Skip to content

Actions: THUDM/slime

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
[Question]将qwen3.5-27B转换为torch dist后，再转换回hf格式，使用sglang部署，在tp=4/8的情况下模型会输出且只输出大量‘！’ Slash Command Handler #963: Issue comment #2222 (comment) created by albaNnaksqr
8s
8s
[Fix] Support origin HF weight backfilling in parallel converter PR Test #2777: Pull request #2270 opened by albaNnaksqr
5m 6s albaNnaksqr:fix/parallel-add-missing-weights
albaNnaksqr:fix/parallel-add-missing-weights
5m 6s
[Fix] Support origin HF weight backfilling in parallel converter pre-commit #4794: Pull request #2270 opened by albaNnaksqr
1m 15s albaNnaksqr:fix/parallel-add-missing-weights
albaNnaksqr:fix/parallel-add-missing-weights
1m 15s
[Fix] Support origin HF weight backfilling in parallel converter conda CI #4028: Pull request #2270 opened by albaNnaksqr
1s albaNnaksqr:fix/parallel-add-missing-weights
albaNnaksqr:fix/parallel-add-missing-weights
1s
Per-comm memory probe (available_memory/memory_stats) is O(#allocator segments) and slows weight sync & training under expandable_segments Slash Command Handler #962: Issue comment #2268 (comment) created by yszhli
9s
9s
Fix O(#segments) per-comm memory probe that slows weight sync & training under expandable_segments pre-commit #4793: Pull request #2269 synchronize by yszhli
1m 17s yszhli:fix/per-comm-memory-probe-perf
yszhli:fix/per-comm-memory-probe-perf
1m 17s
Fix O(#segments) per-comm memory probe that slows weight sync & training under expandable_segments PR Test #2776: Pull request #2269 synchronize by yszhli
3m 43s yszhli:fix/per-comm-memory-probe-perf
yszhli:fix/per-comm-memory-probe-perf
3m 43s
Fix O(#segments) per-comm memory probe that slows weight sync & training under expandable_segments conda CI #4027: Pull request #2269 synchronize by yszhli
1s yszhli:fix/per-comm-memory-probe-perf
yszhli:fix/per-comm-memory-probe-perf
1s
Fix O(#segments) per-comm memory probe that slows weight sync & training under expandable_segments pre-commit #4792: Pull request #2269 opened by yszhli
2m 4s yszhli:fix/per-comm-memory-probe-perf
yszhli:fix/per-comm-memory-probe-perf
2m 4s
Fix O(#segments) per-comm memory probe that slows weight sync & training under expandable_segments PR Test #2775: Pull request #2269 opened by yszhli
3m 29s yszhli:fix/per-comm-memory-probe-perf
yszhli:fix/per-comm-memory-probe-perf
3m 29s
Fix O(#segments) per-comm memory probe that slows weight sync & training under expandable_segments conda CI #4026: Pull request #2269 opened by yszhli
Skipped yszhli:fix/per-comm-memory-probe-perf
yszhli:fix/per-comm-memory-probe-perf
Skipped
Fix model convert when use latest megatron conda CI #4025: Pull request #2267 opened by alexqdh
1s alexqdh:fix_model_convert_use_latest_megatron
alexqdh:fix_model_convert_use_latest_megatron
1s
Fix model convert when use latest megatron pre-commit #4791: Pull request #2267 opened by alexqdh
48s alexqdh:fix_model_convert_use_latest_megatron
alexqdh:fix_model_convert_use_latest_megatron
48s
Fix model convert when use latest megatron PR Test #2774: Pull request #2267 opened by alexqdh
6m 4s alexqdh:fix_model_convert_use_latest_megatron
alexqdh:fix_model_convert_use_latest_megatron
6m 4s
pages build and deployment pages-build-deployment #239: by github-pages Bot
1m 41s gh-pages
gh-pages
1m 41s
Refactor --save-debug-train-data (#2266) Release Documentation #239: Commit 681b3ad pushed by zhuzilin
57s main
main
57s
Refactor --save-debug-train-data (#2266) PR Test #2773: Commit 681b3ad pushed by zhuzilin
4m 19s main
main
4m 19s
Refactor --save-debug-train-data (#2266) pre-commit #4790: Commit 681b3ad pushed by zhuzilin
1m 32s main
main
1m 32s
Refactor --save-debug-train-data PR Test #2772: Pull request #2266 labeled by zhuzilin
5m 23s zilin/save_debug_train_darta
zilin/save_debug_train_darta
5m 23s
Refactor --save-debug-train-data PR Test #2771: Pull request #2266 opened by zhuzilin
16s zilin/save_debug_train_darta
zilin/save_debug_train_darta
16s
Refactor --save-debug-train-data pre-commit #4789: Pull request #2266 opened by zhuzilin
1m 9s zilin/save_debug_train_darta
zilin/save_debug_train_darta
1m 9s
Refactor --save-debug-train-data conda CI #4024: Pull request #2266 opened by zhuzilin
1s zilin/save_debug_train_darta
zilin/save_debug_train_darta
1s
perf: vectorize REINFORCE++ discounted returns Slash Command Handler #961: Issue comment #2205 (comment) created by morluto
7s
7s
fix(npu): bracket IPv6 hosts in distributed init methods (#2199) pre-commit #4788: Commit d38dc29 pushed by zhuzilin
1m 17s main
main
1m 17s
fix(npu): bracket IPv6 hosts in distributed init methods (#2199) PR Test #2770: Commit d38dc29 pushed by zhuzilin
7m 5s main
main
7m 5s