Update README.md
Browse files
README.md
CHANGED
|
@@ -31,10 +31,15 @@ wget https://huggingface.co/kotoba-tech/kotoba-whisper-v1.0-ggml/resolve/main/gg
|
|
| 31 |
3. Run inference using the provided sample audio:
|
| 32 |
|
| 33 |
```bash
|
| 34 |
-
wget
|
| 35 |
make -j && ./main -m models/ggml-kotoba-whisper-v1.0.bin -f sample_ja_speech.wav
|
| 36 |
```
|
| 37 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 38 |
### Quantized Model
|
| 39 |
To use the quantized model, download the quantized GGML weights:
|
| 40 |
|
|
@@ -44,7 +49,6 @@ wget https://huggingface.co/kotoba-tech/kotoba-whisper-v1.0-ggml/resolve/main/gg
|
|
| 44 |
|
| 45 |
Run inference on the sample audio:
|
| 46 |
```bash
|
| 47 |
-
wget
|
| 48 |
make -j && ./main -m models/ggml-kotoba-whisper-v1.0-q5_0.bin -f sample_ja_speech.wav
|
| 49 |
```
|
| 50 |
|
|
|
|
| 31 |
3. Run inference using the provided sample audio:
|
| 32 |
|
| 33 |
```bash
|
| 34 |
+
wget https://huggingface.co/kotoba-tech/kotoba-whisper-v1.0-ggml/resolve/main/sample_ja_speech.wav
|
| 35 |
make -j && ./main -m models/ggml-kotoba-whisper-v1.0.bin -f sample_ja_speech.wav
|
| 36 |
```
|
| 37 |
|
| 38 |
+
Note that it runs only with 16-bit WAV files, so make sure to convert your input before running the tool. For example, you can use ffmpeg like this:
|
| 39 |
+
```
|
| 40 |
+
ffmpeg -i input.mp3 -ar 16000 -ac 1 -c:a pcm_s16le output.wav
|
| 41 |
+
```
|
| 42 |
+
|
| 43 |
### Quantized Model
|
| 44 |
To use the quantized model, download the quantized GGML weights:
|
| 45 |
|
|
|
|
| 49 |
|
| 50 |
Run inference on the sample audio:
|
| 51 |
```bash
|
|
|
|
| 52 |
make -j && ./main -m models/ggml-kotoba-whisper-v1.0-q5_0.bin -f sample_ja_speech.wav
|
| 53 |
```
|
| 54 |
|