harvitronix
commented
6 years ago When you say it got stuck, do you mean it froze? If that's the case, it's likely that you've run out of memory. The LSTM defaults to 2048 units which is quite memory intensive. Reduce it significantly (try just 8 units) to see if that solves the problem, then scale up to the point that it stops working, and use the one before that.
after extracting features by running extract_features.py，then I run train.py,the model is lstm,but when print the input shpe (16,2048),it got stuck at model.add(LSTM(...))，what's wrong?thank you!