README.md

---
base_model: unsloth/mistral-7b-v0.3-bnb-4bit
language:
- ja
license: apache-2.0
tags:
- text-generation-inference
- transformers
- unsloth
- mistral
- trl
- sft
---

# Model Overview:

日本語で質問すると、日本語で回答を得られます。<br>

This is a fine-tuned **unsloth/mistral-7b-v0.3-bnb-4bit** for **Japanese**.<br>
You can ask in Japanese to get the answers in Japanese.<br>
Made possible thanks to [a detailed notebook from Unsloth](https://colab.research.google.com/drive/1tEd1FrOXWMnCU9UIvdYhs61tkxdMuKZu?usp=sharing).
<br>
# Datasets Used:

- "**wikimedia/wikipedia**:" (20231101.ja) for continued pretaining
- "**FreedomIntelligence/alpaca-gpt4-japanese**" for instruction fine tuning

# Inference Template:

```
from transformers import pipeline

pipe = pipeline("text-generation", model="Ryu-m0m/16bit-japanese-finetuned-mistral-7b-v0")

instruction = "侍の歴史を簡単に教えてください。" # Can you give us a brief history of the Samurai?
response = pipe(
    instruction,
    max_length=150,         # Controls the length of the output
    temperature=0.7,        # Controls randomness; lower is more deterministic
    top_k=50,               # Limits sampling pool to top 50 tokens
    top_p=0.9,              # Nucleus sampling, considering tokens up to 90% cumulative probability
    num_return_sequences=1  # Generates only one response
)

print(response[0]['generated_text'])

```

# Contact me

Any questions or quality issues found in the model, please feel free to contact me.

# Uploaded  model

- **Developed by:** Ryu-m0m
- **License:** apache-2.0
- **Finetuned from model :** unsloth/mistral-7b-v0.3-bnb-4bit

This mistral model was trained 2x faster with [Unsloth](https://github.com/unslothai/unsloth) and Huggingface's TRL library.

[<img src="https://raw.githubusercontent.com/unslothai/unsloth/main/images/unsloth%20made%20with%20love.png" width="200"/>](https://github.com/unslothai/unsloth)