Created by: aixier

CPU 8 RAM 32GB GPU v100 显存 16GB 磁盘 100GB 环境配置 Python版本 python3.7 框架版本 PaddlePaddle 1.7.0 @slf12 aistudio@jupyter-115786-193843:~/post_training_quantization_withdata$ sh run_post_training_quanzation.sh -------------------args---------------------- algo: KL batch_nums: 20 batch_size: 3000 is_full_quantize: False model_dir: ../work/PaddleDetection_1/yolov3_dark_freeze/mj_yolov3_darknet model_filename: None params_filename: None save_model_path: yolov3_int8_model use_gpu: True

W0314 12:14:31.721148 637 device_context.cc:237] Please NOTE: device: 0, CUDA Capability: 70, Driver API Version: 10.1, Runtime API Version: 9.0 W0314 12:14:31.725704 637 device_context.cc:245] device: 0, cuDNN Version: 7.3. 2020-03-14 12:14:34,203-INFO: all run batch: 0 2020-03-14 12:14:34,203-INFO: all run batch: 0 2020-03-14 12:14:34,203-INFO: calculate scale factor ... 2020-03-14 12:14:34,203-INFO: calculate scale factor ... Traceback (most recent call last): File "post_training_quantization.py", line 66, in batch_nums=10) File "/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddleslim/quant/quanter.py", line 306, in quant_post post_training_quantization.quantize() File "/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddle/fluid/contrib/slim/quantization/post_training_quantization.py", line 231, in quantize self._calculate_scale_factor() File "/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddle/fluid/contrib/slim/quantization/post_training_quantization.py", line 353, in _calculate_scale_factor data = self._sampling_data[var_name] KeyError: 'stage.3.7.0.conv.weights'