[Bug] Predict method does not finish

[ArlindKadra]

Describe the bug

After training the UIS-RNN algorithm, I call the predict() method with the test_sequences. In this case, the test sequence is a 2-d numpy array, not a list. Somehow the method gets stuck and never finishes.
I checked the input shape and according to the method doc it should work fine.

Maybe I am doing something wrong?

To Reproduce

Commands and arguments

import numpy as np
import uisrnn

SAVED_MODEL_NAME = 'saved_model.uisrnn'


model_args, training_args, inference_args = uisrnn.parse_arguments()
model = uisrnn.UISRNN(model_args)
model.load(SAVED_MODEL_NAME)
test_sequence = np.load('./data/test_sequence.npy', allow_pickle=True)
predicted_cluster_id = model.predict(test_sequence, inference_args)

Logs

Nothing printed.

Serialized models

https://drive.google.com/open?id=1N4JwQQv27Xap-UlvHCNARTiKGT3356AX

Data samples

Data
https://www.kaggle.com/mfekadu/darpa-timit-acousticphonetic-continuous-speech
Embeddings
https://drive.google.com/open?id=1N4JwQQv27Xap-UlvHCNARTiKGT3356AX

Versions

• uis-rnn: 0.1.0
• numpy: 1.18.0
• scipy: 1.4.1
• torch: 1.3.1

Additional context

Debugging the code, I see that everything goes well until this point:

uis-rnn/uisrnn/uisrnn.py

Line 549 in cb3a9c6

for new_beam_rank in range(

The statement after the for loop is finished is not hit.
Using this implementation to generate the embeddings:
https://github.com/HarryVolek/PyTorch_Speaker_Verification

[ArlindKadra]

Further note, preparing everything as a list of sequences works.
If it helps, I can also add the necessary files for this case.

[saumyaborwankar]

Hi @ArlindKadra I am also stuck at the inferencing part could you share the dimensions of the test_sequence variable. Im confused as to what format it requires for inferencing. Thank you

[cfeng16]

Hi @ArlindKadra, I face the same problem now, and I use the AMI corpus dataset. During the training, I concatenate all the training samples as a single numpy matrix, and I prepare the data as a list of sequences during the testing. However, it still can not predict anything, and there are no responses, and I would like to ask if the settings I mentioned above are the same as yours? Thank you!