swift3.0.3微调qwen2.5-vl-7b报错ValueError: model_type: 'qwen2_5_vl' not in [...

[24-solar-terms]

微调qwen2.5-vl-7b
环境：

python 3.10
swift 3.0.3
transformers 4.48.1
torch 2.1.0

根据文档model_type使用qwen2_5_vl
https://swift.readthedocs.io/zh-cn/latest/Instruction/%E6%94%AF%E6%8C%81%E7%9A%84%E6%A8%A1%E5%9E%8B%E5%92%8C%E6%95%B0%E6%8D%AE%E9%9B%86.html

完整训练脚本如下：

CUDA_VISIBLE_DEVICES=0 \
MAX_PIXELS=1073296 \
swift sft \
    --model /huggingface.co/Qwen/Qwen2.5-VL-7B-Instruct \
    --model_type qwen2_5_vl \
    --output_dir /logs/mdocgpt/qwen/debug \
    --dataset /sft_datasets/train.jsonl \
    --val_dataset /sft_datasets/test.jsonl \
    --train_type lora \
    --per_device_train_batch_size 2 \
    --per_device_eval_batch_size 2 \
    --num_train_epochs 10 \
    --save_strategy epoch \
    --gradient_checkpointing_kwargs '{"use_reentrant": false}' \
    --dataloader_num_workers 8

报错：

ValueError: model_type: 'qwen2_5_vl' not in ['emu3_gen', 'emu3_chat', 'baichuan', 'baichuan2', 'modern_bert', 'bert', 'chatglm2', 'chatglm3', 'glm4', 'longwriter_llama3_1', 'codegeex4', 'glm4v', 'cogvlm', 'cogagent_chat', 'cogagent_vqa', 'cogvlm2', 'cogvlm2_video', 'glm_edge', 'glm_edge_v', 'qwen', 'modelscope_agent', 'qwen_audio', 'qwen_vl', 'qwen2', 'qwen2_5', 'qwen2_5_math', 'qwen2_moe', 'qwen2_vl', 'qvq', 'qwen2_audio', 'marco_o1', 'qwq', 'ovis1_6', 'ovis1_6_llama3', 'qwen2_reward', 'qwen2_5_prm', 'qwen2_5_math_reward', 'codefuse_qwen', 'codefuse_codegeex2', 'codefuse_codellama', 'deepseek', 'deepseek_moe', 'deepseek_v2', 'deepseek_v2_5', 'deepseek_vl', 'deepseek_janus', 'deepseek_vl2', 'deepseek_r1', 'deepseek_r1_distill', 'paligemma', 'gemma', 'gemma2', 'internlm', 'internlm2', 'internlm3', 'internvl', 'internvl_phi3', 'internvl2', 'internvl2_phi3', 'internvl2_5', 'xcomposer2_5', 'xcomposer2', 'xcomposer2_4khd', 'xcomposer2_5_ol_audio', 'internlm2_reward', 'llama', 'llama3', 'llama3_1', 'llama3_2', 'llama3_2_vision', 'llama3_1_omni', 'reflection', 'atom', 'mengzi3', 'numina', 'ziya', 'megrez', 'llava_llama3_hf', 'llava1_5_hf', 'llava_onevision_hf', 'llava_next_qwen_hf', 'llama3_llava_next_hf', 'llava1_6_vicuna_hf', 'llava1_6_mistral_hf', 'llava_llama3_1_hf', 'llava1_6_yi_hf', 'llava_next_video_hf', 'llava_next_video_yi_hf', 'llama3_llava_next', 'llava1_6_mistral', 'llava1_6_yi', 'llava_next_qwen', 'grok', 'polylm', 'yuan2', 'orion', 'dbrx', 'bluelm', 'seggpt', 'xverse', 'xverse_moe', 'c4ai', 'aya', 'mamba', 'phi3_vision', 'florence', 'phi3_small', 'phi2', 'phi3', 'phi4', 'phi3_moe', 'minicpm_moe', 'minicpmv', 'minicpmv2_5', 'minicpmv2_6', 'minicpmo2_6', 'minicpm', 'minicpm_chatml', 'minicpm3', 'minimax_vl', 'minimax', 'mistral', 'mixtral', 'mistral_nemo', 'zephyr', 'wizardlm2_moe', 'wizardlm2', 'got_ocr2', 'idefics3', 'pixtral', 'molmoe', 'molmo', 'megrez_omni', 'mplug_owl2', 'mplug_owl2_1', 'mplug_owl3', 'mplug_owl3_241101', 'doc_owl2', 'openbuddy_llama', 'openbuddy_llama3', 'openbuddy_mistral', 'openbuddy_mixtral', 'skywork', 'skywork_o1', 'llama3_2_reward', 'gemma_reward', 'telechat', 'telechat2', 'valley', 'yi_vl', 'yi', 'yi_coder', 'sus']

[Jintao-Huang]

请使用main分支

[Jintao-Huang]

pip install git+https://github.com/modelscope/ms-swift.git

[24-solar-terms]

十分感谢您的回复，已成功