tvm

tvm copied to clipboard

I have encountered an issue while using the relay.transform.DivToMul() optimization in TVM. I don’t know why there exists an inconsistency. I think the two methods should yield the same result...

Jupiterghy

[Bug] RPC Server Python Exception Can't Be Send to RPC Client

6

After upgrade with the V0.15.0, we found the RPC have a bug, if the RPC server have some exception, we can't see the error message in RPC client like below....

Johnson9009

[Frontend][PaddlePaddle] Update the export method of PaddlePaddle Softmax

17

## 中文简介(Chinese Introduction) 当PaddlePaddle模型想要将算子offload到cmsis-nn上时，需要通过[cmsisnn.py](https://github.com/apache/tvm/blob/ff3716b83a72c2ff261c492f259e1fcd260600ce/python/tvm/relay/op/contrib/cmsisnn.py#L75)的 **qnn_softmax_pattern()** 来判断是否符合要求。目前将PaddlePaddle转换为cmsis-nn的图时，会将Softmax拆分为几个不同的算子，此时无法被offload到cmsis-nn。 ## English Introduction(英文简介) When a PaddlePaddle model wants to offload operators to CMSIS-NN, it needs to determine if it meets the requirements through **qnn_softmax_pattern()**...

Zheng-Bicheng

[Bug] [VTA, RPC] Can’t upload custom bit file by RPC on ZCU104

1

I am testing VTA in the following environment. Target : ZCU104 (pynq 2.7) Host : ubuntu 20.04 + TVM(v0.16,dev0) xilinx toos : vivado 2020.1 I successfully synthesized the “vta.bit” file...

muonkmu

This PR introduces API of kernel tuning for external runtimes like cuBLAS/cutlass. It contains initial implementation of tuning algorithm for cuBLAS runtime. By default cublas uses heuristic based approach for...

apeskov

[Bug] Tensorization Failure During Multilevel Tiling with Tensor Intrin

1

### Expected behavior MetaSchedule Tuning Works for the given Conv2d workload ### Actual behavior Triggers an error `ValueError: The block no longer exists in the IRModule` during application of schedule...

zxybazh

[CI][AArch64] Enable ONNX and PyTorch tests on AArch64

2

This makes use of work done in enabling PyTorch and ONNX tests to be run on AArch64 to actually enable the tests to run as part of CI. Co-Authored-By: Nicola...

Liam-Sturge

[Relax] Default purity as unknown, instead of pure

4

Prior to this commit, the constructors for `relax.Function` and `relax.FuncStructInfo` provided a default annotation as a pure function. While this could be overridden, it This commit updates the IR to...

Lunderberg

[Tracking Issue] [WebGPU] Supporting DP4A in WebGPU backend

7

### This issue is to track progress for DP4A (new built-in functions introduced in [`packed_4x8_integer_dot_product`](https://gpuweb.github.io/gpuweb/wgsl/#language-extensions-sec)) on WebGPU Backend - [ ] P1. Title of this piece of the feature (PR...

Jiawei-Shao

Bumps [black](https://github.com/psf/black) from 19.10b0 to 24.3.0. Release notes Sourced from black's releases. 24.3.0 Highlights This release is a milestone: it fixes Black's first CVE security vulnerability. If you run Black...

dependabot[bot]