-
Notifications
You must be signed in to change notification settings - Fork 1.7k
Pull requests: huggingface/candle
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
fix(pickle): avoid unbounded allocation in BinUnicode string read (CWE-770)
#3961
opened Sep 7, 2026 by
li-jin-quan
Loading…
Fix silently incorrect CPU matmul for a stride-zero broadcast batch
#3960
opened Sep 6, 2026 by
akashjainn
Loading…
Specify parquet features to drop redundant deps
#3956
opened Sep 4, 2026 by
ivarflakstad
Member
Loading…
Fix rotary embedding convention in MiMi transformer
#3949
opened Sep 3, 2026 by
jorge-menjivar
Contributor
Loading…
Fix CUDA copy_strided_src over-copying from an offset source view
#3944
opened Sep 2, 2026 by
PCfVW
Loading…
fix: honor decoupled head_dim in quantized_llama and quantized_qwen2
#3917
opened Aug 27, 2026 by
rexlunae
Loading…
feat(cuda): enable runtime dynamic loading and gate MoE cudart linking
#3916
opened Aug 25, 2026 by
Sadik00789
Loading…
fix(onnx): Tile with a zero repeat must empty the dimension
#3913
opened Aug 24, 2026 by
shoemoney
Loading…
fix(quantized): QTensor::quantize ignores layout offset on a narrowed-then-contiguous view
#3912
opened Aug 23, 2026 by
philphauler
Loading…
fix: align F8E4M3 CUDA kernel names with DType::as_str()
#3911
opened Aug 22, 2026 by
philphauler
Loading…
fix(kernels): gate __hmax_nan/__hmin_nan shim on toolkit version
#3909
opened Aug 20, 2026 by
jorge-menjivar
Contributor
Loading…
Add fused chunked-scan kernels for gated-DeltaNet prefill
#3906
opened Aug 18, 2026 by
alytaphoenix
Loading…
fix(gemma3): make value_states contiguous before the KV cache append
#3902
opened Aug 17, 2026 by
sidd-27
Loading…
Previous Next
ProTip!
Exclude everything labeled
bug with -label:bug.