89 lines
3.0 KiB
Markdown
89 lines
3.0 KiB
Markdown
---
|
|
base_model: Qwen/Qwen2.5-7B-Instruct
|
|
datasets:
|
|
- u-10bei/sft_alfworld_trajectory_dataset_v2
|
|
- u-10bei/sft_alfworld_trajectory_dataset_v3
|
|
- u-10bei/sft_alfworld_trajectory_dataset_v4
|
|
- u-10bei/sft_alfworld_trajectory_dataset_v5
|
|
- u-10bei/dbbench_sft_dataset_react
|
|
- u-10bei/dbbench_sft_dataset_react_v2
|
|
- u-10bei/dbbench_sft_dataset_react_v3
|
|
- u-10bei/dbbench_sft_dataset_react_v4
|
|
language:
|
|
- en
|
|
license: apache-2.0
|
|
library_name: transformers
|
|
pipeline_tag: text-generation
|
|
tags:
|
|
- sft
|
|
- agent
|
|
- tool-use
|
|
- alfworld
|
|
- dbbench
|
|
---
|
|
|
|
# qwen25-7b-sft-merged-v5v6-a50
|
|
|
|
This repository provides a **fully merged model** fine-tuned from
|
|
**Qwen2.5-7B-Instruct** using **QLoRA + Unsloth**.
|
|
|
|
Two SFT models (v5 and v6) were trained independently, then combined via
|
|
weight interpolation (alpha=0.5). This is a **complete model** — no adapters
|
|
or additional weights are needed.
|
|
|
|
## Training Objective
|
|
|
|
This model is trained to improve **multi-turn agent task performance**
|
|
on ALFWorld (household tasks) and DBBench (database operations).
|
|
|
|
Loss is applied to **all assistant turns** in the multi-turn trajectory,
|
|
enabling the model to learn environment observation, action selection,
|
|
tool use, and recovery from errors.
|
|
|
|
## Training Configuration
|
|
|
|
- Base model: Qwen/Qwen2.5-7B-Instruct
|
|
- Method: QLoRA (4-bit) + Unsloth, merged into base model
|
|
- Max sequence length: 2048
|
|
- Epochs: 2
|
|
- Learning rate: 5e-5
|
|
- LoRA: r=32, alpha=64
|
|
- Post-training: weight interpolation of v5 and v6 (alpha=0.5)
|
|
|
|
## Usage
|
|
|
|
```python
|
|
from transformers import AutoModelForCausalLM, AutoTokenizer
|
|
import torch
|
|
|
|
model_id = "plotMaker/qwen25-7b-sft-merged-v5v6-a50"
|
|
|
|
tokenizer = AutoTokenizer.from_pretrained(model_id)
|
|
model = AutoModelForCausalLM.from_pretrained(
|
|
model_id,
|
|
torch_dtype=torch.bfloat16,
|
|
device_map="auto",
|
|
)
|
|
```
|
|
|
|
## References
|
|
|
|
- [Model Soups (Wortsman et al., 2022)](https://arxiv.org/abs/2203.05482) — Weight interpolation of fine-tuned
|
|
models
|
|
- [LoRA (Hu et al., 2021)](https://arxiv.org/abs/2106.09685) — Low-Rank Adaptation
|
|
- [NEFTune (Jain et al., 2024)](https://arxiv.org/abs/2310.05914) — Noisy embedding fine-tuning
|
|
- [rsLoRA (Kalajdzievski, 2023)](https://arxiv.org/abs/2312.03732) — Rank-stabilized LoRA scaling
|
|
- [ALFWorld (Shridhar et al., 2021)](https://arxiv.org/abs/2010.03768) — Interactive text-world environments
|
|
- [ReAct (Yao et al., 2023)](https://arxiv.org/abs/2210.03629) — Reasoning and acting in LLMs
|
|
|
|
## Sources & Terms (IMPORTANT)
|
|
|
|
Training data:
|
|
- u-10bei/sft_alfworld_trajectory_dataset_v2 ~ v5
|
|
- u-10bei/dbbench_sft_dataset_react ~ v4
|
|
|
|
Base model: [Qwen/Qwen2.5-7B-Instruct](https://huggingface.co/Qwen/Qwen2.5-7B-Instruct)
|
|
|
|
This repository does NOT redistribute the dataset.
|
|
Users must comply with the dataset license and base model terms.
|