lucidrains vector-quantize-pytorch issues

lucidrains / vector-quantize-pytorch

Vector (and Scalar) Quantization, in Pytorch

MIT License

2.65k stars 215 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Conda Package

#175 iamthebot closed 3 days ago
1
training failed when change the VectorQuantizer to SimVQ

#174 JohnHerry closed 1 week ago
3
Is SimVQ performs even better then VectorQuantizer with ema_update?

#173 JohnHerry closed 1 week ago
1
Simvq

#172 lucidrains closed 1 week ago
0
version 1.19.1 randint() received an invalid combination of arguments - got (int, torch.device, tuple)

#171 yangperasd closed 2 weeks ago
2
A new simple trick to mitigate the codebook collapse

#170 BakerBunker closed 2 weeks ago
69
DDP: RuntimeError: shape mismatch: value tensor of shape [535, 256] cannot be broadcast to indexing result of shape [537, 256]

#169 forwiat opened 2 weeks ago
3
output dimensions for vector quantize

#168 dorsa-zeinali opened 3 weeks ago
2
codebook update

#167 dorsa-zeinali closed 4 weeks ago
5
Title: Issue Regarding the Implementation of **`ResidualFSQ`**

#166 qinxiangyujiayou closed 1 month ago
2
[suggestion] default value of `commitment_loss_weight` in `LFQ`

#165 danelee2601 closed 1 month ago
1
Implement the rotation trick.

#164 cfifty closed 1 month ago
6
Check if we actually training

#163 inspirit closed 1 month ago
1
`shared_codebook=True` not yielding correct results in eval mode

#162 heffernankevin closed 1 month ago
12
Implementation of Regularized Vector Quantization

#161 UkiTenzai closed 1 month ago
4
[BUG] ResidualVQ always initializes MLPs

#160 hummat closed 2 months ago
1
[BUG] LFQ with DDP fails

#159 hummat closed 2 months ago
1
Wrong distributed operation in LFQ

#158 Jason3900 closed 2 months ago
5
What is the difference between Grouped VQ and Multi-codebook VQ when used for Autoregressive LLM training?

#157 JohnHerry opened 2 months ago
0
Remove double l2 normalization

#156 lweitkamp closed 2 months ago
1
There is a future warning

#155 Cyunaam-Wei closed 2 months ago
3
LFQ quantizer quickly converged which leads to codebook lower utilization

#154 Jason3900 closed 2 months ago
3
update autocast

#153 al-jshen closed 2 months ago
1
Difference between FSQ and LFQ

#152 JasonShen-SH opened 3 months ago
1
LatentQuantize exploding loss

#151 jbmaxwell opened 3 months ago
2
Added access to codebook_size and codebook_dim as properties in ResidualVQ

#150 dwromero closed 3 months ago
0
Correct `mask` usage

#149 dwromero closed 3 months ago
10
question about lfq distributed training

#148 syyxsxx closed 4 months ago
1
Seeking clarifications regarding learnable codebook

#147 anupsingh15 opened 4 months ago
0
What's the difference of the distance calculation between the master and fix-lfq-distance branch?

#146 xuzq23 closed 4 months ago
0
LSQ half precision problem #116

#145 denadai2 closed 4 months ago
17
Support half precision for VQ and FSQ

#144 JunityZhan closed 4 months ago
11
[New quantization method] Binary Spherical Quantization

#143 kabachuha closed 4 months ago
12
[BUG] Residual VQ - self.embed.data[ind][mask] = sampled - RuntimeError: shape mismatch:

#142 dwromero closed 4 months ago
17
Multi-GPU Training Hang on Forever

#141 lsyysl9711 closed 5 months ago
1
Fix Issue #139 and refacto of the whole FSQ implementation

#140 MisterBourbaki closed 4 months ago
0
Bug in FSQ, with return_indices to False and image-type feature input

#139 MisterBourbaki opened 5 months ago
0
Small refacto

#138 MisterBourbaki closed 4 months ago
0
FSQ: Added the choice of returning indices

#137 hbenazha closed 5 months ago
1
environment requirement

#136 sivannavis closed 5 months ago
0
a small bug in GroupedResidualVQ

#135 sivannavis closed 5 months ago
3
How to generate image using this repository

#134 tsWen0309 opened 5 months ago
0
Upgrading the examples

#132 MisterBourbaki closed 4 months ago
1
Why do I get almost the same codes after the 1st batch?

#131 tanyz0208 opened 6 months ago
1
Some clarification required

#130 inspirit opened 6 months ago
3
Fix 1e7 in random.randint

#129 leng-yue closed 6 months ago
1
Work on LatentQuantizer

#128 MisterBourbaki closed 6 months ago
0
Will the 16 or 32 codebook dimension hamper the representational capacity of autoencoders?

#127 yilinliu77 opened 6 months ago
2
codebook merge

#126 zhouilu opened 6 months ago
1
Add a simple test suite

#125 MisterBourbaki closed 6 months ago
8