---
library_name: transformers
license: apache-2.0
base_model:
- nbeerbower/flammen16-mistral-7B
datasets:
- wenbopan/Chinese-dpo-pairs
tags:
- experimental
---

![image/png](https://huggingface.co./nbeerbower/flammen13X-mistral-7B/resolve/main/flammen13x.png)

# flammen16-chinese-DPO-7B

A Mistral 7B LLM built from merging pretrained models and finetuning on [Wenbo Pan](https://huggingface.co./wenbopan)'s [Chinese DPO Pairs](https://huggingface.co./datasets/wenbopan/Chinese-dpo-pairs).
Flammen specializes in exceptional character roleplay, creative writing, and general intelligence. 
Please note this is an experimental model and is not recommended for production use.

我是一款基于混合预训练模型并在温博潘的中文DPO对话双方数据上微调的缅德尔7B大语言模型（LLM）。它的特长在于出色的角色扮演、创造性写作和通用智能。请注意，这是一个实验性模型，不适宜生产使用。

### Method

Finetuned using an A100 on Google Colab. 🙏

[Fine-tune a Mistral-7b model with Direct Preference Optimization](https://towardsdatascience.com/fine-tune-a-mistral-7b-model-with-direct-preference-optimization-708042745aac) - [Maxime Labonne](https://huggingface.co./mlabonne)

### Configuration

LoRA, model, and training settings:

```python
# LoRA configuration
peft_config = LoraConfig(
    r=16,
    lora_alpha=16,
    lora_dropout=0.05,
    bias="none",
    task_type="CAUSAL_LM",
    target_modules=['k_proj', 'gate_proj', 'v_proj', 'up_proj', 'q_proj', 'o_proj', 'down_proj']
)

# Model to fine-tune
model = AutoModelForCausalLM.from_pretrained(
    model_name,
    torch_dtype=torch.bfloat16,
    load_in_4bit=True
)
model.config.use_cache = False

# Reference model
ref_model = AutoModelForCausalLM.from_pretrained(
    model_name,
    torch_dtype=torch.bfloat16,
    load_in_4bit=True
)

# Training arguments
training_args = TrainingArguments(
    per_device_train_batch_size=2,
    gradient_accumulation_steps=2,
    gradient_checkpointing=True,
    learning_rate=2e-5,
    lr_scheduler_type="cosine",
    max_steps=1000,
    save_strategy="no",
    logging_steps=1,
    output_dir=new_model,
    optim="paged_adamw_32bit",
    warmup_steps=100,
    bf16=True,
    report_to="wandb",
)

# Create DPO trainer
dpo_trainer = DPOTrainer(
    model,
    ref_model,
    args=training_args,
    train_dataset=dataset,
    tokenizer=tokenizer,
    peft_config=peft_config,
    beta=0.1,
    max_prompt_length=1024,
    max_length=1536,
    force_use_ref_model=True
)

# Fine-tune model with DPO
dpo_trainer.train()
```