Issue while running train.py

[supriyakarambe]

Hi, when I am trying to run train .py , I am getting following error.

Input 'bytes' of 'DecodeRaw' Op has type float32 that does not match expected type of string.

Would you please help me to resolve this error

[qizhao1]

I also have met this problem.

[supriyakarambe]

Are you able to resolve that issue??

[redoc700]

Did you solve the issue? Any update will be helpful.

[kushalps]

In data_load.py, add function: def load_npy(filename): filename = filename.decode("utf-8") return np.load(filename)

in same file at line 124-126, call this function instead of lambda, i.e. like mel = tf.py_func(load_pny, ......)

This worked for me.

[redoc700]

thanks!

On Fri, Aug 3, 2018 at 11:59 AM, kushalps [email protected] wrote:

In data_load.py, add function: def load_npy(filename): filename = filename.decode("utf-8") return np.load(filename)

in same file at line 124-126, call this function instead of lambda, i.e. like mel = tf.py_func(load_pny, ......)

This worked for me.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/Kyubyong/deepvoice3/issues/18#issuecomment-410158120, or mute the thread https://github.com/notifications/unsubscribe-auth/Ak3oMct_L7rHFfjFsh51NGcUbvOdqbHJks5uM-3dgaJpZM4TCuQg .

[redoc700]

Can you send me sniped of modified code bcz I didn't find any lamda function In data_load.py as u mentioned

On Fri 3 Aug, 2018, 10:32 PM Pratap Verma, [email protected] wrote:

thanks!

On Fri, Aug 3, 2018 at 11:59 AM, kushalps [email protected] wrote:

In data_load.py, add function: def load_npy(filename): filename = filename.decode("utf-8") return np.load(filename)

in same file at line 124-126, call this function instead of lambda, i.e. like mel = tf.py_func(load_pny, ......)

This worked for me.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/Kyubyong/deepvoice3/issues/18#issuecomment-410158120, or mute the thread https://github.com/notifications/unsubscribe-auth/Ak3oMct_L7rHFfjFsh51NGcUbvOdqbHJks5uM-3dgaJpZM4TCuQg .

[Tortoise17]

Dear friend.

mel = tf.py_func(load_pny, ......)

can you share, what is after load_npy, ... as I have syntex error still. Please if you can share the complete three lines.

[guyco87]

Its just a typo, this is why you get syntex error - instead of load_pny you should write load_npy.

This is how lines 124-126 should look like after you correct the typo:

mel = tf.py_func(load_npy, [mel], tf.float32) # (None, n_mels) done = tf.py_func(load_npy, [done], tf.int32) # (None,) mag = tf.py_func(load_npy, [mag], tf.float32) # (None, 1+n_fft/2)