Jian Guo
Apr 25, 2018, 7:41:36 AM4/25/18
to Caffe Users
Hi, I'm new to caffe and I want to apply LSTM to my deep learning project. But I am puzzled about the blob's shape of the input and output. I have two questions here.
1. Suppose I have a blob of (T, N, m) as input, and I set the num_output of recurrent_param to 200, then what's the shape of the LSTM's output? Is it (T, N, 200) or something else?
2. Is it possible to pass a four dimension al blob to the LSTM as input? If it could, then what is the shape of the output?
1. Suppose I have a blob of (T, N, m) as input, and I set the num_output of recurrent_param to 200, then what's the shape of the LSTM's output? Is it (T, N, 200) or something else?
2. Is it possible to pass a four dimension al blob to the LSTM as input? If it could, then what is the shape of the output?
Reply all
Reply to author
Forward
0 new messages