Automatic Speech Recognition
Transformers
TensorBoard
Safetensors
Croatian
whisper
Generated from Trainer
Instructions to use GoranS/whisper-large-v3-turbo-hr-parla with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use GoranS/whisper-large-v3-turbo-hr-parla with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("automatic-speech-recognition", model="GoranS/whisper-large-v3-turbo-hr-parla")# Load model directly from transformers import AutoProcessor, AutoModelForSpeechSeq2Seq processor = AutoProcessor.from_pretrained("GoranS/whisper-large-v3-turbo-hr-parla") model = AutoModelForSpeechSeq2Seq.from_pretrained("GoranS/whisper-large-v3-turbo-hr-parla", device_map="auto") - Notebooks
- Google Colab
- Kaggle
Update README.md
Browse files
README.md
CHANGED
|
@@ -24,21 +24,27 @@ private dataset, both with augmented additions.
|
|
| 24 |
|
| 25 |
It achieves the following results on the evaluation set:
|
| 26 |
- Loss: 0.0816
|
| 27 |
-
- Wer:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 28 |
|
| 29 |
-
## Model description
|
| 30 |
-
|
| 31 |
-
More information needed
|
| 32 |
-
|
| 33 |
-
## Intended uses & limitations
|
| 34 |
-
|
| 35 |
-
More information needed
|
| 36 |
-
|
| 37 |
-
## Training and evaluation data
|
| 38 |
-
|
| 39 |
-
More information needed
|
| 40 |
-
|
| 41 |
-
## Training procedure
|
| 42 |
|
| 43 |
### Training hyperparameters
|
| 44 |
|
|
|
|
| 24 |
|
| 25 |
It achieves the following results on the evaluation set:
|
| 26 |
- Loss: 0.0816
|
| 27 |
+
- Wer: 3.52%
|
| 28 |
+
|
| 29 |
+
|
| 30 |
+
### WER comparsion
|
| 31 |
+
|
| 32 |
+
| Dataset | Model | WER (%) |
|
| 33 |
+
|---------------------------------|----------------------------------------------------|----------|
|
| 34 |
+
| google/fleurs hr_hr test | openai/whisper-base | 65.99 |
|
| 35 |
+
| google/fleurs hr_hr test | openai/whisper-large-v3-turbo | 12.73 |
|
| 36 |
+
| google/fleurs hr_hr test | slsolucije/whisper-large-v3-turbo-hr-parla-lora-merged | 9.93 |
|
| 37 |
+
| google/fleurs hr_hr test | **GoranS/whisper-large-v3-turbo-hr-parla** | **8.66** |
|
| 38 |
+
| GoranS/stt-croatian_99k_265_2 test | openai/whisper-large-v3-turbo | 22.93 |
|
| 39 |
+
| GoranS/stt-croatian_99k_265_2 test | slsolucije/whisper-large-v3-turbo-hr-parla-lora-merged | 19.02 |
|
| 40 |
+
| GoranS/stt-croatian_99k_265_2 test | **GoranS/whisper-large-v3-turbo-hr-parla** | **18.44** |
|
| 41 |
+
| GoranS/stt-croatian-sl-31k test | openai/whisper-large-v3-turbo | 21.62 |
|
| 42 |
+
| GoranS/stt-croatian-sl-31k test | slsolucije/whisper-large-v3-turbo-hr-parla-lora-merged | 17.07 |
|
| 43 |
+
| GoranS/stt-croatian-sl-31k test | **GoranS/whisper-large-v3-turbo-hr-parla** | **16.97** |
|
| 44 |
+
| parla_867k_2483_0.5 test | openai/whisper-large-v3-turbo | 10.23 |
|
| 45 |
+
| parla_867k_2483_0.5 test | slsolucije/whisper-large-v3-turbo-hr-parla-lora-merged | 4.58 |
|
| 46 |
+
| parla_867k_2483_0.5 test | **GoranS/whisper-large-v3-turbo-hr-parla** | **3.52** |
|
| 47 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 48 |
|
| 49 |
### Training hyperparameters
|
| 50 |
|