-
Notifications
You must be signed in to change notification settings - Fork 187
Pull requests: NVIDIA-NeMo/Automodel
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
fix(fsdp2): guard uninitialized accumulated grads
#2744
opened Jun 23, 2026 by
akoumpa
Contributor
Loading…
3 tasks done
cp: Trigger Testing CICD
ci: address thrift cve bump to 0.23.0 (2736) into r0.5.0
cherry-pick
Run CICD
#2742
opened Jun 23, 2026 by
svcnvidia-nemo-ci
Contributor
Loading…
fix(qwen3-moe): export grouped HF expert weights
r0.5.0
Auto-cherrypick to release branch. Apply before merge; cherrypick happens after merge.
#2741
opened Jun 23, 2026 by
akoumpa
Contributor
Loading…
build: install TileKernels for DeepSeek V4
r0.5.0
Auto-cherrypick to release branch. Apply before merge; cherrypick happens after merge.
#2740
opened Jun 23, 2026 by
akoumpa
Contributor
Loading…
ci: update claude review guidelines
#2739
opened Jun 23, 2026 by
akoumpa
Contributor
Loading…
3 tasks
fix(checkpoint): resolve tie_word_embeddings top-level-first to match HF tying
community-request
#2732
opened Jun 23, 2026 by
Achyuthan-S
Contributor
Loading…
feat(datasets): add DP-aware stateful dataloader
community-request
#2730
opened Jun 23, 2026 by
huahuajhu
Loading…
2 of 3 tasks
fix(ci): reduce mixtral release smoke batch
r0.5.0
Auto-cherrypick to release branch. Apply before merge; cherrypick happens after merge.
#2728
opened Jun 23, 2026 by
akoumpa
Contributor
Loading…
3 tasks done
fix(qwen3_moe): step-0 NaN in MXFP8 packed finetune — expert unload + fused RoPE
r0.5.0
Auto-cherrypick to release branch. Apply before merge; cherrypick happens after merge.
#2722
opened Jun 22, 2026 by
hemildesai
Contributor
Loading…
Security: Potential Path Traversal in Dataset Loading
community-request
#2713
opened Jun 22, 2026 by
tomaioo
Contributor
Loading…
test: add per-test timeout via pytest-timeout
#2710
opened Jun 22, 2026 by
akoumpa
Contributor
Loading…
2 tasks
fix(loss): avoid pkg_resources in linear CE
r0.5.0
Auto-cherrypick to release branch. Apply before merge; cherrypick happens after merge.
#2700
opened Jun 22, 2026 by
akoumpa
Contributor
Loading…
3 tasks done
feat(nemotron_v3): support dense Nemotron-H (Nano 4B)
community-request
#2670
opened Jun 20, 2026 by
stanley1208
Contributor
•
Draft
3 tasks done
fix(vlm): guard validation forward against cuDNN fused-MHA SDPA backend
r0.5.0
Auto-cherrypick to release branch. Apply before merge; cherrypick happens after merge.
fix(deepseek-v4): avoid bf16 -inf overflow in additive attention mask
r0.5.0
Auto-cherrypick to release branch. Apply before merge; cherrypick happens after merge.
#2658
opened Jun 20, 2026 by
akoumpa
Contributor
Loading…
fix(qwen3_5): route dense MTP through SDPA + block-causal mask for pack
r0.5.0
Auto-cherrypick to release branch. Apply before merge; cherrypick happens after merge.
#2656
opened Jun 20, 2026 by
akoumpa
Contributor
Loading…
fix(deepseek-v4): restore batch axis for packed-sequence (THD) forward
r0.5.0
Auto-cherrypick to release branch. Apply before merge; cherrypick happens after merge.
#2651
opened Jun 20, 2026 by
akoumpa
Contributor
Loading…
ci: Update transformers to latest version 5.12.1
#2632
opened Jun 18, 2026 by
svcnvidia-nemo-ci
Contributor
Loading…
fix(checkpoint): write consolidated safetensors without append
community-request
#2627
opened Jun 18, 2026 by
huahuajhu
Loading…
3 tasks done
feat(magi): honor AttnMaskSpec on the HF attention backend
#2622
opened Jun 17, 2026 by
HuiyingLi
Contributor
Loading…
fix(loss): support THD/packed layout in FusedLinearCrossEntropy
r0.5.0
Auto-cherrypick to release branch. Apply before merge; cherrypick happens after merge.
#2615
opened Jun 17, 2026 by
akoumpa
Contributor
Loading…
fix(models): audit fp32 protected tensors
r0.5.0
Auto-cherrypick to release branch. Apply before merge; cherrypick happens after merge.
#2598
opened Jun 16, 2026 by
yuhezhang-ai
Contributor
Loading…
Previous Next
ProTip!
Add no:assignee to see everything that’s not assigned.