Skip to content

Pull requests: huggingface/candle

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
Assigned to nobody Loading
Sort

Pull requests list

Remove ineffective KV-cache truncation
#3957 opened Sep 4, 2026 by 9r0x Loading…
Specify parquet features to drop redundant deps
#3956 opened Sep 4, 2026 by ivarflakstad Member Loading…
Fix CUDA argsort for large f32 rows
#3955 opened Sep 4, 2026 by sizzlecar Loading…
Fix rotary embedding convention in MiMi transformer
#3949 opened Sep 3, 2026 by jorge-menjivar Contributor Loading…
Add XLMRobertaModel::forward_with_cancel
#3933 opened Aug 31, 2026 by thomnico Loading…
Add quantized Gemma 4 GGUF support
#3932 opened Aug 31, 2026 by gabgiani Loading…
Add qmatmul bench terminal report
#3925 opened Aug 31, 2026 by shergin Loading…
candle Intel GPU / SYCL support
#3923 opened Aug 28, 2026 by Hahihula Loading…
fix(cuda): add context to kernel lookup errors
#3920 opened Aug 28, 2026 by aiqubits Loading…
nomic-bert: use fused Metal SDPA in attention
#3918 opened Aug 27, 2026 by jasonthird Loading…
feat(qwen3_vl): add clear_kv_cache
#3914 opened Aug 24, 2026 by sidd-27 Loading…
Add the LFM2-MoE model (LFM2.5-8B-A1B)
#3910 opened Aug 22, 2026 by Jacqkues Contributor Loading…
fix(kernels): gate __hmax_nan/__hmin_nan shim on toolkit version
#3909 opened Aug 20, 2026 by jorge-menjivar Contributor Loading…
Use cudnn for conv2d backprop
#3905 opened Aug 17, 2026 by EthanAlmloff Contributor Loading…
feat(gemma3): add embed_tokens() and forward_embeds()
#3903 opened Aug 17, 2026 by sidd-27 Loading…
ProTip! Exclude everything labeled bug with -label:bug.