README.md

---
tags:
- faster-whisper
- ctranslate2
- quantized
- int8
- speech-recognition
- automatic-speech-recognition
license: apache-2.0
---

# faster-whisper-base-int8

INT8-quantized [CTranslate2](https://github.com/OpenNMT/CTranslate2) export of
`openai/whisper-base`, converted for use with
[faster-whisper](https://github.com/SYSTRAN/faster-whisper).

- **Quantization**: INT8 (weights compressed from FP32 → INT8; ~75% size reduction)
- **Target hardware**: CPU, tested down to 2 cores / 4GB RAM
- **Source checkpoint**: [`openai/whisper-base`](https://huggingface.co/openai/whisper-base)

## Usage

```python
from faster_whisper import WhisperModel

model = WhisperModel(
    "devxyasir/faster-whisper-base-int8",
    device="cpu",
    compute_type="int8",
    cpu_threads=2,
    num_workers=1,
)

segments, info = model.transcribe("audio.mp3", beam_size=1, vad_filter=True)
for segment in segments:
    print(segment.text)
```