YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

qwensft-v5

Qwen/Qwen2.5-7B-Instruct をベースに、AgentBench DB Bench タスク向けに QLoRA SFT を行ったモデルです。

Base Model

  • Qwen/Qwen2.5-7B-Instruct

Training

  • 手法: QLoRA SFT (4-bit quantization + LoRA, merged)
  • LoRA: r=16, alpha=32, dropout=0.05
  • 学習率: 3e-5
  • エポック数: 1
  • データセット: WikiSQL スキーマベースの合成データ (1,500 samples)
    • SELECT: 600 samples
    • INSERT: 600 samples
    • UPDATE: 300 samples
  • 精度: BF16
  • 学習環境: Google Colab (T4 GPU)

Usage

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained("mirukumiruku/qwensft-v5", torch_dtype="bfloat16", device_map="auto")
tokenizer = AutoTokenizer.from_pretrained("mirukumiruku/qwensft-v5")

License

This model inherits the license of the base model Qwen/Qwen2.5-7B-Instruct.

Downloads last month
4
Safetensors
Model size
8B params
Tensor type
BF16
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support