onnx模型转成paddle报错

[Jackson8888]

模型用于目标检测

模型来源： 模型vgg16_bn 使用pytorch -> onnx, 在使用onnx转换成paddle时报错

Now translating model from onnx to paddle. model ir_version: 6, op version: 9 shape inferencing ... Warning: Unsupported operator adaptive_avg_pool2d. No schema registered for this operator. Stopping at incomplete shape inference at adaptive_avg_pool2d: x2paddle_143 node inputs: name: "x2paddle_141" type { tensor_type { elem_type: 1 shape { dim { dim_value: 1 } dim { dim_value: 512 } dim { dim_value: 16 } dim { dim_value: 16 } } } }

name: "x2paddle_142" type { tensor_type { elem_type: 7 shape { dim { dim_value: 2 } } } }

node outputs: name: "x2paddle_143" type { tensor_type { elem_type: 0 } }

!!!!!!!!!! Warning: Unsupported operator adaptive_avg_pool2d. No schema registered for this operator. shape inferenced. Traceback (most recent call last): File "/Users/baijunjie/.pyenv/versions/env/bin/x2paddle", line 10, in sys.exit(main()) File "/Users/baijunjie/.pyenv/versions/3.7.3/envs/env/lib/python3.7/site-packages/x2paddle/convert.py", line 264, in main onnx2paddle(args.model, args.save_dir, args.paddle_type) File "/Users/baijunjie/.pyenv/versions/3.7.3/envs/env/lib/python3.7/site-packages/x2paddle/convert.py", line 171, in onnx2paddle model = ONNXDecoder(model_path) File "/Users/baijunjie/.pyenv/versions/3.7.3/envs/env/lib/python3.7/site-packages/x2paddle/decoder/onnx_decoder.py", line 420, in init self.graph = ONNXGraph(onnx_model) File "/Users/baijunjie/.pyenv/versions/3.7.3/envs/env/lib/python3.7/site-packages/x2paddle/decoder/onnx_decoder.py", line 183, in init self.collect_value_infos() File "/Users/baijunjie/.pyenv/versions/3.7.3/envs/env/lib/python3.7/site-packages/x2paddle/decoder/onnx_decoder.py", line 381, in collect_value_infos TENSOR_TYPE_TO_NP_TYPE[item.type.tensor_type.elem_type], KeyError: 0

[SunAhong1993]

方便把onnx模型发送到sunyanfang01@baidu.com吗？

[hong19860320]

我在转onnx的ssd-mobilenetv1模型到paddle模型也遇到这个问题。 ssd_mobilenet_v1_10.onnx.zip

[nissansz]

可以分享一下 vgg16_bn 的onnx模型吗？