|
需求:
和开发沟通,需要帮忙提交一个redmine进行跟踪
现象:
在pytorch模型生成rknn时,开启量化do_quantization=True时报错,当do_quantization=False时正常。
报错日志:
D Process convolution_at_1639_204_204 ...
D RKNN output shape(convolution): (10 256 160 4)
D Process attach_convolution_at_1639_204_out0_214 ...
D RKNN output shape(output): (10 256 160 4)
I Build FairMOT complete.
E Catch exception when building RKNN model!
E Traceback (most recent call last):
E File "rknn/api/rknn_base.py", line 578, in rknn.api.rknn_base.RKNNBase.build
E File "rknn/api/rknn_base.py", line 1509, in rknn.api.rknn_base.RKNNBase._quantize2
E File "rknn/base/RKNNlib/RK_nn.py", line 546, in rknn.base.RKNNlib.RK_nn.RKnn.quantize
E File "rknn/base/RKNNlib/app/medusa/quantization.py", line 129, in rknn.base.RKNNlib.app.medusa.quantization.Quantization.run
E File "rknn/base/RKNNlib/app/medusa/quantization.py", line 51, in rknn.base.RKNNlib.app.medusa.quantization.Quantization._run_quantization
E File "rknn/base/RKNNlib/range/analyze_range.py", line 405, in rknn.base.RKNNlib.range.analyze_range.sparse_network
E File "rknn/base/RKNNlib/range/analyze_range.py", line 313, in rknn.base.RKNNlib.range.analyze_range.analyze_operation
E File "rknn/base/RKNNlib/range/analyze_range.py", line 290, in rknn.base.RKNNlib.range.analyze_range._a_add
E ValueError: operands could not be broadcast together with shapes (4096,) (64,)
Build FairMOT failed!
完整日志:附件
|
本帖子中包含更多资源
您需要 登录 才可以下载或查看,没有帐号?立即注册
x
|