Closed GenTxt closed 7 years ago
That part is not fully tested. Could you try with a very small dataset?
Sung
On Wed, Jan 18, 2017 at 8:48 PM, GenTxt notifications@github.com wrote:
Hi:
Training and standard sampling working perfectly in tensorflow 12.1 on 64 bit ubuntu with gpu.
Problem occurs when choosing "beam search pick" for sampling. After 15 minutes of trying to generate default 200 words the following warning appears:
/home/aaron/Desktop/word-rnn-tensorflow/beam.py:26: RuntimeWarning: divide by zero encountered in log cand_scores = np.array(live_scores)[:, None] - np.log(self.probs)
25 minutes still no terminal output or redirect to text file.
Any ideas how I can fix this?
Thanks
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/hunkim/word-rnn-tensorflow/issues/49, or mute the thread https://github.com/notifications/unsubscribe-auth/AA3DV8NsbfBUUfG4ksofLLOqi5L1sIBJks5rTgohgaJpZM4Lm2Fj .
Original corpus was 6.3 megs. Trained with tinyshakespeare 1.1 megs.
Script generated random prime word (which did not happen before) but 20 minutes later there was still no terminal output for 200 words.
Previous warning message did not appear.
Original reference: https://github.com/hunkim/word-rnn-tensorflow/pull/29
I've attempted a solution in #51.
I've pulled @victorkwan solution into master with latest update. Please check and open new issue if problems persist.
Hi:
Training and standard sampling working perfectly in tensorflow 12.1 on 64 bit ubuntu with gpu.
Problem occurs when choosing "beam search pick" for sampling. After 15 minutes of trying to generate default 200 words the following warning appears:
/home/aaron/Desktop/word-rnn-tensorflow/beam.py:26: RuntimeWarning: divide by zero encountered in log cand_scores = np.array(live_scores)[:, None] - np.log(self.probs)
25 minutes still no terminal output or redirect to text file.
Any ideas how I can fix this?
Thanks