[Question] 为什么使用Bilstm+Crf时，实际有33个类别，但CRF转移矩阵尺寸大小为35*35，导致无法预测
LiangYuHai opened this issue · 0 comments
LiangYuHai commented
如下报错
(0) Invalid argument: Input to reshape is a tensor with 399000 values, but the requested shape has 376200
[[{{node layer_crf_dense/Tensordot}}]]
[[layer_crf/mul/_2999]]
(1) Invalid argument: Input to reshape is a tensor with 399000 values, but the requested shape has 376200
[[{{node layer_crf_dense/Tensordot}}]]
[Paste requirements.txt file here]
检查model_info.json,发现layer_crf的output_dim为33，但转移概率矩阵却是35*35
[A clear and concise description of what you want to know.]