README.md

---
library_name: peft
base_model: tokyotech-llm/Swallow-MX-8x7b-NVE-v0.1
language:
- ja
license: apache-2.0
tags:
- text-generation-inference
- transformers
- trl
- mixtral
datasets:
- kunishou/amenokaku-code-instruct
license_name: mixtral
---

# Uploaded  model

- **Developed by:** taoki
- **License:** apache-2.0
- **Finetuned from model :** tokyotech-llm/Swallow-MX-8x7b-NVE-v0.1


# Usage

```python
import torch
from transformers import AutoModelForCausalLM, AutoTokenizer
from peft import PeftModel

model_name = "tokyotech-llm/Swallow-MX-8x7b-NVE-v0.1"
tokenizer = AutoTokenizer.from_pretrained(model_name, torch_dtype=torch.bfloat16, device_map="auto")
model = AutoModelForCausalLM.from_pretrained(model_name, load_in_4bit=True, torch_dtype=torch.bfloat16)

model = PeftModel.from_pretrained(model, "taoki/Swallow-MX-8x7b-NVE-v0.1-qlora-amenokaku-code-adapter")

prompt="""### Instruction:
紫式部と清少納言の作風をjsonで出力してください。
### Response:
"""

input_ids = tokenizer.encode(
    prompt,
    add_special_tokens=False,
    return_tensors="pt"
)
tokens = model.generate(
    input_ids.to(device=model.device),
    max_new_tokens=1024,
    temperature=0.99,
    top_p=0.95,
    do_sample=True,
)

out = tokenizer.decode(tokens[0], skip_special_tokens=True)
print(out)
```

# Output
````
### Instruction:
紫式部と清少納言の作風をjsonで出力してください。
### Response:
```json
{
    "紫式部": "貴人に会って、その人が話していることを思い出しながら奏でると、これにまさる楽器はありません。」,
    "清少納言": "人によってあげくはなく、おのずからかなしくゆくほどに、かなしみは深くなりゆきなさるなり。」
}
```
````

# Framework versions

- PEFT 0.9.0