Summary
Summary: a task in LegoFlow-SWE (Harbor dataset). Add support for fine-tuning InternVL2.5 and InternVL3 series multimodal models (e.g., OpenGVLab/InternVL2 5-8B , InternVL2 5-1B-MPO-hf ). Currently, LlamaFactory does not provide a template or multimodal plugin for these models, preventing users…
The task
Add support for fine-tuning InternVL2.5 and InternVL3 series multimodal models (e.g., `OpenGVLab/InternVL2_5-8B`, `InternVL2_5-1B-MPO-hf`). Currently, LlamaFactory does not provide a template or multimodal plugin for these models, preventing users from performing SFT, LoRA, or full fine-tuning.
Part of Lego-X/LegoFlow-SWE.