Search results

加载parquet数据集出错 pyarrow.lib.ArrowNotImplementedError: Nested data conversions not implemented for chunked array outputs

bugSomething isn't workingSomething isn't working

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

Status: Open.

#7310 In hiyouga/LLaMA-Factory;

· Youxu-Shi opened on Mar 14, 2025

8*H20,8块H20 无法使用DPO训练qwen2.5-vl-7B模型，300帧视频理解

bugSomething isn't workingSomething isn't working

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

Status: Open.

#7309 In hiyouga/LLaMA-Factory;

· h123fire opened on Mar 14, 2025

运行llamafactory-cli chat指令，在Loading checkpoint shards后发生异常

bugSomething isn't workingSomething isn't working

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

Status: Open.

#7307 In hiyouga/LLaMA-Factory;

· Kobebryantss opened on Mar 14, 2025

Questions about qwen template modification

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

Status: Open.

#7306 In hiyouga/LLaMA-Factory;

· gom168 opened on Mar 14, 2025

多机多卡训练时卡住

bugSomething isn't workingSomething isn't working

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

Status: Open.

#7300 In hiyouga/LLaMA-Factory;

· loliw opened on Mar 14, 2025

Integration with verl?

enhancementNew feature or requestNew feature or request

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

Status: Open.

#7298 In hiyouga/LLaMA-Factory;

· zinccat opened on Mar 14, 2025

Dimension mismatch in attention

bugSomething isn't workingSomething isn't working

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

Status: Open.

#7293 In hiyouga/LLaMA-Factory;

· YusukeO opened on Mar 13, 2025

H200全量训练DeepSeek-R1-Distill-Llama-70B，采用zero3(batch_size=1)时溢出。采用zero3+offload（优化器、参数）时，显存占用少（141G显存只占用27G)，cpu占用高

bugSomething isn't workingSomething isn't working

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

Status: Open.

#7282 In hiyouga/LLaMA-Factory;

· github-eliviate opened on Mar 13, 2025

多模态数据集多的时候，数据加载失败

bugSomething isn't workingSomething isn't working

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

Status: Open.

#7266 In hiyouga/LLaMA-Factory;

· zhaop-l opened on Mar 12, 2025

streaming 训练卡在第一个step

bugSomething isn't workingSomething isn't working

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

Status: Open.

#7261 In hiyouga/LLaMA-Factory;

· zfr00 opened on Mar 12, 2025

GPU Imbalanced Loading

bugSomething isn't workingSomething isn't working

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

Status: Open.

#7250 In hiyouga/LLaMA-Factory;

· WillDreamer opened on Mar 11, 2025

微调 DeepSeek-R1 蒸馏模型，在 Chat 加载秩表现出色，但在导出部署到 Ollama 后问答准确率大幅下降

bugSomething isn't workingSomething isn't working

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

Status: Open.

#7238 In hiyouga/LLaMA-Factory;

· Nehcknarf opened on Mar 11, 2025