Skip to content

Pull requests: nod-ai/shark-ai

New pull request New
38 Open 695 Closed
38 Open 695 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
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

Bump IREE requirement pins to their latest versions.
#867 opened Jan 24, 2025 by shark-pr-automator bot Loading…
4
[sharktank][llama] Adds parity to hf's rotary_embedding layer and a test to maintain it
#863 opened Jan 24, 2025 by dan-garvey Loading…
1
[Sharktank][Llama][FP8] Minimal changes for numerically correct fp8
#859 opened Jan 22, 2025 by dan-garvey • Draft
[Llama] Kv cache new layout
#858 opened Jan 22, 2025 by Groverkss • Draft
[Llama] Do not allow configurable partitions for KVCache
#856 opened Jan 21, 2025 by Groverkss Loading…
3
[shortfin_apps/llm] Add locks to protect trie nodes, node reference counts, and the trie structure itself from race conditions
#838 opened Jan 17, 2025 by renxida • Draft
Bump iree dependencies forward to include barrier changes
#834 opened Jan 16, 2025 by rsuderman Loading…
4
[llama] Generate causal mask better
#832 opened Jan 16, 2025 by Groverkss • Draft
1
Add barrier for cross device synchronization
#830 opened Jan 15, 2025 by rsuderman Loading…
Added experimental mm padding for cache behavior
#825 opened Jan 14, 2025 by rsuderman Loading…
1
Llm attn op override
#805 opened Jan 9, 2025 by dan-garvey Loading…
Add [pinned, unpinnned] matrix to ci_eval.yaml.
#767 opened Jan 7, 2025 by ScottTodd • Draft
5
Re-raise exception after logging it in SDXL test.
#766 opened Jan 6, 2025 by ScottTodd • Draft
1
Adds FLUX.1 service to shortfin apps.
#742 opened Jan 3, 2025 by monorimet Loading…
Implement script to cleanup user-facing commands
#733 opened Jan 1, 2025 by vinayakdsci Loading…
47
Fixed mixed precision contraction semantics for mmt_block_scaled_offset_q4_unsigned
#720 opened Dec 20, 2024 by Groverkss Loading…
Enable tokenizers in shortfin packages on Linux x86_64.
#688 opened Dec 12, 2024 by ScottTodd Loading…
1
3
Expanded sharded support for alternative sharding mechanisms
#680 opened Dec 12, 2024 by rsuderman Loading…
3
Fix iree_hal_buffer_subspan after change in iree.
#671 opened Dec 10, 2024 by AWoloszyn Loading…
Convert ci-shark-ai.yml to pkgci_shark_ai.yml.
#625 opened Nov 27, 2024 by ScottTodd • Draft
Create pkgci.yml and pkgci_build_packages.yml.
#589 opened Nov 22, 2024 by ScottTodd • Draft
8
[test] Change punet int8 model version
#569 opened Nov 19, 2024 by nithinsubbiah Loading…
[shortfin] Implement async alloc/dealloc of buffers.
#507 opened Nov 14, 2024 by stellaraccident Loading…
[shortfin] Add port specifier to send_request.py
#500 opened Nov 13, 2024 by AWoloszyn Loading…
Add IREE numerics test for Llama 3.1 8B FP16 TP8
#394 opened Oct 31, 2024 by sogartar Loading…
1
ProTip! Add no:assignee to see everything that’s not assigned.