Skip to content

Pull requests: InternLM/lmdeploy

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
Assignee
Filter by who’s assigned
Sort

Pull requests list

fix awq tp bug Something isn't working
#3435 opened Apr 15, 2025 by RunningLeon Loading…
optimize internvit
#3433 opened Apr 15, 2025 by caikun-pjlab Loading…
bump version to v0.8.0
#3432 opened Apr 14, 2025 by lvhan028 Loading…
Zmz/prefill without permute by dlblas
#3430 opened Apr 14, 2025 by hellozmz Draft
update fp8 weight loader
#3405 opened Apr 8, 2025 by grimoire Loading…
[WIP]: vl prefix caching
#3389 opened Apr 3, 2025 by RunningLeon Loading…
add twomicrobatch support enhancement New feature or request
#3381 opened Apr 1, 2025 by SHshenhao Loading…
Add Gloo communication to turobmind enhancement New feature or request
#3362 opened Mar 28, 2025 by irexyc Loading…
Create SECURITY.md
#3333 opened Mar 25, 2025 by ybdesire Loading…
Improve turbomind's prefix cache BC-breaking improvement
#3332 opened Mar 25, 2025 by lvhan028 Loading…
6 of 8 tasks
LMDeploy Distserve enhancement New feature or request
#3304 opened Mar 22, 2025 by JimyMa Loading…
add deepseekv3 doc documentation Improvements or additions to documentation WIP
#3265 opened Mar 17, 2025 by CUHKSZzxy Loading…
support loading model with user input params (turbomind) enhancement New feature or request
#3204 opened Mar 3, 2025 by irexyc Loading…
Support GRPO improvement
#3146 opened Feb 16, 2025 by tastelikefeet Loading…
Mutual exclusion for vision encoder & LLM
#3126 opened Feb 10, 2025 by lzhangzz Loading…
add internvl2.5-78b mmmu evaluation for testing
#2980 opened Dec 31, 2024 by lvhan028 Loading…
support Turbomind ep enhancement New feature or request
#2883 opened Dec 12, 2024 by irexyc Loading…
Support Medusa speculative decoding enhancement New feature or request
#2859 opened Dec 5, 2024 by AllentDan Loading…
ProTip! Follow long discussions with comments:>50.