Open mxtx0509 opened 6 years ago
when I try to slove the problem I fond that update the rnn could solve the problem .but when I update the rnn ,there is another question . The question is that
/home/mz/torch/install/bin/luajit: /home/mz/torch/install/share/lua/5.1/torch/File.lua:343: unknown Torch class
I am reproducing the code of [deep-soli radar], the same problem you encountered, if rnn uses version [torch/rnn] and appears [attempt to call field 'Recurrent' (a nil value)], replace it with [ Element-Research/rnn] has the same problem as you [rnn/Sequencer.lua:90: attempt to index field 'tableoutput' (a nil value) ]
Did you solve this problem please?
@Acciwu see this,https://github.com/simonwsw/deep-soli/issues/7#issue-218389900,then you will meet "invalid arguments: CudaTensor CudaTensor number",see this https://github.com/simonwsw/deep-soli/issues/5#issue-218385384,then everything is ok
I am so sorry,my issues is that /home/mz/torch/install/bin/luajit: /home/mz/torch/install/share/lua/5.1/rnn/Sequencer.lua:90: attempt to index field 'tableoutput' (a nil value) stack traceback: /home/mz/torch/install/share/lua/5.1/rnn/Sequencer.lua:90: in function 'forward' ./net/rnntrain.lua:28: in function 'batchEval' ./net/train.lua:24: in function 'epochEval' ./net/train.lua:46: in function 'train' net/main.lua:45: in main chunk [C]: in function 'dofile' ...e/mz/torch/install/lib/luarocks/rocks/trepl/scm-1/bin/th:150: in main chunk [C]: at 0x004065d0
I think I don't have the true input ,but i don't know how to do ?could you tell me? the code is as same as provide. @simonwsw @graulef