I run python train.py and get error with File "../utils/alias_multinomial.py", line 57

[thucnt]

Traceback (most recent call last): File "train.py", line 36, in main() File "train.py", line 32, in main save_every=20, grad_clip=5.0 File "../utils/training.py", line 127, in train neg_loss, dirichlet_loss = model(doc_indices, pivot_words, target_words) File "/Users/macbook/anaconda/lib/python3.6/site-packages/torch/nn/modules/module.py", line 325, in call result = self.forward(input, **kwargs) File "../utils/lda2vec_loss.py", line 72, in forward neg_loss = self.neg(pivot_words, target_words, doc_vectors, w) File "/Users/macbook/anaconda/lib/python3.6/site-packages/torch/nn/modules/module.py", line 325, in call result = self.forward(input, **kwargs) File "../utils/lda2vec_loss.py", line 142, in forward noise = self.multinomial.draw(batch_sizewindow_sizeself.num_sampled) File "../utils/alias_multinomial.py", line 57, in draw b = torch.bernoulli(q) RuntimeError: invalid argument 1: must be >= 0 and <= 1 at /Volumes/OSX/Downloads/pytorch/aten/src/TH/THRandom.c:300

[TropComplique]

I just tested the code and didn't get any errors. Have you changed any parameters in the code? Do you have up-to-date pytorch installed?

[thucnt]

Thanks for your response. But I have upgrade the latest pytorch using condo -c soumith pytorch and get a new error:

RuntimeError: invalid argument 1: must be >= 0 and <= 1 at /Users/soumith/miniconda2/conda-bld/pytorch_1503975723910/work/torch/lib/TH/THRandom.c:270

Because I run your code on my computer without GPU so I have to change your code by deleting anything relating to cuda(). Because I have just used pytorch, please tell me is it right?

[TropComplique]

Hi. Have you changed any parameters in the code?

[TropComplique]

Let's debug together. Add to utils/alias_multinomial.py before line 57 (line with b = torch.bernoulli(q)) the following print(q).

[thucnt]

This is result before program raise error: 0.3298 0.5141 0.6557 ⋮
0.3008 0.8198 0.8300 [torch.FloatTensor of size 1075200] main() File "/Volumes/OSX/Users/thucnt/git/lda2vec-pytorch/20newsgroups/train.py", line 32, in main save_every=20, grad_clip=5.0 File "/Volumes/OSX/Users/thucnt/git/lda2vec-pytorch/utils/training.py", line 127, in train neg_loss, dirichlet_loss = model(doc_indices, pivot_words, target_words) File "/Users/macbook/anaconda/envs/pytorch/lib/python3.6/site-packages/torch/nn/modules/module.py", line 224, in call result = self.forward(input, **kwargs) File "/Volumes/OSX/Users/thucnt/git/lda2vec-pytorch/utils/lda2vec_loss.py", line 72, in forward neg_loss = self.neg(pivot_words, target_words, doc_vectors, w) File "/Users/macbook/anaconda/envs/pytorch/lib/python3.6/site-packages/torch/nn/modules/module.py", line 224, in call result = self.forward(input, **kwargs) File "/Volumes/OSX/Users/thucnt/git/lda2vec-pytorch/utils/lda2vec_loss.py", line 142, in forward noise = self.multinomial.draw(batch_sizewindow_sizeself.num_sampled) File "/Volumes/OSX/Users/thucnt/git/lda2vec-pytorch/utils/alias_multinomial.py", line 58, in draw b = torch.bernoulli(q) RuntimeError: invalid argument 1: must be >= 0 and <= 1 at /Users/soumith/miniconda2/conda-bld/pytorch_1503975723910/work/torch/lib/TH/THRandom.c:270

[TropComplique]

Now try print(q.min(), q.max()).

[thucnt]

number of documents: 12829 number of windows: 1023189 number of topics: 25 vocabulary size: 7460 word embedding dim: 50 number of batches: 143

epoch 1 0%| | 0/143 [00:00<?, ?it/s]Min and Max 0.20730416476726532 1.0001881122589111

[TropComplique]

Add q = q.clamp(0.0, 1.0) before b = torch.bernoulli(q). It must solve the problem.

[thucnt]

Wow. A great support!!! Thank you very much. It runs now.

[Oliveche]

Hello， thucnt Would you like to send me a copy of your modified program on CPU, thank you.

[Oliveche]

I want to know If I want to run on CPU, is it ok if I just delete the.Cuda () in the code?

[TropComplique]

@Oliveche I don't recommend to run this on CPU because it will be very slow. But yeah, if you delete cuda in right places the code will run on CPU.

[Oliveche]

Thanks! I have deleted. But there is still the following problem，

number of documents: 12829 number of windows: 1023189 number of topics: 25 vocabulary size: 7460 word embedding dim: 50 number of batches: 143

epoch 1 0%| | 0/143 [00:00<?, ?it/s]

Please help me , thank you very much.