search

rolczynski / Automatic-Speech-Recognition

🎧 Automatic Speech Recognition: DeepSpeech & Seq2Seq (TensorFlow)
GNU Affero General Public License v3.0
223 stars 64 forks source link

shape problem #17

Closed Fakhraddin closed 4 years ago

Fakhraddin commented 4 years ago

I have an error when I run the basic,py 'ValueError: Error when checking input: expected X to have 3 dimensions, but got array with shape (41440, 1)' error.

rolczynski commented 4 years ago

Hey, sorry for the late response. Have you resolved this problem?

Fakhraddin commented 4 years ago

Hi ,I haven't yet.

Drazcat commented 4 years ago

same issue here, the weird part is that every time i try to train, the shape duplicates its value. i started with a shape(20720, 1), and ended with an input shape in the millions. Its a very weird bug and first time i see it

rolczynski commented 4 years ago

Solved - I had to change the wrap preprocess in the pipeline.