-
Notifications
You must be signed in to change notification settings - Fork 7.4k
OpenSep 30, 2026
No due date
•Last updated 47% complete
List view
0 of 21 selected 0 issues of 21 selected
[feature] implement TeaCache
performanceAnything related to performance improvements, profiling and benchmarkingAnything related to performance improvements, profiling and benchmarkingroadmapAdd to current release roadmapAdd to current release roadmapStatus: Open.#12589 In huggingface/diffusers;- Status: Open.#14705 In huggingface/diffusers;
GGUF: enable_sequential_cpu_offload() raises KeyError: None — any GGUF model larger than VRAM is unrunnable
bugSomething isn't workingSomething isn't workingggufIssues and PRs related to GGUFIssues and PRs related to GGUFStatus: Open.#14691 In huggingface/diffusers;RMSNorm crashes on NPU when elementwise_affine=False (weight=None): npu_rms_norm requires a real gamma tensor
bugSomething isn't workingSomething isn't workingStatus: Open.#14590 In huggingface/diffusers;- Status: Open.#14533 In huggingface/diffusers;
Add SANA-WM camera-controlled image-to-video pipeline
documentationImprovements or additions to documentationImprovements or additions to documentationno-issue-neededfor PRs that do not require link to an issuefor PRs that do not require link to an issuesize/LPR with diff > 200 LOCPR with diff > 200 LOCStatus: Open (in progress).huggingface/diffusersnumber 13881#13881 In huggingface/diffusers;[Modular]improve loop block
documentationImprovements or additions to documentationImprovements or additions to documentationsize/LPR with diff > 200 LOCPR with diff > 200 LOCStatus: Open (in progress).[modular] improve auto offload
documentationImprovements or additions to documentationImprovements or additions to documentationsize/LPR with diff > 200 LOCPR with diff > 200 LOCStatus: Open (in progress).[modular] LTX-2.5: two-stage generation as one pipeline
documentationImprovements or additions to documentationImprovements or additions to documentationsize/LPR with diff > 200 LOCPR with diff > 200 LOCStatus: Open (in progress).- Status: Open.#14316 In huggingface/diffusers;
Add Echo modular pipeline
documentationImprovements or additions to documentationImprovements or additions to documentationsize/LPR with diff > 200 LOCPR with diff > 200 LOCStatus: Open (in progress).huggingface/diffusersnumber 14696#14696 In huggingface/diffusers;Consolidate torch device backend dispatch
documentationImprovements or additions to documentationImprovements or additions to documentationsize/LPR with diff > 200 LOCPR with diff > 200 LOCStatus: Open (in progress).[core] Shard tensor-parallel checkpoints on load and save
documentationImprovements or additions to documentationImprovements or additions to documentationsize/LPR with diff > 200 LOCPR with diff > 200 LOCStatus: Open (in progress).huggingface/diffusersnumber 14544#14544 In huggingface/diffusers;[Modular] Avoid downloading weights when loading from an existing local path
documentationImprovements or additions to documentationImprovements or additions to documentationsize/MPR with diff < 200 LOCPR with diff < 200 LOCStatus: Open (in progress).Add single file support for Minimax H3
size/MPR with diff < 200 LOCPR with diff < 200 LOCStatus: Open (in progress).Allow tensor parallelism and context parallelism in one ParallelConfig
no-issue-neededfor PRs that do not require link to an issuefor PRs that do not require link to an issuesize/LPR with diff > 200 LOCPR with diff > 200 LOCStatus: Open (in progress).chore: add one additional copied from in qwenimage 2.1 vae
size/SPR with diff < 50 LOCPR with diff < 50 LOCStatus: Open (in progress).huggingface/diffusersnumber 14810#14810 In huggingface/diffusers;[tests] refactor the pipeline-level attention processor tests
size/LPR with diff > 200 LOCPR with diff > 200 LOCStatus: Open (in progress).[tests] remove
test_pipelines_common.pysize/LPR with diff > 200 LOCPR with diff > 200 LOCStatus: Open (in progress).feat: cp support in qwenimage 2.1
size/MPR with diff < 200 LOCPR with diff < 200 LOCStatus: Open (in progress).huggingface/diffusersnumber 14817#14817 In huggingface/diffusers;Add single file support for Krea 2
documentationImprovements or additions to documentationImprovements or additions to documentationsize/MPR with diff < 200 LOCPR with diff < 200 LOCStatus: Open (in progress).