Skip to content

Actions: NVIDIA-NeMo/RL

Actions

Create PR to main with cherry-pick from release

Actions

Loading...
Loading

Show workflow options

Create status badge

Loading
cherry-pick-release-commit.yml
654 workflow runs
654 workflow runs

Filter by Event

Filter by Status

Filter by Branch

Filter by Actor

docs: Add repo overview diagram (#1403) Create PR to main with cherry-pick from release #554: Commit e762237 pushed by terrykong
29s main
main
29s
feat: Overlap param iteration and broadcast in non-colocated refit (#… Create PR to main with cherry-pick from release #553: Commit 73e0c09 pushed by terrykong
23s main
main
23s
fix: Fix policy worker placement when using unified placement group (… Create PR to main with cherry-pick from release #552: Commit d843f02 pushed by terrykong
22s main
main
22s
feat: refit refactoring with zmq and overlapping (#1267) Create PR to main with cherry-pick from release #551: Commit 3a69c21 pushed by terrykong
24s main
main
24s
feat: support truncated importance sampling (#1348) Create PR to main with cherry-pick from release #550: Commit f2de476 pushed by terrykong
23s main
main
23s
feat: Add debugger flag that can be turned on via RAY_DEBUG=legacy (#… Create PR to main with cherry-pick from release #549: Commit f286857 pushed by terrykong
17s main
main
17s
feat: add Megatron support for on-policy distillation (#1324) Create PR to main with cherry-pick from release #548: Commit 73c8725 pushed by terrykong
29s main
main
29s
docs: Update README to include NVIDIA NeMo Framework link (#1392) Create PR to main with cherry-pick from release #547: Commit 9a22e2c pushed by terrykong
12s main
main
12s
docs: update latest news list (#1390) Create PR to main with cherry-pick from release #546: Commit 1b3c12d pushed by terrykong
9s main
main
9s
fix: more robust fp8 rollout metric check (#1307) Create PR to main with cherry-pick from release #545: Commit 85eeb8d pushed by terrykong
13s main
main
13s
fix: fix mcore train_iters in grpo (#1383) Create PR to main with cherry-pick from release #544: Commit 905a224 pushed by terrykong
18s main
main
18s
feat: Support DAPO dynamic sampling and reward shaping (#602) Create PR to main with cherry-pick from release #543: Commit 7bd853a pushed by terrykong
18s main
main
18s
fix: Fix non-colocated refit when vLLM model parallel size is larger … Create PR to main with cherry-pick from release #542: Commit dee3fd9 pushed by terrykong
23s main
main
23s
fix: update the custom vllm instructions (#1116) Create PR to main with cherry-pick from release #541: Commit 9da0317 pushed by terrykong
14s main
main
14s
test: Update on-policy distillation release tests (#1363) Create PR to main with cherry-pick from release #540: Commit 638bc52 pushed by terrykong
15s main
main
15s
fix: Fix the logger error in non-colocated sync-grpo code path (#1355) Create PR to main with cherry-pick from release #539: Commit 96656c3 pushed by terrykong
23s main
main
23s
fix: Megatron worker to have locked dependencies (#1315) Create PR to main with cherry-pick from release #538: Commit 8f6e00e pushed by terrykong
14s main
main
14s
fix: grpo early exit edge case (#1361) Create PR to main with cherry-pick from release #537: Commit 0a769cc pushed by terrykong
21s main
main
21s
fix: Replace decode-based prefix matching with EOS-boundary splicing … Create PR to main with cherry-pick from release #536: Commit 5c67023 pushed by terrykong
15s main
main
15s
chore: add chat_template_kwargs in default train configs (#1353) Create PR to main with cherry-pick from release #535: Commit 15a0343 pushed by terrykong
16s main
main
16s
perf: Add num_workers in DPO, GRPO and SFT for loading data (#1314) Create PR to main with cherry-pick from release #534: Commit 355aa98 pushed by terrykong
18s main
main
18s
feat: Add Penguin env (#1327) Create PR to main with cherry-pick from release #533: Commit 4db1704 pushed by terrykong
11s main
main
11s
fix: Fix checkpoint conversion error for qwen 30b-a3b (#1335) Create PR to main with cherry-pick from release #532: Commit eb5bb0f pushed by terrykong
13s main
main
13s
ci: add a descriptive error message for the no-test state (#1318) Create PR to main with cherry-pick from release #531: Commit 6d1d711 pushed by terrykong
10s main
main
10s
test: disable dpo mistral nightly until transformers upgrades past 4.… Create PR to main with cherry-pick from release #530: Commit 53129d4 pushed by terrykong
14s main
main
14s