-
Notifications
You must be signed in to change notification settings - Fork 216
Pull requests: radixark/miles
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
rocm: disable gradient_accumulation_fusion on gfx950 in test_qwen2.5_0.5B_gsm8k_async
#1160
opened May 19, 2026 by
sreerohi
Loading…
rocm: disable gradient_accumulation_fusion on gfx950 in test_qwen2.5_0.5B_gsm8k
#1159
opened May 19, 2026 by
sreerohi
Loading…
rocm: disable gradient_accumulation_fusion on gfx950 in test_sglang_config_mixed_offload_ft
#1158
opened May 19, 2026 by
sreerohi
Loading…
rocm: disable gradient_accumulation_fusion on gfx950 in test_sglang_config_mixed_offload
#1157
opened May 19, 2026 by
sreerohi
Loading…
rocm: disable gradient_accumulation_fusion on gfx950 in test_sglang_config
#1156
opened May 19, 2026 by
sreerohi
Loading…
rocm: disable gradient_accumulation_fusion on gfx950 in test_qwen2.5_0.5B_gsm8k_async_short
#1155
opened May 19, 2026 by
sreerohi
Loading…
rocm: disable gradient_accumulation_fusion on gfx950 in test_qwen2.5_0.5B_gsm8k_short
#1154
opened May 19, 2026 by
sreerohi
Loading…
SendRecv-Broadcast-Based Disaggregated Weight Sync
#1152
opened May 19, 2026 by
zyzshishui
Contributor
•
Draft
fix: expert_overlap and pp issues in nemotron-3-super
#1150
opened May 19, 2026 by
Zhichenzzz
Contributor
Loading…
[CI] Add dense true-on-policy E2E gate
run-ci-true-on-policy
Run dense true-on-policy E2E CI
#1147
opened May 18, 2026 by
maocheng23
Contributor
Loading…
[DO NOT MERGE] test ci-image
run-ci-image
#1146
opened May 18, 2026 by
maocheng23
Contributor
Loading…
[TITO] Refactor/session verify args consume miles params Namespace
run-ci-sglang
#1142
opened May 16, 2026 by
guapisolo
Collaborator
Loading…
feat: add chunked log-probs operator from hidden states
#1136
opened May 15, 2026 by
Yukinanaa
Loading…
Fix
PYTHONBUFFERED typo: should be PYTHONUNBUFFERED
#1135
opened May 15, 2026 by
timgianitsos
Loading…
fix: support Qwen3-VL Megatron training inputs
#1133
opened May 14, 2026 by
nikhilbarhate99
Loading…
[loss refactor] [3] file restructure and rename
run-ci-megatron
#1132
opened May 14, 2026 by
yueming-yuan
Collaborator
Loading…
[TITO] deepseek v32 support
run-ci-sglang
#1131
opened May 14, 2026 by
guapisolo
Collaborator
Loading…
utils/ppo_utils: fix OPSM mask using sequence-level advantage instead of per-token comparison
#1127
opened May 14, 2026 by
Dev-X25874
Loading…
Previous Next
ProTip!
Mix and match filters to narrow down what you’re looking for.