-
Notifications
You must be signed in to change notification settings - Fork 265
Pull requests: NVIDIA-NeMo/Automodel
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
test(ci): report Jensen-Shannon checkpoint parity metrics
#3620
opened Aug 22, 2026 by
yuhezhang-ai
Contributor
•
Draft
3 tasks done
perf(checkpoint): bound quantized DCP load memory
#3619
opened Aug 21, 2026 by
yuhezhang-ai
Contributor
•
4/4
•
Draft
perf(checkpoint): load standard HF safetensors with DCP
#3616
opened Aug 21, 2026 by
yuhezhang-ai
Contributor
•
3/4
•
Draft
build(deps): AUT-1625 bump transformers to 5.15.1
Run CICD
Trigger Testing CICD
#3615
opened Aug 21, 2026 by
svcnemo-autobot
Contributor
Loading…
test(pp): add deepseek_v4 PP and EP parity tests [2/4]
#3613
opened Aug 21, 2026 by
athitten
Contributor
Loading…
perf(checkpoint): avoid full CPU loads for single-GPU MoE
#3610
opened Aug 21, 2026 by
yuhezhang-ai
Contributor
•
2/4
•
Draft
feat(tokenization): add optional gigatoken tokenizer backend
community-request
#3603
opened Aug 21, 2026 by
aminehd
Contributor
Loading…
fix(retrieval): support Sentence Transformers v6 metadata
#3600
opened Aug 20, 2026 by
oliverholworthy
Contributor
Loading…
fix(perf): recognize RTX PRO 6000 for MFU
community-request
#3598
opened Aug 20, 2026 by
ralovets
Loading…
3 tasks done
fix(checkpoint): load base weights under DDP and MegatronFSDP
community-request
waiting-on-customer
Waiting on the original author to respond
#3597
opened Aug 20, 2026 by
ralovets
Loading…
2 of 3 tasks
feat(qwen image): dmd2 preprocess fixes
#3593
opened Aug 20, 2026 by
linnanwang
Contributor
Loading…
2 of 3 tasks
refactor(checkpoint): make the target-module rename hook a StateDictAdapter method
community-request
#3591
opened Aug 19, 2026 by
stanley1208
Contributor
Loading…
fix(retrieval): do not lose metric records on preemption or error
community-request
waiting-on-customer
Waiting on the original author to respond
#3583
opened Aug 19, 2026 by
sahel-sh
Contributor
Loading…
1 of 3 tasks
perf(checkpoint): reduce allocating grouped MoE load overhead
#3580
opened Aug 18, 2026 by
yuhezhang-ai
Contributor
•
1/4
Loading…
feat(diffusion): periodic validation loss during training
community-request
#3579
opened Aug 18, 2026 by
beccohov
Contributor
Loading…
3 tasks done
refactor(packing): replace global HF monkeypatch with typed metadata
community-request
#3577
opened Aug 18, 2026 by
beccohov
Contributor
Loading…
2 of 3 tasks
fix(vlm): support packed THD Step3.7 pipeline training
community-request
waiting-on-customer
Waiting on the original author to respond
#3572
opened Aug 18, 2026 by
new-TonyWang
Loading…
fix(mtp): generalize context parallelism across MTP models
#3570
opened Aug 18, 2026 by
HuiyingLi
Contributor
Loading…
test(checkpoint): expand parity metrics and phase coverage
#3567
opened Aug 17, 2026 by
yuhezhang-ai
Contributor
Loading…
Previous Next
ProTip!
Filter pull requests by the default branch with base:main.