Stuck on an issue?

Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.

And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.

Error related to text mask

See original GitHub issue

File "/fsx/dalle2/.dalle_env_38/lib64/python3.8/site-packages/torch/nn/modules/module.py", line 1130, in _call_impl
losses = self.p_losses(unet, image, times, image_embed = image_embed, text_encodings = text_encodings, text_mask = text_mask, lowres_cond_img = lowres_cond_img, predict_x_start = predict_x_start, learned_variance = learned_variance, is_latent_diffusion = is_latent_diffusion, noise_scheduler = noise_scheduler)
File "/fsx/dalle2/DALLE2-pytorch/dalle2_pytorch/dalle2_pytorch.py", line 2340, in p_losses
return forward_call(*input, **kwargs)
File "/fsx/dalle2/DALLE2-pytorch/dalle2_pytorch/dalle2_pytorch.py", line 1817, in forward
model_output = unet(
File "/fsx/dalle2/.dalle_env_38/lib64/python3.8/site-packages/torch/nn/modules/module.py", line 1130, in _call_impl
text_keep_mask = text_mask & text_keep_mask
RuntimeError: The size of tensor a (0) must match the size of tensor b (26) at non-singleton dimension 0
return forward_call(*input, **kwargs)
File "/fsx/dalle2/DALLE2-pytorch/dalle2_pytorch/dalle2_pytorch.py", line 1817, in forward
text_keep_mask = text_mask & text_keep_mask
RuntimeError: The size of tensor a (0) must match the size of tensor b (26) at non-singleton dimension 0

Recently introduced but not sure which commit

Issue Analytics

State:
Created a year ago
Comments:49 (22 by maintainers)

Top GitHub Comments

1reaction

nousrcommented, Jul 13, 2022

awesome! ill give it a shot tomorrow, i’m gonna sign off for the afternoon.

On Jul 12, 2022, at 5:51 PM, Phil Wang @.***> wrote:

@nousr threw in yet another stability measure 349aaca

worst comes to worst, we can always rollback

thanks for letting me know about the dreaded NaNs!

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.

1reaction

lucidrainscommented, Jul 13, 2022

@nousr threw in yet another stability measure https://github.com/lucidrains/DALLE2-pytorch/commit/349aaca56fe66a6f0fc6720c91ee5d7fa1e36f93

worst comes to worst, we can always rollback

thanks for letting me know about the dreaded NaNs!