NPROC_PER_NODE=8 \ CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7,8 \ swift sft \ --model $MODEL_PATH_IL_SECOND_STAGE \ --run_name 'Qwen2.5-VL-7B-Instruct_nodule_data_llm_train ...
custom training data format for Chain-of-Thought (CoT) reasoning and 'Think with Images' in Qwen3-VL
Hi there! First off, huge thanks to the Qwen3-VL team for such an amazing contribution—really appreciate the work you’ve put into this. I’m just getting started in this field, and right now I’m trying ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results
Feedback