When I put bie path in ktc.kneron_inference instead of onnx path it gives segmentation error

William

When I put bie path in ktc.kneron_inference API instead of onnx path it gives segmentation error for example Littleonnx model which I converted for Kl720

below is the error code

(base) root@edf75761b2af://workspace/william# python kl720_image_inference.py 

/workspace/miniconda/lib/python3.7/site-packages/numpy/__init__.py:156: UserWarning: mkl-service package failed to import, therefore Intel(R) MKL initialization ensuring its correct out-of-the box operation under condition when Gnu OpenMP had already been loaded by Python process is not assured. Please install mkl-service package, see http://github.com/IntelPython/mkl-service

 from . import _distributor_init

Using TensorFlow backend.

Segmentation fault (core dumped)

Please let me know how to fix this error

PowHuang

If you use Vmware , try add CPU and RAM to setting

William

I am using Docker version 20.10.7, build 20.10.7-0ubuntu1~18.04.1

How to fix the issue for docker??

PowHuang

Could you share your kl720_image_inference.py file?

William

# Code for Inference

import ktc

import onnx

from PIL import Image

import numpy as np

def preprocess(input_file):

  image = Image.open(input_file)

  image = image.convert("RGB")

  image = Image.fromarray(np.array(image)[...,::-1])

  img_data = np.array(image.resize((112, 96), Image.BILINEAR)) / 255

  return img_data

input_data = [preprocess("/workspace/examples/LittleNet/pytorch_imgs/Abdullah_0001.png")]

#inf_results = ktc.kneron_inference(input_data, onnx_file="/workspace/examples/LittleNet/LittleNet.onnx", input_names=["data_out"])

#inf_results = ktc.kneron_inference(input_data, nef_file="/data1/batch_compile/models_720.nef", input_names=["data_out"])

inf_results = ktc.kneron_inference(input_data, bie_file="/data1/output.bie", input_names=["data_out"])

print('Section 3 E2E simulator result:')

print(inf_results)

PowHuang

Same as another case. ktc.kneron_inference 720 bie, must add platform=720

inf_results = ktc.kneron_inference(input_data, bie_file="/data1/output.bie", platform=720, input_names=["data_out"])

William

Thanks , Its working now.