Upload README.md with huggingface_hub

README.md

@@ -0,0 +1,41 @@
+---
+language: he
+tags:
+  - whisper
+  - hebrew
+  - acft
+  - speech-recognition
+library_name: transformers
+base_model: yoad/whisper-tiny
+---
+
+# whisper-yoad-tiny-he-acft
+
+ACFT (Audio Context Fine-Tuning) applied to [yoad/whisper-tiny](https://huggingface.co/yoad/whisper-tiny) for Hebrew speech recognition.
+
+ACFT aligns partial-context encoder representations with full-context ones, improving short-utterance inference (e.g., keyboard dictation).
+
+## Evaluation
+
+WER on ivrit-ai/whisper-training test split (2000 samples, no normalization):
+
+| Model | WER |
+|-------|-----|
+| yoad/whisper-tiny (base) | 0.3978 |
+| yoad/whisper-tiny + ACFT (this model) | 0.3993 |
+
+## Training
+
+- **Method:** ACFT (encoder MSE alignment)
+- **Dataset:** google/fleurs he_il
+- **Epochs:** 8
+- **Device:** Apple MPS (M4 Pro)
+
+## Usage
+
+```python
+from transformers import WhisperForConditionalGeneration, WhisperProcessor
+
+model = WhisperForConditionalGeneration.from_pretrained("amitkot/whisper-yoad-tiny-he-acft")
+processor = WhisperProcessor.from_pretrained("amitkot/whisper-yoad-tiny-he-acft")
+```