ base_model: openai/whisper-base
 tags:
 - generated_from_trainer
 datasets:
-- deepdml/igbo-dict-16khz
+- google/fleurs
 - deepdml/igbo-dict-expansion-16khz
+- deepdml/igbo-dict-16khz
 metrics:
 - wer
 model-index:
@@ -21,11 +22,13 @@ model-index:
       type: google/fleurs
       config: ig_ng
       split: test
+      args: ig_ng
     metrics:
     - name: Wer
       type: wer
-      value: 155.96350889807658
+      value: 54.948739128322245
 ---
+
 <!-- 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. -->
 
@@ -33,8 +36,9 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [openai/whisper-base](https://huggingface.co/openai/whisper-base) on the google/fleurs dataset.
 It achieves the following results on the evaluation set:
-- Loss: 1.6790
-- Wer: 155.9635
+- Loss: 1.0933
+- Wer: 54.9487
+- Cer: 21.3532
 
 ## Model description
 
@@ -55,22 +59,22 @@ More information needed
 The following hyperparameters were used during training:
 - learning_rate: 1e-05
 - train_batch_size: 64
-- eval_batch_size: 8
+- eval_batch_size: 64
 - seed: 42
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: linear
-- lr_scheduler_warmup_steps: 500
+- lr_scheduler_warmup_ratio: 0.04
 - training_steps: 5000
 
 ### Training results
 
-| Training Loss | Epoch  | Step | Validation Loss | Wer      |
-|:-------------:|:------:|:----:|:---------------:|:--------:|
-| 0.2318        | 0.2    | 1000 | 1.3526          | 68.0029  |
-| 0.0786        | 1.0814 | 2000 | 1.5104          | 123.0631 |
-| 0.0627        | 1.2814 | 3000 | 1.5945          | 166.5873 |
-| 0.0317        | 2.1628 | 4000 | 1.6534          | 141.3940 |
-| 0.0321        | 3.0442 | 5000 | 1.6790          | 155.9635 |
+| Training Loss | Epoch  | Step | Validation Loss | Wer     | Cer     |
+|:-------------:|:------:|:----:|:---------------:|:-------:|:-------:|
+| 0.2087        | 0.2    | 1000 | 0.8427          | 54.4143 | 20.1160 |
+| 0.0734        | 1.0814 | 2000 | 0.9702          | 55.5707 | 21.6200 |
+| 0.0609        | 1.2814 | 3000 | 1.0272          | 54.0256 | 20.4927 |
+| 0.0336        | 2.1628 | 4000 | 1.0804          | 54.4337 | 20.4677 |
+| 0.0341        | 3.0442 | 5000 | 1.0933          | 54.9487 | 21.3532 |
 
 
 ### Framework versions
@@ -79,14 +83,3 @@ The following hyperparameters were used during training:
 - Pytorch 2.3.0+cu121
 - Datasets 2.19.1
 - Tokenizers 0.19.1
-
-## Citation
-
-```bibtex
-@misc{deepdml/whisper-base-ig-mix-norm,
-      title={Fine-tuned Whisper base ASR model for speech recognition in Igbo},
-      author={Jimenez, David},
-      howpublished={\url{https://huggingface.co/deepdml/whisper-base-ig-mix-norm}},
-      year={2025}
-    }
-```

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:698788d295ec7c648a79ee1082c461e800e329b4f4adad747692a65ceecebe49
+oid sha256:e6e7fab85f5e5b8b4db3eb17ef1b3c862f6e3684a8c2599a37b403b0f82ea364
 size 290403936