README.md

---
base_model:
- tokyotech-llm/Swallow-70b-instruct-hf
- allenai/tulu-2-dpo-70b
tags:
- mergekit
- merge
language:
  - en
  - ja
library_name: transformers
pipeline_tag: text-generation
license: llama2
model_type: llama
---
# Superswallow-70b-v0.2

**Important Notice:**

This model partially utilizes the parameters of Tulu V2 DPO finetuned based on Llama 2, so it may inherit the AI2 ImpACT license. Please use the model keeping in mind that there may be changes regarding the license if AI2 contacts me.

The [AI2 ImpACT license](https://allenai.org/impact-license) includes information about data artifacts and model artifacts, but does not cover the case of directly applying parts of the LLM parameters of a model artifact to other models. However, I respect their research and great work, so I will change the license immediately if AI2 contacts me.

## Description

This is a merge of pre-trained language models created using [mergekit](https://github.com/cg123/mergekit). The model was created by injecting the ability to follow user intent from [Tulu 2 DPO](https://arxiv.org/abs/2311.10702) into the [Swallow](https://zenn.dev/tokyotech_lm/articles/d6cb3a8fdfc907) instract model.

It was a proof of concept for merging LLMs trained in other languages, and paid close attention to preserving the linguistic capabilities of the merge-based model.

As far as I know, Swallow is the full set Llama 2 model(7B, 13B, 70B) that can output the most beautiful Japanese. Therefore, I used it as the base model for merging this time. Thank you for their wonderful work.

## Test environment

This model was tested using [text-generation-webui](https://github.com/oobabooga/text-generation-webui/tree/main). I use preset `simple-1` and `Null preset` for Generation.

### Recommendation

Use `simple-1` settings:
- temperature: 0.7
- top_p: 0.9
- repetition_penalty: 1.15
- top_k: 20

### Tested `temperature` Range

- temperature: 0.3 - 1.0

### Tested `repetition_penalty` Range

- repetition_penalty: 1.0 - 1.15

## Prompt template

All prompt templates are available as well.

### Tulu Style

```
<|user|>
Your message here!
<|assistant|>

```

For best results, format all inputs in this manner. **Make sure to include a newline after `<|assistant|>`, this can affect generation quality quite a bit.**

### Swallow Style (Alpaca format)

```
以下に、あるタスクを説明する指示があり、それに付随する入力が更なる文脈を提供しています。リクエストを適切に完了するための回答を記述してください。

### 指示:
{instruction}

### 応答:

```

## Use the instruct model

```
import torch
from transformers import AutoTokenizer, AutoModelForCausalLM

model_name = "nitky/Superswallow-70b-v0.2"

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


PROMPT_DICT = {
    "prompt_input": (
        "以下に、あるタスクを説明する指示があり、それに付随する入力が更なる文脈を提供しています。"
        "リクエストを適切に完了するための回答を記述してください。\n\n"
        "### 指示:\n{instruction}\n\n### 入力:\n{input}\n\n### 応答:"

    ),
    "prompt_no_input": (
        "以下に、あるタスクを説明する指示があります。"
        "リクエストを適切に完了するための回答を記述してください。\n\n"
        "### 指示:\n{instruction}\n\n### 応答:"
    ),
}

def create_prompt(instruction, input=None):
    """
    Generates a prompt based on the given instruction and an optional input.
    If input is provided, it uses the 'prompt_input' template from PROMPT_DICT.
    If no input is provided, it uses the 'prompt_no_input' template.

    Args:
        instruction (str): The instruction describing the task.
        input (str, optional): Additional input providing context for the task. Default is None.

    Returns:
        str: The generated prompt.
    """
    if input:
        # Use the 'prompt_input' template when additional input is provided
        return PROMPT_DICT["prompt_input"].format(instruction=instruction, input=input)
    else:
        # Use the 'prompt_no_input' template when no additional input is provided
        return PROMPT_DICT["prompt_no_input"].format(instruction=instruction)

# Example usage
instruction_example = "以下のトピックに関する詳細な情報を提供してください。"
input_example = "東京工業大学の主なキャンパスについて教えてください"
prompt = create_prompt(instruction_example, input_example)

input_ids = tokenizer.encode(
    prompt,
    add_special_tokens=False,
    return_tensors="pt"
)

tokens = model.generate(
    input_ids.to(device=model.device),
    max_new_tokens=200,
    temperature=0.7,
    top_p=0.9,
    repetition_penalty=1.15,
    top_k=20,
    do_sample=True,
)

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

```

## Merge Details
### Merge Method

This model was merged using the [DARE](https://arxiv.org/abs/2311.03099) [TIES](https://arxiv.org/abs/2306.01708) and the SLERP merge method using [tokyotech-llm/Swallow-70b-instruct-hf](https://huggingface.co/tokyotech-llm/Swallow-70b-instruct-hf) as a base.

### Models Merged

The following models were included in the merge:
* [allenai/tulu-2-dpo-70b](https://huggingface.co/allenai/tulu-2-dpo-70b)

### Configuration

The following YAML configuration was used to produce this model:

```yaml
models:
  - model: tokyotech-llm/Swallow-70b-instruct-hf
    # no parameters necessary for base model
  - model: allenai/tulu-2-dpo-70b # follow user intent
    parameters:
      density: 1
      weight:
      - filter: mlp.down_proj
        value: [0.45, 0.10, 0.45, 0.10, 0.45, 0.10, 0.45, 0.10, 0.10]
      - filter: mlp.gate_proj 
        value: [0.70, 0.10, 0.45, 0.10, 0.45, 0.10, 0.45, 0.10, 0.10]
      - filter: mlp.up_proj
        value: [0.70, 0.10, 0.45, 0.10, 0.45, 0.10, 0.45, 0.10, 0.10]
      - filter: self_attn
        value: [0.70, 0.45, 0.10, 0.45, 0.10, 0.45, 0.10, 0.45, 0.45]
      - value: 0 # fallback for rest of tensors.
merge_method: dare_ties
base_model: tokyotech-llm/Swallow-70b-instruct-hf
dtype: bfloat16
tokenizer_source: union
name: Superswallow-70b-v0.2-flavor
---
slices:
  - sources:
      - model: nitky/Superswallow-70b-baseline
        layer_range: [0, 80]
      - model: Superswallow-70b-v0.2-flavor
        layer_range: [0, 80]
merge_method: slerp
base_model: nitky/Superswallow-70b-baseline
parameters:
  t: # model stabilization
    - filter: self_attn
      value: [0, 0.5, 0.3, 0.7, 1]
    - filter: mlp
      value: [1, 0.5, 0.7, 0.3, 0]
    - value: 0.5 # fallback for rest of tensors
dtype: bfloat16
name: Superswallow-70b-v0.2

```