rosinality/vq-vae-2-pytorch issues and pull requests

#79 - Decoder Top vs Upsamping layer

Issue - State: open - Opened by space1panda 9 months ago

#78 - 上层特征量化后如何与下层融合问题

Issue - State: closed - Opened by Ontheroad123 10 months ago

#77 - How to get latent space vector

Issue - State: open - Opened by 27yw 11 months ago

#76 - Image blurriness problem with VQ-VAE

Issue - State: open - Opened by easonoob over 1 year ago

#75 - How to resume training from a half-trained checkpoint

Issue - State: closed - Opened by easonoob over 1 year ago - 1 comment

#74 - Dataset format problems

Issue - State: closed - Opened by easonoob over 1 year ago - 1 comment

#73 - How to condition the sampler

Issue - State: open - Opened by parthnatekar over 1 year ago

#72 - Update vqvae.py

Pull Request - State: open - Opened by provos over 1 year ago

#71 - The result is a noisy image

Issue - State: open - Opened by MapleSalmon7 about 2 years ago

#70 - Stuck at epoch 1 iter 1 when train vqvae with multi-gpu

Issue - State: open - Opened by JamesZhutheThird about 2 years ago - 11 comments

#69 - question about latent_loss_weight

Issue - State: closed - Opened by Sharpiless over 2 years ago

#68 - How to to train pixelsnail?

Issue - State: closed - Opened by preethiseshadri518 almost 3 years ago - 3 comments

#67 - how to distributed train?

Issue - State: open - Opened by Dududu233 almost 3 years ago - 5 comments

#66 - PixelSNAIL overfitting issue

Issue - State: open - Opened by vipul109 over 3 years ago - 8 comments

#65 - Support for torch.cuda.amp in VQ-VAE training

Issue - State: open - Opened by vvvm23 over 3 years ago - 6 comments

#64 - Runtime Error While Training PixelSNAIL

Issue - State: open - Opened by barisbatuhan over 3 years ago - 1 comment

#63 - VQ-VAE-2 with three levels

Issue - State: closed - Opened by vvvm23 over 3 years ago - 5 comments

#62 - Unable to load the model when training VQVAE with multiple GPUs

Issue - State: closed - Opened by luyvlei over 3 years ago - 3 comments

#61 - Changing the latent map sizes

Issue - State: open - Opened by SURABHI-GUPTA over 3 years ago - 3 comments

#60 - ResBlock with inplace relu?

Issue - State: open - Opened by rfeinman over 3 years ago - 5 comments

#59 - Video Prediction Results

Issue - State: open - Opened by pravn over 3 years ago - 1 comment

#58 - Unbelievable results of the PixelSnail network

Issue - State: open - Opened by ZhanYangen over 3 years ago - 8 comments

#57 - About the loss of VQVAE2 network

Issue - State: open - Opened by ZhanYangen over 3 years ago - 4 comments

#56 - Image File Extension

Issue - State: open - Opened by SURABHI-GUPTA almost 4 years ago - 29 comments

#55 - Checkpoint for PixelSNAIL

Issue - State: open - Opened by betterze almost 4 years ago

#54 - Top/Bottom unexpected information in Stage1

Issue - State: closed - Opened by SURABHI-GUPTA almost 4 years ago - 4 comments

#53 - Any plan to open-source pre-trained top/bottom models of pixel-snail

Issue - State: open - Opened by shawnthu almost 4 years ago

#52 - Images with different size (e.g. 128*128)

Issue - State: closed - Opened by JCL823 almost 4 years ago - 2 comments

#51 - Codebook for different levels

Issue - State: open - Opened by SURABHI-GUPTA almost 4 years ago - 1 comment

#50 - AttributeError: 'CodeRow' object has no attribute 'file'

Issue - State: closed - Opened by JCL823 almost 4 years ago

#49 - Training Hypterparameters of PixelSnail for VQ-VAE experiments

Issue - State: open - Opened by fostiropoulos almost 4 years ago - 3 comments

#48 - Reconstructions on LFW

Issue - State: closed - Opened by SURABHI-GUPTA about 4 years ago - 14 comments

#47 - [train_vqvae] Can I know rationale why training sample shows as duplicate?

Issue - State: open - Opened by kimdn about 4 years ago - 5 comments

#46 - [extract_code] error in lmdb

Issue - State: open - Opened by kimdn about 4 years ago - 1 comment

#45 - [train_vqvae] multiple gpu seems not work as expected

Issue - State: open - Opened by kimdn about 4 years ago - 1 comment

#44 - [train_vqvae] generated images look blank

Issue - State: open - Opened by kimdn about 4 years ago - 2 comments

#43 - A question about image crispness with VQ-VAE2

Issue - State: open - Opened by francoisruty about 4 years ago - 11 comments

#42 - [Question] Which K (or number of embeddings) to choose for VQ?

Issue - State: open - Opened by jjmartinezro about 4 years ago - 6 comments

#41 - Error when running Pixel Snail on a custom dataset

Issue - State: closed - Opened by xandernewton about 4 years ago - 2 comments

#40 - about one_hot

Issue - State: open - Opened by Sky-zhan about 4 years ago - 2 comments

#39 - Output of autoregressive prior

Issue - State: closed - Opened by harrygcoppock about 4 years ago - 6 comments

#38 - train hierarchical vq-vae

Issue - State: open - Opened by mahmud1006 about 4 years ago - 2 comments

#37 - How to get the dataset?

Issue - State: open - Opened by RyannDaGreat about 4 years ago - 10 comments

#36 - using vq-vae-2 for scoring probability of an image

Issue - State: open - Opened by zoharbarzelay about 4 years ago - 5 comments

#35 - weight_v meaning

Issue - State: open - Opened by Dashi-1997 about 4 years ago - 1 comment

#34 - vqvae reconstructed images are too blurry

Issue - State: open - Opened by lelechen63 over 4 years ago - 3 comments

#33 - Unable to get good results

Issue - State: open - Opened by yellowwoods12 over 4 years ago - 3 comments

#32 - Fidelity to the VQ-VAE-2 paper

Issue - State: closed - Opened by natoucs-datagen over 4 years ago - 1 comment

#31 - Error in code for PixelSnail (Proper masking)

Issue - State: open - Opened by rakhimovv over 4 years ago - 2 comments

#30 - vqvae can't reconstruction

Issue - State: closed - Opened by LonglongaaaGo over 4 years ago - 3 comments

#29 - How To Generate Images

Issue - State: closed - Opened by impulsecorp over 4 years ago - 2 comments

#28 - Quantization process is wrong

Issue - State: closed - Opened by danieltudosiu over 4 years ago - 4 comments

#27 - Struggling to figure out why model learns for 3 images but not 4

Issue - State: open - Opened by steve-1820 over 4 years ago - 1 comment

#26 - Has anyone reproduced the same result as the author of the stage-1?

Issue - State: open - Opened by ZexinYan over 4 years ago

#25 - Update pixelsnail_mnist.py

Pull Request - State: closed - Opened by david-macleod over 4 years ago

#24 - train_vqvae with 1 channel - error

Issue - State: closed - Opened by pythonmobile over 4 years ago - 1 comment

#23 - Ask for pre-trained models

Issue - State: open - Opened by sangwoomo over 4 years ago

#22 - Results of Pixelsnail are not good

Issue - State: open - Opened by hasamkhalid almost 5 years ago - 3 comments

#21 - Cannot reconstruct when use mel spectrum as data

Issue - State: open - Opened by LiNaihan almost 5 years ago - 1 comment

#20 - why we are including z_e(x) also for decoder input

Issue - State: closed - Opened by shiva1393 almost 5 years ago - 1 comment

#19 - Training procedure?

Issue - State: closed - Opened by gslaller almost 5 years ago - 2 comments

#18 - Targeted Image Generation

Issue - State: open - Opened by hasamkhalid almost 5 years ago

#17 - Is q_latent_loss missing?

Issue - State: closed - Opened by nimaous almost 5 years ago - 1 comment

#16 - The hardware required.

Issue - State: closed - Opened by nobodypengium about 5 years ago - 2 comments

#15 - Unexpected information in top/bottom levels

Issue - State: closed - Opened by pabloppp about 5 years ago - 13 comments

#14 - Latent maps sizes

Issue - State: closed - Opened by takada-lmd about 5 years ago - 3 comments

#13 - how to do with video?

Issue - State: open - Opened by 1306107097 about 5 years ago - 1 comment

#12 - can't pickle Environment objects

Issue - State: open - Opened by MichalZajac about 5 years ago - 5 comments

#11 - class-conditional samples

Issue - State: open - Opened by edgarschnfld about 5 years ago - 1 comment

#10 - Is is reasonable when the loss increases on validation set?

Issue - State: closed - Opened by BenjaminChoou about 5 years ago - 1 comment

#9 - Conditioned Sample

Issue - State: open - Opened by Arstanley about 5 years ago - 11 comments

#8 - Support multi GPU training on stage 1

Issue - State: closed - Opened by Corea about 5 years ago - 1 comment

#7 - Error when the number of data is not a multiple of batch size

Issue - State: closed - Opened by Corea about 5 years ago - 1 comment

#6 - Error on train_pixelsnail.py after last update

Issue - State: closed - Opened by Corea about 5 years ago - 1 comment

#5 - Any tips for stage 2 training

Issue - State: closed - Opened by 1Konny about 5 years ago - 2 comments

#4 - [Question] What is PixelSnail? How to Train it?

Issue - State: open - Opened by EibrielInv about 5 years ago - 35 comments

#3 - VQ-VAE Loss Function

Issue - State: closed - Opened by Arstanley over 5 years ago - 2 comments

#2 - AttributeError: module 'torch.nn.functional' has no attribute 'one_hot'

Issue - State: closed - Opened by cdq14 over 5 years ago - 2 comments

#1 - State of Current Implementation

Issue - State: closed - Opened by pclucas14 over 5 years ago - 2 comments

GitHub / rosinality/vq-vae-2-pytorch issues and pull requests