write-rnn-tensorflow

write-rnn-tensorflow copied to clipboard

FileNotFoundError: [Errno 2] No such file or directory: './data/strokes_training_data.cpkl' error

1

sbswarna

Generated samples look bad, but training seems ok

12

I'm getting crappy-looking samples from networks that seem to train properly. Loss converges nicely (see last screenshot), but samples don't compare with @hardmaru's results. Here are some examples... Any ideas...

grisaitis

why loss is inf? how to repair?

Instantnoodles-madman

u = expand(expand(np.array([i for i in range(args.U)], dtype=np.float32), 0, args.K), 0, args.batch_size)Shapes (5, 1, 41) and () are incompatible

1

Instantnoodles-madman

number of cells should be args.num_layers

1

yaylinda

NotFoundError, Tensor name [...] not found in checkpoint files save/model.ckpt-11000

3

First, thanks a lot for this very cool code! Running the pretrained model with the suggested command: python sample.py --filename example_name --sample_length 1000 Produces this error: WARNING:tensorflow:: Using a concatenated...

lk251

Better to provide a requirements.txt

1

I think it's better to provide a `requirements.txt`, indicating the python libraries and version it needed .

drcege

Hi, thanks for the nice work. In the paper, the training loss can go down to -1000. I can only get to -5. I wonder how your training loss curve...

leinxx

ValueError: Trying to share variable rnnlm/multi_rnn_cell/cell_0/basic_lstm_cell/kernel, but specified shape (512, 1024) and found shape (259, 1024).

3

When I run train.py, got this errors. I think is kind of version problems, my tf version is 1.3 File "/home/lxt/tf_project/HyperNetwork/write-rnn-tensorflow/model.py", line 50, in __init__ outputs, state_out = tf.contrib.legacy_seq2seq.rnn_decoder(inputs, self.state_in,...

lxtGH

When i tried to run train.py file..... I got the following error **"IOError: [Errno 2] No such file or directory: './data/strokes_training_data.cpkl'** Please provide some suggestions

anushaGundapaneni