PaddlePaddle
自己训练的推理模型，图片验证的时候出现了这样的错误?
欢迎您反馈PaddleHub使用问题，非常感谢您对PaddleHub的贡献！ 在留下您的问题时，辛苦您同步提供如下信息：
- 版本、环境信息 1）PaddleHub和PaddlePaddle版本：请提供您的PaddleHub和PaddlePaddle版本号，例如PaddleHub1.4.1，PaddlePaddle1.6.2 2）系统环境：请您描述系统类型，例如Linux/Windows/MacOS/，python版本
- 复现信息：如为报错，请给出复现环境、复现步骤 paddle 1.8.5 paddlehub 1.7.1
D:\soft_path\Anaconda\lib\site-packages\win32\lib\pywintypes.py:3: DeprecationWarning: the imp module is deprecated in favour of importlib; see the module's documentation for alternative uses
import sys, os,imp
W0315 15:06:41.720773 21248 analysis_predictor.cc:1059] Deprecated. Please use CreatePredictor instead.
e[37m--- fused 0 scale with matmule[0m
e[37m--- Fused 0 ReshapeTransposeMatmulMkldnn patternse[0m
e[37m--- Fused 0 ReshapeTransposeMatmulMkldnn patterns with transpose's xshapee[0m
e[37m--- Fused 0 ReshapeTransposeMatmulMkldnn patterns with reshape's xshapee[0m
e[37m--- Fused 0 ReshapeTransposeMatmulMkldnn patterns with reshape's xshape with transpose's xshapee[0m
e[37m--- Fused 0 MatmulTransposeReshape patternse[0m
e[37m--- fused 0 batch norm with relu activatione[0m
Traceback (most recent call last):
File "E:/Projection/unit3/predict_module/demo2.py", line 11, in
File "D:\soft_path\Anaconda\lib\site-packages\paddle\fluid\framework.py", line 2604, in append_op
op = Operator(
File "D:\soft_path\Anaconda\lib\site-packages\paddle\fluid\layer_helper.py", line 43, in append_op
return self.main_program.current_block().append_op(*args, **kwargs)
File "D:\soft_path\Anaconda\lib\site-packages\paddle\fluid\layers\nn.py", line 2922, in conv2d
helper.append_op(
File "E:\Face_Detection\PaddleSeg-release-v0.7.0\PaddleSeg-release-v0.7.0\contrib\HumanSeg\nets\libs.py", line 128, in conv
return fluid.layers.conv2d(*args, **kargs)
File "E:\Face_Detection\PaddleSeg-release-v0.7.0\PaddleSeg-release-v0.7.0\contrib\HumanSeg\nets\libs.py", line 153, in separate_conv
input = conv(
File "E:\Face_Detection\PaddleSeg-release-v0.7.0\PaddleSeg-release-v0.7.0\contrib\HumanSeg\nets\backbone\xception.py", line 268, in xception_block
data = separate_conv(
File "E:\Face_Detection\PaddleSeg-release-v0.7.0\PaddleSeg-release-v0.7.0\contrib\HumanSeg\nets\backbone\xception.py", line 199, in middle_flow
data, short_cuts = self.xception_block(
File "E:\Face_Detection\PaddleSeg-release-v0.7.0\PaddleSeg-release-v0.7.0\contrib\HumanSeg\nets\backbone\xception.py", line 81, in __call__
data = self.middle_flow(data)
File "E:\Face_Detection\PaddleSeg-release-v0.7.0\PaddleSeg-release-v0.7.0\contrib\HumanSeg\nets\deeplabv3p.py", line 366, in build_net
data, decode_shortcuts = backbone_net(image)
File "E:\Face_Detection\PaddleSeg-release-v0.7.0\PaddleSeg-release-v0.7.0\contrib\HumanSeg\models\humanseg.py", line 852, in build_net
model_out = model.build_net(inputs)
File "E:\Face_Detection\PaddleSeg-release-v0.7.0\PaddleSeg-release-v0.7.0\contrib\HumanSeg\models\load_model.py", line 43, in load_model
model.test_inputs, model.test_outputs = model.build_net(
File "export.py", line 37, in export
model = models.load_model(args.model_dir)
File "export.py", line 43, in <module>
export(args)
ResourceExhaustedError: Fail to alloc memory of 263380608 size.
[Hint: p should not be null.] (at D:\v2.0.0\paddle\paddle\fluid\memory\detail\system_allocator.cc:69)
[operator < conv2d > error]
Process finished with exit code 1
希望寻求帮助
您好，请问是使用paddleseg时候有问题吗，如果是，请移步到paddleseg提issue，感谢您的配合https://github.com/PaddlePaddle/PaddleSeg
您可以到百度aistudio上上传，您的报错信息是不是内存或者显存不足导致的，建议您在百度aistudio上建立一个复现环境，若该问题不是由于paddlehub提供的模型导致的，建议您到paddle相应的repo下提出问题，以便得到更好更专业的回复。 从您目前的报错信息来看，我的建议是：
para_state_dict = paddle.load(args.model_dir)
model_state_dict = model.state_dict()
keys = model_state_dict.keys()
num_params_loaded = 0
for k in keys:
if k not in para_state_dict:
print("{} is not in pretrained model".format(k))
elif list(para_state_dict[k].shape) != list(
model_state_dict[k].shape):
print("[SKIP] Shape of pretrained params {} doesn't match.(Pretrained: {}, Actual: {})"
.format(k, para_state_dict[k].shape,
model_state_dict[k].shape))
else:
model_state_dict[k] = para_state_dict[k]
num_params_loaded += 1
model.set_dict(model_state_dict)
print('load model success')
看下模型和参数load是否存在不匹配问题
