frameworks, license, tasks

frameworks	license	tasks
Pytorch	Apache License 2.0	text-generation

Fine-tuning the qwen2-7b-instruct model using the msagent-pro dataset and the loss_scale technique with swift, the script is as follows:

NPROC_PER_NODE=8 \
CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7 \
MASTER_PORT=29500 \
swift sft \
    --model_type qwen2-7b-instruct \
    --learning_rate 2e-6 \
    --sft_type full \
    --dataset msagent-pro \
    --gradient_checkpointing true \
    --gradient_accumulation_steps 8 \
    --deepspeed default-zero3 \
    --use_loss_scale true \
    --save_strategy epoch \
    --batch_size 1 \
    --num_train_epochs 1 \
    --max_length 4096 \
    --preprocess_num_proc 4 \
    --use_loss_scale true \
    --loss_scale_config_path agent-flan \
    --ddp_backend nccl \

Comparison with the Original Model on the ToolBench Evaluation Set

Model	ToolBench (in-domain)					ToolBench (out-of-domain)
	Plan.EM	Act.EM	HalluRate (lower is better)	Avg.F1	R-L	Plan.EM	Act.EM	HalluRate (lower is better)	Avg.F1
qwen2-7b-instruct	74.11	54.74	4.16	46.53	8.51	73.17	57.67	3.84	48.58
qwen2-7b-agent-instruct	83.37	60.01	2.58	54.41	26.34	82.57	60.14	1.79	55.25

For detailed explanations of the evaluation metrics, please refer to document

您可以通过如下git clone命令，或者ModelScope SDK来下载模型

SDK下载

#安装ModelScope
pip install modelscope

#SDK模型下载
from modelscope import snapshot_download
model_dir = snapshot_download('swift/qwen2-7b-agent-instruct')

Git下载

#Git模型下载
git clone https://www.modelscope.cn/swift/qwen2-7b-agent-instruct.git

如果您是本模型的贡献者，我们邀请您根据模型贡献文档，及时完善模型卡片内容。