README.md

---
library_name: transformers
model_name: shisa-v2-llama3.3-70b
license: llama3.3
datasets:
- shisa-ai/shisa-v2-sharegpt
- shisa-ai/deepseekv3-ultrafeedback-armorm-dpo
language:
- ja
- en
base_model: shisa-ai/shisa-v2-llama3.3-70b
pipeline_tag: text-generation
tags:
- mlx
---

# mlx-community/shisa-v2-llama3.3-70b-mlx-fp16

The Model [mlx-community/shisa-v2-llama3.3-70b-mlx-bf16](https://huggingface.co/mlx-community/shisa-v2-llama3.3-70b-mlx-bf16) was converted to MLX format from [shisa-ai/shisa-v2-llama3.3-70b](https://huggingface.co/shisa-ai/shisa-v2-llama3.3-70b) using mlx-lm version **0.22.3**.

## Use with mlx

```bash
pip install mlx-lm
```

```python
from mlx_lm import load, generate

model, tokenizer = load("bibproj/shisa-v2-llama3.3-70b-mlx-fp16")

prompt="hello"

if hasattr(tokenizer, "apply_chat_template") and tokenizer.chat_template is not None:
    messages = [{"role": "user", "content": prompt}]
    prompt = tokenizer.apply_chat_template(
        messages, tokenize=False, add_generation_prompt=True
    )

response = generate(model, tokenizer, prompt=prompt, verbose=True)
```