Skip to content

Pull requests: nod-ai/shark-ai

New pull request New
115 Open 1,193 Closed
115 Open 1,193 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Assigned to nobody Loading
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

[shortfin] Fix runner label for nightly workflow.
#1556 opened Jun 2, 2025 by ScottTodd Loading…
[sharktank] in device creation add arg to specify tensor parallelism
#1554 opened Jun 2, 2025 by sogartar Loading…
Bump IREE requirement pins to 3.5.0rc20250602
#1553 opened Jun 2, 2025 by shark-pr-automator bot Loading…
@amd-vivekag
remove redundant object
#1552 opened Jun 1, 2025 by dan-garvey Loading…
2
[shortfin] Implement run_in_executor for PyWorkerEventLoop
#1551 opened Jun 1, 2025 by vinayakdsci Loading…
6
[WIP][sharktank] Add topk hip kernel support
#1550 opened May 31, 2025 by jinchen62 Loading…
[sharktank] Fix deepseek sharding
#1547 opened May 30, 2025 by archana-ramalingam Loading…
2
Refactor Token Selection
#1546 opened May 30, 2025 by stbaione Loading…
@stbaione
[sharktank] Fix pipeline parallelism perplexity regression for toy deepseek
#1545 opened May 30, 2025 by archana-ramalingam • Draft
Add Conv1/3DLayer and LayerNorm with normalized_shape for Wan2.1 model
#1541 opened May 30, 2025 by AmosLewis Loading…
[sharktank] Fix which prefill logits eager perplexity calculations are using
#1540 opened May 30, 2025 by Alex-Vasile Loading…
1 task
1
Remove argmax and topk functions from PagedLlmModelV1
#1536 opened May 29, 2025 by stbaione Loading…
Re-enable and fix rotary.py
#1534 opened May 29, 2025 by rsuderman Loading…
[RoPE] Stick to hf RoPE implementation
#1532 opened May 29, 2025 by Groverkss • Draft
[mlir_kernel] Use mlir_kernel for injecting Wave kernels
#1530 opened May 29, 2025 by aviator19941 Loading…
5
Re-enable the quark parity test
#1518 opened May 27, 2025 by KyleHerndon Loading…
[SharkTank] exporting prefill llama4 toy model to mlir
#1517 opened May 27, 2025 by oyazdanb • Draft
4
Cache Storage to avoid allocate pinned memory for every prefill step and every decode step
#1516 opened May 27, 2025 by dezhiAmd Loading…
6
Thread pool token selection enhancement New feature or request
#1515 opened May 27, 2025 by stbaione Loading…
@stbaione
Handle more datatypes gracefully in the dump_gguf tool
#1508 opened May 22, 2025 by KyleHerndon Loading…
1
Fix multi hypothesis streaming
#1496 opened May 21, 2025 by zphoenixrises • Draft
[sharktank] Add missing argument device affinity for tensor-parallel Resnet block test
#1490 opened May 21, 2025 by sogartar • Draft
Create CODEOWNERS
#1478 opened May 20, 2025 by Groverkss Loading…
Integrate MLIR kernel for KV cache gather
#1468 opened May 16, 2025 by rsuderman • Draft
[Shortfin][LLM] Add initial support for disaggregated invocations
#1463 opened May 16, 2025 by vinayakdsci • Draft
ProTip! Add no:assignee to see everything that’s not assigned.