This repository provides a LoRA (Low-Rank Adaptation) adapter for [Base Model Name]. This model has been specifically fine-tuned through a multi-stage process, starting from SFT (Supervised Fine-Tuning) followed by DPO (Direct Preference Optimization) to enhance its reasoning and alignment capabilities.

Training Configuration

Base model: Qwen/Qwen3-4B-Instruct-2507
SFT Phase Firest, the base model was trained using the [Dataset Name/Description] to acquire domain-specific knowledge. (See: Hi-Satoh/sft-base4-qwen-adapter)
DPO Phase
Method: DPO (Direct Preference Optimization)
Epochs: 1.5
Learning rate: 1e-06
Beta: 0.1
Max sequence length: 2048
LoRA Config: r=8, alpha=16 (merged into base)

Usage

Since this is a merged model, you can use it directly with transformers.

python

from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

model_id = "your_id/your-repo-name"

tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    torch_dtype=torch.float16,
    device_map="auto"
)

# Test inference
prompt = "Your question here"
inputs = tokenizer.apply_chat_template([{"role": "user", "content": prompt}], tokenize=True, add_generation_prompt=True, return_tensors="pt").to("cuda")
outputs = model.generate(**inputs, max_new_tokens=512)
print(tokenizer.decode(outputs[0]))

Sources & License (IMPORTANT)

Training Data: [u-10bei/dpo-dataset-qwen-cot]
License: MIT License. (As per dataset terms).
Compliance: Users must follow the original base model's license terms.

Hi-Satoh/sft-base4-dpo-base3-qwen-cot-merged

＜【課題】sft-base4-dpo-base3-qwen-cot-merged＞

Training Objective

Training Configuration

Usage

Sources & License (IMPORTANT)

Hi-Satoh/
sft-base4-dpo-base3-qwen-cot-merged