executorch
executorch copied to clipboard
pytorch
On-device AI across mobile, embedded and edge for PyTorch
By default outputs are re-ordered to correct order during TOSA lowering. However this is seen as a workaround as it should not be needed. Furthermore the output issue is not...
cc @freddan80 @per @zingo @oscarandersson8218 @digantdesai
### 🐛 Describe the bug Hello, I am trying to run a YOLO-NAS model on a GPU using a Vulkan backend. I can lower and export the model to run...
### Summary Add support for delegating view_copy in the XNNPACK delegate via the XNN static_reshape operator. This includes support for up to one dynamic dimension. It also includes conditional support...
This PR is auto-generated nightly by [this action](https://github.com/pytorch/executorch/blob/main/.github/workflows/nightly.yml). Update the pinned pytorch hash.
Summary: Check that self.dim() is smaller than kTensorDimensionLimit, otherwise we may write past the end of the out tensor. Differential Revision: D86356264
Summary: For model hand_tracking_dpe_v26 we have this metadata string that is being deserialized. "L__self__,,on_device_ai.helios.compiler_tests.e2e_test_gen.configs.hand_track ing_dpe_v26.hand_tracking_dpe_v26.Test_dpe_v26_dpefeedforward_only_por_qconfig _executorch;140317725696144,mods,nimble.ml_libs.problems.mview.models.dpe_infe rence_v2.DPEFeedForward;140317579086400,getattr(getattr(L['self'].mods._image_ encoder._image_backbone._layers, '0'), '3'),torch.nn.modules.pooling.MaxPool2d" This string has multiple entries separated by ";" with format...
Summary: `set_all_to_value(out_data, out_step_len, value)` loops over the padding and copies a row at a time. Add some checks to make sure we do not exceed the out tensor (out_data) size...
Differential Revision: D86588171