README.md

---
library_name: transformers
license: mit
language:
- ja
- en
---

# Stockmark-2-100B-Instruct

![image/jpeg](https://cdn-uploads.huggingface.co/production/uploads/607ef1c3e758c3c5a2959eab/AbyPvKu-FBY6RDYhGi1KX.jpeg)

## Model description

**Stockmark-2-100B-Instruct** is a 100-billion-parameter large language model built from scratch, with a particular focus on Japanese. It was pre-trained on approximately 2.0 trillion tokens of data, consisting of 60% English, 30% Japanese, and 10% code. Following pretraining, the model underwent post-training (SFT and DPO) with synthetic data in Japanese to enhance its ability to follow instructions. This version improves instruction-following ability and adds support for long-context (32k), compared to the previous version ([Stockmark-2-100B-Instruct-beta](https://huggingface.co/stockmark/Stockmark-2-100B-Instruct-beta)).

This project was supported by [GENIAC](https://www.meti.go.jp/policy/mono_info_service/geniac/index.html).

### Features

- Model Type: Causal Language Model
- Number of Parameters: 96B
- Number of Layers: 86
- Number of Attention Heads (GQA): 72 for Q and 8 for KV
- Context Length: 32k
- Supported Languages: Japanese and English

## Model performance

### Japanese MT-bench
| Model | Average | coding | extraction | humanities | math | reasoning | roleplay | stem |
|:---:|:---:|:---:|:---:|:---:|:---:|:---:|:---:|:---:|
| Stockmark-2-100B-Instruct | 7.87 | 7.07 | 8.35 | 8.73 | 7.57 | 5.45 | 8.65 | 8.33 | 8.83 |
| Stockmark-2-100B-Instruct-beta | 7.71 | 6.73 | 8.23 | 8.63 | 7.01 | 5.85 | 8.54 | 8.07 | 8.61 |


## How to use

### transformers
```python
import torch
from transformers import AutoModelForCausalLM, AutoTokenizer

model_name = "stockmark/Stockmark-2-100B-Instruct"

tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name, device_map="auto", torch_dtype="bfloat16")

instruction = "自然言語処理とは？"
input_ids = tokenizer.apply_chat_template(
    [{"role": "user", "content": instruction}],
    add_generation_prompt=True,
    return_tensors="pt"
).to(model.device)

with torch.inference_mode():
  tokens = model.generate(
    input_ids,
    max_new_tokens = 512,
    do_sample = True,
    temperature = 0.7,
    top_p = 0.95
  )
    
output = tokenizer.decode(tokens[0], skip_special_tokens=True)
print(output)
```

### vLLM
```python
from vllm import LLM, SamplingParams

llm = LLM(
    model="stockmark/Stockmark-2-100B-Instruct",
    tensor_parallel_size=4,
    dtype="bfloat16"
)

sampling_params = SamplingParams(
    temperature=0.7,
    top_p=0.95,
    max_tokens=512
)

conversation = [{"role": "user", "content": "自然言語処理とは？"}]

outputs = llm.chat(conversation, sampling_params=sampling_params)

for output in outputs:
    generated_text = output.outputs[0].text
    print(generated_text)
```

## Libraries used for training
- Pretraining: [NVIDIA/Megatron-LM](https://github.com/NVIDIA/Megatron-LM)
- Posttraining: [huggingface/trl](https://github.com/huggingface/trl)

## License

[MIT](https://opensource.org/licenses/MIT)

## Developed by

[Stockmark Inc.](https://stockmark.co.jp/)