-
Notifications
You must be signed in to change notification settings - Fork 836
Pull requests: apple/coremltools
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
Add support for PyTorch non-finite predicates
#2834
opened Aug 24, 2026 by
binaydhakal
Contributor
Loading…
Add torch.cdist to the PyTorch converter
#2832
opened Aug 24, 2026 by
binaydhakal
Contributor
Loading…
Fix: shape ops are never folded into quantized weights, leaving them mirrored vs dense
#2831
opened Aug 24, 2026 by
kasper0406
Contributor
Loading…
Use FindPython3 consistently for source builds
#2830
opened Aug 23, 2026 by
cdeil
Contributor
Loading…
Acquire the GIL when releasing NumPy-backed Core ML inputs
#2829
opened Aug 23, 2026 by
cdeil
Contributor
Loading…
Fix weight_id collision when materializing multiple source functions
#2828
opened Aug 23, 2026 by
kasper0406
Contributor
Loading…
Fix SIGSEGV: release prediction input arrays with the GIL held
#2827
opened Aug 21, 2026 by
kasper0406
Contributor
Loading…
Fix rms_norm eps inflation from max-|x| rescale; guard all-zero input
#2822
opened Aug 19, 2026 by
nschlaepfer
Loading…
Add InstanceNorm3d alias-conversion regression coverage
#2820
opened Aug 17, 2026 by
sfeng01-tech
Loading…
Support PyTorch angle and trig ops (deg2rad, rad2deg, hypot)
#2817
opened Aug 15, 2026 by
jaseem-kurikkal
Contributor
Loading…
Decompose Softplus in TF frontend to avoid fp16 overflow on ANE
#2814
opened Aug 10, 2026 by
axiom-of-choice
Loading…
Read torch rand's dtype from the argument that holds it
#2813
opened Aug 9, 2026 by
LeSingh1
Contributor
Loading…
Fall back instead of baking a symbolic dim into softplus and nll_loss
#2812
opened Aug 9, 2026 by
LeSingh1
Contributor
Loading…
State the limitation when chunk / unbind meet a dynamic dim
#2811
opened Aug 9, 2026 by
LeSingh1
Contributor
Loading…
Use the first dim of size 3 when torch.cross has no dim
#2810
opened Aug 9, 2026 by
LeSingh1
Contributor
Loading…
Tell apart torch overloads that TorchScript reports under one node kind
#2809
opened Aug 9, 2026 by
LeSingh1
Contributor
Loading…
Support a symbolic batch dim in torch narrow and group_norm
#2808
opened Aug 9, 2026 by
LeSingh1
Contributor
Loading…
fuse_layernorm_or_instancenorm: check gamma / beta shapes, not just their rank
#2807
opened Aug 9, 2026 by
LeSingh1
Contributor
Loading…
Bound gather_nd / scatter_nd indices by the dims they actually address
#2806
opened Aug 9, 2026 by
LeSingh1
Contributor
Loading…
const_deduplication: compare constexpr inputs by value, not by str(ndarray)
#2805
opened Aug 9, 2026 by
LeSingh1
Contributor
Loading…
loop_invariant_elimination: require the body to return the loop var itself
#2804
opened Aug 9, 2026 by
LeSingh1
Contributor
Loading…
Make squeeze and slice_by_size const folding agree with their type inference
#2803
opened Aug 9, 2026 by
LeSingh1
Contributor
Loading…
Const fold scatter_along_axis and scaled_dot_product_attention correctly
#2802
opened Aug 9, 2026 by
LeSingh1
Contributor
Loading…
Previous Next
ProTip!
What’s not been updated in a month: updated:<2026-07-25.