Search results

qwen2vl- Slow preprocessing of multimodal data

bugSomething isn't workingSomething isn't working

pendingThis problem is yet to be addressedThis problem is yet to be addressed

Status: Open.

#8807 In hiyouga/LLaMA-Factory;

· xxddccaa opened on Aug 2, 2025

Qwen2.5-VL-7B-Instruct 勾选冻结模型中的语言模型和冻结模型中的视觉编码器，取消勾选冻结模型中的多模态投影器训练报错

bugSomething isn't workingSomething isn't working

pendingThis problem is yet to be addressedThis problem is yet to be addressed

Status: Open.

#8804 In hiyouga/LLaMA-Factory;

· qoop1145 opened on Aug 1, 2025

微调Kimi-VL时添加新tokens出现数组越界错误

bugSomething isn't workingSomething isn't working

pendingThis problem is yet to be addressedThis problem is yet to be addressed

Status: Open.

#8803 In hiyouga/LLaMA-Factory;

· yyhycx opened on Aug 1, 2025

多轮对话是否支持自定义选择某几轮不计算loss

bugSomething isn't workingSomething isn't working

pendingThis problem is yet to be addressedThis problem is yet to be addressed

Status: Open.

#8802 In hiyouga/LLaMA-Factory;

· zht8506 opened on Aug 1, 2025

v100 32g集群多节点微调出错 使用的是官方镜像hiyouga/llamafactory latest aef88c6874a2 NVIDIA-SMI 550.163.01 Driver Version: 550.163.01 CUDA Version: 12.4

bugSomething isn't workingSomething isn't working

pendingThis problem is yet to be addressedThis problem is yet to be addressed

Status: Open.

#8801 In hiyouga/LLaMA-Factory;

· bayunmm opened on Aug 1, 2025

internvl3-8B模型实行视频微调后，采用vllm推理结果不一致

enhancementNew feature or requestNew feature or request

pendingThis problem is yet to be addressedThis problem is yet to be addressed

Status: Open.

#8800 In hiyouga/LLaMA-Factory;

· zzb213213 opened on Aug 1, 2025

InternVL3-8B-hf ValueError: Processor was not found, please check and update your model file.

bugSomething isn't workingSomething isn't working

pendingThis problem is yet to be addressedThis problem is yet to be addressed

Status: Open.

#8799 In hiyouga/LLaMA-Factory;

· vl199967 opened on Aug 1, 2025

AssertionError: leaf_module_classes must be a list of types, got ['DeepseekV3MoE']

bugSomething isn't workingSomething isn't working

pendingThis problem is yet to be addressedThis problem is yet to be addressed

Status: Open.

#8797 In hiyouga/LLaMA-Factory;

· lvjiajun1120 opened on Aug 1, 2025

是否支持多模态reward模型训练

bugSomething isn't workingSomething isn't working

pendingThis problem is yet to be addressedThis problem is yet to be addressed

Status: Open.

#8796 In hiyouga/LLaMA-Factory;

· JoshonSmith opened on Jul 31, 2025

deepseed微调qwen2.5vl无法继续断点训练

bugSomething isn't workingSomething isn't working

pendingThis problem is yet to be addressedThis problem is yet to be addressed

Status: Open.

#8794 In hiyouga/LLaMA-Factory;

· Franklin-L opened on Jul 31, 2025

Flash Attention 2 on ROCm is SLOWER during sft training.

bugSomething isn't workingSomething isn't working

pendingThis problem is yet to be addressedThis problem is yet to be addressed

Status: Open.

#8793 In hiyouga/LLaMA-Factory;

· marksverdhei opened on Jul 31, 2025

Inconsistencies Between InternVL-1B and Qwen-VL-3B in Multi-Image DPO Training

solvedThis problem has been already solvedThis problem has been already solved

Status: Open.

#8789 In hiyouga/LLaMA-Factory;

· hexiao0275 opened on Jul 30, 2025