Created by: liangruofei

12 %run tools/train.py -c configs/ppyolo/ppyolo_r18vd.yml
--eval 运行时长: 92毫秒 结束时间: 2020-08-17 20:21:11 ---------------------------------------------------------------------------Exception Traceback (most recent call last)~/PaddleDetection-0.4/tools/train.py in 366 help='The profiler output file path. (used for benchmark)') 367 FLAGS = parser.parse_args() --> 368 main() ~/PaddleDetection-0.4/tools/train.py in main() 102 with fluid.program_guard(train_prog, startup_prog): 103 with fluid.unique_name.guard(): --> 104 model = create(main_arch) 105 if FLAGS.fp16: 106 assert (getattr(model.backbone, 'norm_type', None) ~/PaddleDetection-0.4/ppdet/core/workspace.py in create(cls_or_name, **kwargs) 248 target = global_config[target_key] 249 if isinstance(target, SchemaDict): --> 250 kwargs[k] = create(target_key) 251 elif hasattr(target, 'dict'): # serialized object 252 kwargs[k] = target ~/PaddleDetection-0.4/ppdet/core/workspace.py in create(cls_or_name, **kwargs) 256 # (e.g., list, dict) from within the created module instances 257 kwargs = copy.deepcopy(kwargs) --> 258 return cls(**kwargs) ~/PaddleDetection-0.4/ppdet/modeling/anchor_heads/yolo_head.py in init(self, conv_block_num, norm_decay, num_classes, anchors, anchor_masks, drop_block, coord_conv, iou_aware, iou_aware_factor, block_size, keep_prob, yolo_loss, spp, nms, weight_prefix_name, downsample, scale_x_y, clip_bbox) 78 scale_x_y=1.0, 79 clip_bbox=True): ---> 80 check_version("1.8.4") 81 self.conv_block_num = conv_block_num 82 self.norm_decay = norm_decay ~/PaddleDetection-0.4/ppdet/utils/check.py in check_version(version) 75 return 76 if version_installed[i] < version_split[i]: ---> 77 raise Exception(err) 78 79 Exception: PaddlePaddle version 1.8.4 or higher is required, or a suitable develop version is satisfied as well. Please make sure the version is good with your code. 而我用的aitudio的系统是1.8.0，新建也没有1.8.4的