用quant_post量化了一个yolov3_darknet的的模型,推理加载出错
Created by: aixier
环境: paddle1.7 出错信息: PaddleCheckError: OP(LoadCombine) fail to open file ..\yolov3_darknet_quant_924_params_, please check whether the mod el file is complete or damaged. at [D:\1.6.1\paddle\paddle/fluid/operators/load_combine_op.h:46]
量化代码
def quantize(): val_reader = mjreader.custom_reader(images_lists, data_dir, input_size,mode) place = fluid.CUDAPlace(0) exe = fluid.Executor(place) quant_post( executor=exe, model_dir='../work/PaddleDetection/yolov3_freeze/yolov3_darknet', quantize_model_path='./yolov3_darknet_quant_924/', sample_generator=val_reader, model_filename='__model__', params_filename='__params__', batch_size=16, batch_nums=20) def main(): quantize()
用quant_post量化了一个yolov3_darknet的的模型,得到的模型如何进行加载推理