Skip to content

Actions: modelscope/ms-swift

Actions

All workflows

Actions

Loading...
Loading

Showing runs from all workflows
2,500+ workflow runs
2,500+ workflow runs

Filter by Workflow

⌥ + click/return to exclude

Filter by Event

⌥ + click/return to exclude

Filter by Status

⌥ + click/return to exclude

Filter by Branch

⌥ + click/return to exclude

Filter by Actor

⌥ + click/return to exclude
Fix NaN for full-parameter SFT with DeepSpeed ZeRO-1/2 on torch<2.9 (AdamW int32 flat-partition overflow) citest-npu #3610: Pull request #10023 synchronize by kiscad
Action required kiscad:main
kiscad:main
Action required
Fix NaN for full-parameter SFT with DeepSpeed ZeRO-1/2 on torch<2.9 (AdamW int32 flat-partition overflow) Lint test #18487: Pull request #10023 synchronize by kiscad
Action required kiscad:main
kiscad:main
Action required
Fix NaN for full-parameter SFT with DeepSpeed ZeRO-1/2 on torch<2.9 (AdamW int32 flat-partition overflow) citest #13516: Pull request #10023 synchronize by kiscad
Action required kiscad:main
kiscad:main
Action required
fix(megatron): forward mtp_shared_weights to megatron-bridge citest-npu #3609: Pull request #10025 opened by Excelius-Wang
47m 37s Excelius-Wang:fix/mtp-shared-weights
Excelius-Wang:fix/mtp-shared-weights
47m 37s
fix(megatron): forward mtp_shared_weights to megatron-bridge citest #13515: Pull request #10025 opened by Excelius-Wang
50m 36s Excelius-Wang:fix/mtp-shared-weights
Excelius-Wang:fix/mtp-shared-weights
50m 36s
fix(megatron): forward mtp_shared_weights to megatron-bridge Lint test #18486: Pull request #10025 opened by Excelius-Wang
1m 38s Excelius-Wang:fix/mtp-shared-weights
Excelius-Wang:fix/mtp-shared-weights
1m 38s
Bug: unexpected behavior with special characters in input Qoder Assistant #535: Issue comment #10024 (comment) created by Groudonbrulant
1s
1s
Fix NaN for full-parameter SFT with DeepSpeed ZeRO-1/2 on torch<2.9 (AdamW int32 flat-partition overflow) citest-npu #3608: Pull request #10023 opened by kiscad
Action required kiscad:main
kiscad:main
Action required
Fix NaN for full-parameter SFT with DeepSpeed ZeRO-1/2 on torch<2.9 (AdamW int32 flat-partition overflow) citest #13514: Pull request #10023 opened by kiscad
Action required kiscad:main
kiscad:main
Action required
Fix NaN for full-parameter SFT with DeepSpeed ZeRO-1/2 on torch<2.9 (AdamW int32 flat-partition overflow) Lint test #18485: Pull request #10023 opened by kiscad
Action required kiscad:main
kiscad:main
Action required
Increasing the IMAGE_MAX_TOKEN_NUM does not increase the number of steps. Qoder Assistant #534: Issue comment #9749 (comment) created by z0o0ey
1s
1s
perf(rlhf): vectorize padding-free log-probability restore citest #13513: Pull request #10021 opened by Ruihan11
14m 27s Ruihan11:perf/rlhf-padding-free-restore
Ruihan11:perf/rlhf-padding-free-restore
14m 27s
perf(rlhf): vectorize padding-free log-probability restore Lint test #18484: Pull request #10021 opened by Ruihan11
1m 18s Ruihan11:perf/rlhf-padding-free-restore
Ruihan11:perf/rlhf-padding-free-restore
1m 18s
perf(rlhf): vectorize padding-free log-probability restore citest-npu #3607: Pull request #10021 opened by Ruihan11
26m 49s Ruihan11:perf/rlhf-padding-free-restore
Ruihan11:perf/rlhf-padding-free-restore
26m 49s
4.3.0版本使用swift infer + vllm backend推理Kimi-k25模型报错”AttributeError: 'KimiK25VisionConfig' object has no attribute 'hidden_size'. Did you mean: 'mm_hidden_size'?“ Qoder Assistant #533: Issue comment #9772 (comment) created by z0o0ey
2s
2s
sft 垂域 文生文 训练中断 Qoder Assistant #532: Issue comment #9788 (comment) created by z0o0ey
1s
1s
多级多卡进行DPO训练时batch为{} Qoder Assistant #531: Issue comment #5601 (comment) created by z0o0ey
2s
2s
使用多卡训练分类模型出现tensor对齐情况 Qoder Assistant #530: Issue comment #5469 (comment) created by z0o0ey
1s
1s
支持训练时在多个验证集上独立评估 Qoder Assistant #529: Issue comment #5467 (comment) created by z0o0ey
1s
1s
LoRA合并qwen2.5-vl-7B模型在 ms-swift 3.6.4 与 3.7 表现差异: 3.7版本合并的模型在多槽位grounding任务上显著优于3.6.4版本合并的模型 Qoder Assistant #528: Issue comment #5430 (comment) created by z0o0ey
1s
1s
DPO 损失函数计算NaN问题查找 Qoder Assistant #527: Issue comment #5418 (comment) created by z0o0ey
2s
2s
同样的配置，启动多次实验，有的实验日志中输出token_acc，有的没有 Qoder Assistant #526: Issue comment #5388 (comment) created by z0o0ey
10s
10s
llava-next-video微调 Qoder Assistant #525: Issue comment #5322 (comment) created by z0o0ey
1s
1s
LoRA batch推理，报错：Length of adapter_names should be the same as the number of inputs, but got 1 and 10 respectively Qoder Assistant #524: Issue comment #5285 (comment) created by z0o0ey
1s
1s
蒸馏Qwen2.5-Omni模型报错：IndexError: max(): Expected reduction dim 1 to have non-zero size. Qoder Assistant #523: Issue comment #4755 (comment) created by z0o0ey
1s
1s