-
Notifications
You must be signed in to change notification settings - Fork 571
Pull requests: EricLBuehler/mistral.rs
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
Fix: Resolve NaN Logit Memory Leak During Sequential Benchmarks (Issue #2095)
#2108
opened Apr 14, 2026 by
glaziermag
Contributor
Loading…
fix(gguf): safely propagate runtime errors for unknown architectures
#2106
opened Apr 14, 2026 by
glaziermag
Contributor
Loading…
fix(quant): resolve dtype mismatch casting F8E4M3 to BF16 in UnquantLinear (#2072)
#2096
opened Apr 10, 2026 by
glaziermag
Contributor
Loading…
fix(quant): prevent device mismatch in GEMV guard and UnquantLinear forward
#2089
opened Apr 9, 2026 by
Jamesrobertsonldn
Loading…
fix: correct Qwen chunked CUDA index bounds (#1815)
#2083
opened Apr 9, 2026 by
glaziermag
Contributor
Loading…
Fix Gemma softcap F16 overflow NaN and scheduler hang (#2058)
#2076
opened Apr 8, 2026 by
glaziermag
Contributor
Loading…
fix: Idefics3 encoder cache panic when do_image_splitting is enabled
#2074
opened Apr 8, 2026 by
romnn
Loading…
Fix Responses API background=true + stream=true panic (#1945)
#2068
opened Apr 6, 2026 by
glaziermag
Contributor
Loading…
fix(metal): pass --sdk and -std to air-to-metallib link step in build scripts
#2067
opened Apr 6, 2026 by
setoelkahfi
Contributor
Loading…
fix(core): use
from_env for sandboxed apps
#2064
opened Apr 6, 2026 by
setoelkahfi
Contributor
Loading…
Add tensor parallelism support for GDN layers + fix UQFF artifact count
#2054
opened Apr 4, 2026 by
ormandj
Loading…
feat(gguf): add Qwen3.5 (qwen3-next) hybrid MoE GGUF loader
#2049
opened Apr 2, 2026 by
emanueleDiVizio
Loading…
feat(metal): fused MoE expert dispatch with Q4K kernels for Metal
#2048
opened Apr 2, 2026 by
emanueleDiVizio
Loading…
fix(metal): GDN bfloat16, PA scheduler, error handling, MLX SDPA fixes
#2047
opened Apr 2, 2026 by
emanueleDiVizio
Loading…
fix(core): Resolve PagedAttention VRAM leak and scheduler deadlock during OOM
#2045
opened Apr 2, 2026 by
glaziermag
Contributor
Loading…
fix(paged-attn): resolve scheduler queue loop deadlock under memory pressure
#2043
opened Mar 31, 2026 by
glaziermag
Contributor
Loading…
Re-architect FCFS Priorities and Bypass Completion Bucket Discrimination
#2034
opened Mar 28, 2026 by
glaziermag
Contributor
Loading…
Fix PagedAttention Scheduler O(N^2) Thrashing
#2031
opened Mar 26, 2026 by
glaziermag
Contributor
Loading…
Address possible memory leak during model sleep/unload
#2030
opened Mar 25, 2026 by
glaziermag
Contributor
Loading…
Attempt to fix server panic in /re_isq endpoint (#1959)
#2025
opened Mar 25, 2026 by
glaziermag
Contributor
Loading…
chore(deps): bump tar from 0.4.44 to 0.4.45
dependencies
Pull requests that update a dependency file
rust
Pull requests that update rust code
#2014
opened Mar 21, 2026 by
dependabot
bot
Loading…
Previous Next
ProTip!
no:milestone will show everything without a milestone.