search

VeriSilicon / TIM-VX

VeriSilicon Tensor Interface Module
Other
216 stars 85 forks source link

Import question about tvm_vsi version!!! #436

Closed 2050airobert closed 1 year ago

2050airobert commented 1 year ago

Hi, 1 Does tvm vsi_npu version support int8 full quant on the a311d NPU? 2 I mean input is just an onnx model, and then could we covert it with TVM , finally put the result model onto the a311d NPU. 3 Perlay or perchannnel quant supported now ? If not ,when it can ,tks ! 4 Is there any detailed document or weblink about how to run the cnn model with the TVM vsi_npu？ BR

sunshinemyson commented 1 year ago
  1. int8 full quantize can be support.
  2. Not understand
  3. depends on hw version. It can support in latest design.
  4. https://github.com/VeriSilicon/tvm/tree/vsi_npu/tests/python/contrib/test_vsi_npu