librosa.util.exceptions.ParameterError: Audio buffer is not finite everywhere

[greatforu]

I have trained my own dataset
Tacotron Model = 890k with loss 0.405 and WaveRNN model 1024k with loss value 3.32
Output generated with these models is not audible at all but it was generating something and this output was generated when tacotron is 810k steps trained but now (890k) this is giving me this error

(espnetENV) ubuntu@ubuntu-16:~/drive_a/mayank/fatchord/WaveRNN$ python gen_tacotron.py --input_text "hey there" wavernn
Using device: cuda

Initialising WaveRNN Model...

voc id: mayank_own_data
Trainable Parameters: 4.234M

Initialising Tacotron Model...

tts id: mayank_own_data
Trainable Parameters: 11.088M
+----------+---+--------------+---------+-----------------+----------------+-----------------+
| Tacotron | r | Vocoder Type | WaveRNN | Generation Mode | Target Samples | Overlap Samples |
+----------+---+--------------+---------+-----------------+----------------+-----------------+
| 890k | 2 | WaveRNN | 1024k | Batched | 11000 | 550 |
+----------+---+--------------+---------+-----------------+----------------+-----------------+

| Generating 1/1
| ████████████████ 48000/48400 | Batch Size: 4 | Gen Rate: 1.8kHz | Traceback (most recent call last):
File "gen_tacotron.py", line 165, in
voc_model.generate(m, save_path, batched, hp.voc_target, hp.voc_overlap, hp.mu_law)
File "/home/ubuntu/drive_a/mayank/fatchord/WaveRNN/models/fatchord_version.py", line 260, in generate
save_wav(output, save_path)
File "/home/ubuntu/drive_a/mayank/fatchord/WaveRNN/utils/dsp.py", line 23, in save_wav
librosa.output.write_wav(path, x.astype(np.float32), sr=hp.sample_rate)
File "", line 2, in write_wav
File "/home/ubuntu/drive_a/espnetENV/lib/python3.6/site-packages/librosa/util/decorators.py", line 58, in __wrapper
return func(*args, **kwargs)
File "/home/ubuntu/drive_a/espnetENV/lib/python3.6/site-packages/librosa/output.py", line 239, in write_wav
util.valid_audio(y, mono=False)
File "/home/ubuntu/drive_a/espnetENV/lib/python3.6/site-packages/librosa/util/utils.py", line 275, in valid_audio
raise ParameterError('Audio buffer is not finite everywhere')
librosa.util.exceptions.ParameterError: Audio buffer is not finite everywhere

Any help would be appreciated!
Thanks