79 lines
2.4 KiB
Markdown
79 lines
2.4 KiB
Markdown
|
---
|
||
|
|
base_model: Qwen/Qwen3-4B-Instruct-2507
|
||
|
|
datasets:
|
||
|
|
- 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
|
||
|
|
pipeline_tag: text-generation
|
||
|
|
tags:
|
||
|
|
- unsloth
|
||
|
|
- agent
|
||
|
|
- tool-use
|
||
|
|
- alfworld
|
||
|
|
---
|
||
|
|
|
||
|
|
# Qwen3-4B Agent Trajectory (v17)
|
||
|
|
|
||
|
|
This repository provides a **fully merged model** fine-tuned from **Qwen/Qwen3-4B-Instruct-2507** using Unsloth.
|
||
|
|
|
||
|
|
Unlike standard adapter repositories, this repository contains the **merged weights**, meaning you do not need to load the base model separately.
|
||
|
|
|
||
|
|
## Training Objective
|
||
|
|
|
||
|
|
This model is trained to improve **multi-turn agent task performance**
|
||
|
|
on ALFWorld (household tasks).
|
||
|
|
|
||
|
|
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.
|
||
|
|
|
||
|
|
## Data Processing
|
||
|
|
|
||
|
|
- Train/Validation Split: 95% / 5%
|
||
|
|
- Random Seed: 3407 (used for shuffling and initialization)
|
||
|
|
- Loss Masking: Loss was computed only on the assistant's responses. User prompts and observations were masked during training (`train_on_responses_only` was applied to `<|im_start|>assistant\n`).
|
||
|
|
|
||
|
|
## Training Configuration
|
||
|
|
- **Base model**: Qwen/Qwen3-4B-Instruct-2507
|
||
|
|
- **Method**: LoRA + Unsloth (Merged in 16-bit)
|
||
|
|
- **Max sequence length**: 8192
|
||
|
|
- **Epochs**: 1
|
||
|
|
- **Learning rate**: 3e-06
|
||
|
|
- **LoRA**: r=16, alpha=32
|
||
|
|
- **PER_DEVICE_TRAIN_BATCH_SIZE** = 4
|
||
|
|
- **GRAD_ACCUM** = 4
|
||
|
|
- **WARMUP_RATIO** = 0.15
|
||
|
|
- **WEIGHT_DECAY** = 0.05
|
||
|
|
- **NEFTUNE_NOISE_ALPHA** = 5.0
|
||
|
|
- **VAL_RATIO** = 0.05
|
||
|
|
|
||
|
|
## Usage
|
||
|
|
|
||
|
|
```python
|
||
|
|
from transformers import AutoModelForCausalLM, AutoTokenizer
|
||
|
|
import torch
|
||
|
|
|
||
|
|
model_id = "choco800/qwen3-4b-agent-v17"
|
||
|
|
|
||
|
|
tokenizer = AutoTokenizer.from_pretrained(model_id)
|
||
|
|
model = AutoModelForCausalLM.from_pretrained(
|
||
|
|
model_id,
|
||
|
|
torch_dtype=torch.bfloat16,
|
||
|
|
device_map="auto",
|
||
|
|
)
|
||
|
|
```
|
||
|
|
## Sources & Terms (IMPORTANT)
|
||
|
|
|
||
|
|
Training data:
|
||
|
|
- u-10bei/dbbench_sft_dataset_react (available on Hugging Face Hub)
|
||
|
|
- u-10bei/dbbench_sft_dataset_react_v2 (available on Hugging Face Hub)
|
||
|
|
- u-10bei/dbbench_sft_dataset_react_v3 (available on Hugging Face Hub)
|
||
|
|
- u-10bei/dbbench_sft_dataset_react_v4
|
||
|
|
|
||
|
|
Dataset License: MIT License. These datasets are used and distributed under the terms of the MIT License.
|
||
|
|
Compliance: Users must comply with the dataset licenses and the base model's original terms of use (Apache 2.0).
|