kendryte / nncase

Open deep learning compiler stack for Kendryte AI accelerators ✨
Apache License 2.0
748 stars 181 forks source link

RuntimeError 没有任何提示 #671

Closed Nagisakurumi closed 2 years ago

Nagisakurumi commented 2 years ago

BUG : Traceback (most recent call last): File ".\export.py", line 32, in convert.convert() File "F:\workspace\python\目标检测\detection\convert_model\convert_base.py", line 112, in convert k510model = self.convert_to_k510model(samplify_onnx_model) File "F:\workspace\python\目标检测\detection\convert_model\convert_base.py", line 90, in convert_to_k510model compiler = nncase.Compiler(compile_options) RuntimeError

compile_options

compile_options : (编译参数) compile_options = nncase.CompileOptions() compile_options.target = 'k510' compile_options.input_type = 'float32' # or 'uint8' 'int8' compile_options.output_type = 'float32' # or 'uint8' 'int8'. Only work in PTQ compile_options.output_range = [] # Only work in PTQ and output type is not "float32" compile_options.preprocess = True # if False, the args below will unworked compile_options.swapRB = True compile_options.input_shape = self.input_shape # keep layout same as input layout [1, 3, 320, 320] compile_options.input_layout = 'NCHW' compile_options.model_layout = '' # default is empty. Specific it when tflite model with "NCHW" layout and Onnx(Caffe) model with "NHWC" layout compile_options.mean = [0, 0, 0] compile_options.std = [1, 1, 1] compile_options.input_range = [0, 1] compile_options.letterbox_value = 0. # pad what you want compile_options.dump_ir = True compile_options.dump_asm = True compile_options.dump_dir = 'tmp'

compiler = nncase.Compiler(compile_options)

sunnycase commented 2 years ago

请按issue模板填写。

sunnycase commented 2 years ago

请按issue模板填写。