Repository Issue Activity (beta)

#2825 SIGSEGV in Espresso cast_kernel_cpu / vDSP_vflt32 for transformer text encoders with a dynamic sequence dimension (static shape works)

#2823 Removed `torch.ao.quantization.quantizer` imports in coremltools/optimize/torch break on torch 2.11 — request for decision on removal/deprecation of CoreMLQuantizer

#2826 TorchConverter: states=[] cannot express "this model has no states"

#2821 rms_norm translation inflates eps by max(|x|)^2 — percent-level errors on spiky activations even at FLOAT32 / CPU_ONLY

#2340 Support BICUBIC interpolation for image resizing

#1747 Implement PyTorch op `upsample_bicubic2d`

#2666 CoreML CPU InstanceNorm3d behaves differently depending on track_running_stats

#2772 avg_pool1d/avg_pool2d with ceil_mode=True silently return incorrect values

#2586 SKLearn regression model with multiple outputs

#2579 Unsupported a / x.shape[0]

#2565 coremltools gives incorrect output for torch.ops.aten.diagonal.default

#2226 Please add support for `torch.tensor_split`

#2761 noop_elimination silently drops the floor of floor_div(x, 1) for float inputs

#2646 Support Python 3.14

#2764 EnumeratedShapes converts with ClassifierConfig

#2751 Map MIL ops back to their originating nn.Module for debugging and profiling

#2625 Large output errors when converting MobileNetV3 in FP16

#2687 Softplus on Apple Neural Engine has a hard fp16 discontinuity at x ≈ 10.4 (output drops to 0)

#2759 cumsum with exclusive=True crashes when constant-folded (TypeError: Cannot construct a dtype from an array)

#2130 Input error during converting torch.fft.irfft

#2628 MIL pow op for 2.0 to 0.0 is 0.0

#1793 RuntimeError: PyTorch convert function for op 'torchvision::roi_align' not implemented.

#2758 predict() with CPU_AND_NE silently SIGKILLs the process on a large fully-ANE-eligible mlprogram (ANECCompile failure not surfaced at load)

#1889 Support for deform_conv2d operation from PyTorch

#2757 MLComputePlan.load_from_path() aborts the whole process (uncatchable SIGABRT) when given an .mlpackage path