chore: update README.md content
Browse files
README.md
CHANGED
@@ -301,6 +301,30 @@ For direct use with `unsloth`, you can easily get started with the following ste
|
|
301 |
print(results)
|
302 |
```
|
303 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
304 |
### Instructions
|
305 |
|
306 |
Here are specific instructions and explanations for each use case.
|
|
|
301 |
print(results)
|
302 |
```
|
303 |
|
304 |
+
#### Use with MLX
|
305 |
+
|
306 |
+
For direct use with `mlx`, you can easily get started with the following steps.
|
307 |
+
|
308 |
+
- Firstly, you need to install unsloth via the command below with `pip`.
|
309 |
+
|
310 |
+
```bash
|
311 |
+
pip install mlx-lm
|
312 |
+
```
|
313 |
+
|
314 |
+
- Right now, you can start using the model directly.
|
315 |
+
```python
|
316 |
+
from mlx_lm import load, generate
|
317 |
+
|
318 |
+
model, tokenizer = load("ghost-x/ghost-8b-beta-1608-mlx")
|
319 |
+
messages = [
|
320 |
+
{"role": "system", "content": ""},
|
321 |
+
{"role": "user", "content": "Why is the sky blue ?"},
|
322 |
+
]
|
323 |
+
prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
|
324 |
+
response = generate(model, tokenizer, prompt=prompt, verbose=True)
|
325 |
+
```
|
326 |
+
|
327 |
+
|
328 |
### Instructions
|
329 |
|
330 |
Here are specific instructions and explanations for each use case.
|