Workflow runs · Qihoo360/360-LLaMA-Factory · GitHub

Actions

All workflows

Actions

Loading...
Loading

Showing runs from all workflows

21 workflow runs

21 workflow runs

Support QAT training of gpt-oss label_issue #30: Issue #81 opened by shelterwff-byte

7s

7s

[Bug] AttributeError when training text-only models (e.g., Qwen2) with Sequence Parallelism label_issue #29: Issue #78 opened by chenzihan1

12s

12s

Qwen3-VL support label_issue #28: Issue #77 opened by Osilly

10s

10s

加载数据集报错 label_issue #27: Issue #76 opened by Zheng-Liming

11s

11s

sequence_parallel_attention makes model _update_causal_mask go wrong way ? label_issue #26: Issue #75 opened by PrometheusComing

11s

11s

function call dataset can not be trained label_issue #25: Issue #74 opened by Unknown-Body

9s

9s

Sliding Window Attention is enabled but not implemented for sequence_parallel_attention; unexpected results may be encountered. label_issue #24: Issue #73 opened by sherlock1987

10s

10s

Loss scale mismatch when enabling sequence parallelism label_issue #23: Issue #72 opened by DSTTSD

10s

10s

长序列训练显存未减少 label_issue #22: Issue #70 opened by nickle-fang

10s

10s

sft训练Qwen3卡在attention label_issue #21: Issue #69 opened by nickle-fang

14s

14s

开启zigzag-ring，训练速度很慢 label_issue #20: Issue #68 opened by xuchanganan

8s

8s

moe : RuntimeError: v must have shape (batch_size, seqlen_k, num_heads_k, head_size) label_issue #19: Issue #66 opened by liujia-cc

13s

13s

add sequence_parallel_size and error. label_issue #18: Issue #63 opened by guotong1988

8s

8s

RuntimeError: The size of tensor a (16648) must match the size of tensor b (16384) at non-singleton dimension 2 label_issue #17: Issue #62 opened by mrjin645

8s

8s

使用多模态数据集微调Qwen2.5-VL，KeyError: 'sequence_parallel_attention'。问题出现在模型加载阶段，Transformers库中的Qwen2.5-VL模型不支持 sequence_parallel_attention 这种注意力实现 label_issue #16: Issue #61 opened by hybdxxw

10s

10s

src/data/template.py 出现未被定义的变量 label_issue #15: Issue #60 opened by diansheng

7s

7s

code和requirement.txt中对transformers的要求不一致 label_issue #14: Issue #59 opened by Huihao-Yang

16s

16s

使用流式加载报错，传递了额外参数：TypeError: IterableDataset.map() got an unexpected keyword argument 'num_proc' label_issue #13: Issue #57 opened by CxsGhost

12s

12s

Support for Sequence Parallelism (SP) in the Multimodal Pipeline? label_issue #12: Issue #56 opened by Eisenhower

12s

12s

对于sp，当用ulysses替换ring_flash_attn时,为什么loss很大 label_issue #11: Issue #55 opened by cangshuli

7s

7s

Please update the QR code for wechat label_issue #10: Issue #54 opened by BeverYang

8s

8s