[TVM Discuss] [Questions] Is there any speed comparison of quantization on cpu

> [[topi] add ARM v8.2 udot (uint8) support > #3978](https://github.com/apache/incubator-tvm/pull/3978) This works if you have a machine/device with ARM v8.2 and DOT instruction. Rasp3b and 4b don't have it. --- [Visit Topic](https://discuss.tvm.ai/t/is-there-any-speed-comparison-of-quan

[TVM Discuss] [Questions] WSL Windows 10 subsystem Ubuntu 18.04

It works with some caveats. WSL doesn't give any hw device access, so things like CUDA won't work and are constrained to CPU only. --- [Visit Topic](https://discuss.tvm.ai/t/wsl-windows-10-subsystem-ubuntu-18-04/6388/2) to respond. You are receiving this because you enabled mailing list

[TVM Discuss] [Questions] Nnvm issue when trying to convert pytorch conv2d to nnvm conv2d

nnvm._base.NNVMError: Required parameter channels of int is not presented, in operator conv2d(name="", kernel_size="(5, 5)", strides="(1, 1)", out_layout="NCHW") --- [Visit Topic](https://discuss.tvm.ai/t/nnvm-issue-when-trying-to-convert-pytorch-conv2d-to-nnvm-conv2d/6389/1) to respond.

[TVM Discuss] [Questions] Is there any speed comparison of quantization on cpu

@anijain2305 @masahi [ [topi] add ARM v8.2 udot (uint8) support #3978](https://github.com/apache/incubator-tvm/pull/3978) as this commit said, arm platform support udot(uint8), can I reckon that arm can achieve int8-speedup for udot(uint8) support, then what is the right open method? -

[TVM Discuss] [Questions] Is there any speed comparison of quantization on cpu

@anijain2305 thanks a lot. I thought the tvm relay quantize is the same as tvm model converted from pre-quantized. I also test tvm-int8 model from pytorch qat model, the speed is the same as tvm-relay-quantize-int8 model. I really have no idea how to get 1.3x -1.5x speedup no matter pre-q

[TVM Discuss] [Questions] WSL Windows 10 subsystem Ubuntu 18.04

Hi there, Does anyone use WSL Windows 10 subsystem Ubuntu 18.04 to run TVM well? Thank you! Mason --- [Visit Topic](https://discuss.tvm.ai/t/wsl-windows-10-subsystem-ubuntu-18-04/6388/1) to respond. You are receiving this because you enabled mailing list mode. To unsubscribe from thes

[TVM Discuss] [Questions] How to generate docs locally

Hi, How do I generate the [docs](https://tvm.apache.org/docs/) locally. I would like to build the docs of a particular tag or version. --- [Visit Topic](https://discuss.tvm.ai/t/how-to-generate-docs-locally/6387/1) to respond. You are receiving this because you enabled mailing list mode.

[TVM Discuss] [Questions] Is there any speed comparison of quantization on cpu

I have mostly worked on pre-quantized models. So, I cant comment on the performance of Relay quantized model through ARM. There might be few missing pieces there. I am planning to write a tutorial by next week on how to read pre-quantized models from TFLite. You can also try @masahi tutorial

[TVM Discuss] [Questions] Incorrect generated function after PartitionGraph pass

This should be resolved by this PR: https://github.com/apache/incubator-tvm/pull/5320 :) --- [Visit Topic](https://discuss.tvm.ai/t/incorrect-generated-function-after-partitiongraph-pass/6380/2) to respond. You are receiving this because you enabled mailing list mode. To unsubscribe fro

[TVM Discuss] [Questions] Incorrect generated function after PartitionGraph pass

Wow, perfect timing! Thanks :) --- [Visit Topic](https://discuss.tvm.ai/t/incorrect-generated-function-after-partitiongraph-pass/6380/3) to respond. You are receiving this because you enabled mailing list mode. To unsubscribe from these emails, [click here](https://discuss.tvm.ai/email/

[TVM Discuss] [Questions] Incorrect generated function after PartitionGraph pass

Hi @zhiics @comaniac, I am using BYOC to offload transformers to external codegen tools. These transformers are composite functions. I had been using this feature well with my manually-generated annotation passes, but when I merge the latest changes to go through the `AnnotateGraph -> Partiti

[TVM Discuss] [Questions] Is there any speed comparison of quantization on cpu

[quote="anijain2305, post:27, topic:6256, full:true"] For rasp3 and rasp4, we saw 1.3x - 1.5x performance speedup going from FP32 to Int8. The link comparing QNNPACK and TVM is not upstream'd yet. If I understand correctly, it will be sometime before the authors of that work will be able to m

[TVM Discuss] [Questions] Can we schedule bunch of OPS in CPU and other some in GPU ? while running inference in TVM

Hi Expert, I have just started looking into the TVM framework. I am exploring possibilities like how do we get best latency numbers using TVM. As a part of this I wanted to know that, is there anyway user can attached device info per OPS? Also can use create multiple graphs (like one with Obj

[TVM Discuss] [Questions] Dynamic input shape

relay.frontend.from_onnx supports dynamic input_shape ? ```Python relay.frontend.from_onnx (onnx_model, shape=shape_dict) ``` where the shape_dict should be given，otherwise，How to set the params for different shapes, because mytest onnx model is support for dynamic shape. Thanks a lot! -

[TVM Discuss] [Questions] How to map nn.conv2d to VTA?

I'm studying the VTA design and how it is being mapped to TVM. The resnet18 tutorial is good, however, the resnet18 itself is too complicated to follow. Instead, I'm trying with a simple nn.conv2d + nn.relu network as below: ``` def conv2d(data, weight=None, **kwargs): name = kwargs.get("n