You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Looking at the "many_to_many" example for RNNs, I don't think this is a true many-to-many architecture. The input is many (None, timesteps, num_input) but the output is not many (None, num_classes). I think the output should have the shape (None, timesteps, num_classes) for this to actually be a many-to-many model. Notice how the output for the many-to-one example has the same dimensionality as the many-to-many output.
The text was updated successfully, but these errors were encountered:
Nhìn vào ví dụ "many_to_many" cho RNN, tôi không nghĩ đây là một kiến trúc nhiều-nhiều thực sự. Đầu vào là nhiều (None, timesteps, num_input) nhưng đầu ra không nhiều (None, num_classes). Tôi nghĩ rằng đầu ra nên có hình dạng (Không có, lần lượt, num_classes) để đây thực sự là một mô hình nhiều-nhiều. Lưu ý cách đầu ra cho ví dụ nhiều-một có cùng chiều với đầu ra nhiều-nhiều.
Looking at the "many_to_many" example for RNNs, I don't think this is a true many-to-many architecture. The input is many (None, timesteps, num_input) but the output is not many (None, num_classes). I think the output should have the shape (None, timesteps, num_classes) for this to actually be a many-to-many model. Notice how the output for the many-to-one example has the same dimensionality as the many-to-many output.
The text was updated successfully, but these errors were encountered: