search

pclucas14 / pixel-cnn-pp

Pytorch Implementation of OpenAI's PixelCNN++
Other
345 stars 76 forks source link

Pre-trained Model Request #3

Closed lxuniverse closed 6 years ago

lxuniverse commented 6 years ago

Hi:

Thanks for your amazing job.

Is it possible that you give the access of your pre-trained model?

Thanks!

pclucas14 commented 6 years ago

Hi,

sorry for the late reply, I just saw this. I will try to upload a few, just need to figure out what is the best way to proceed.

-Lucas

pclucas14 commented 6 years ago

Let me know if this works for you https://mega.nz/#F!W7IhST7R!PV7Pbet8Q07GxVLGnmQrZg

mmazeika commented 6 years ago

When I use load_state_dict instead of load_part_of_model, I get the error

KeyError: 'unexpected key "module.down_layers.0.u_stream.0.conv_input.conv.bias" in state_dict'

load_part_of_model doesn't load anything for me. Was the model code changed at some point?

mmazeika commented 6 years ago

Problem solved. There needed to be a DataParallel wrapper around the model in order to load the parameters.

rsbandhu commented 4 years ago

Hi Lucas, thanks for posting some of the pretrained models. Based on your code, it appears that the pretrained models are after 789, 889 epochs. Is that correct? In the code, it appears much longer than that. Is it possible to share the trained model after higher epoch counts, around 5000?

Also, is it possible to use this pre-trained model for another dataset such as SVHN or do I need to train them from scratch?

Thanks for your response