google-research/big_vision
Official codebase used to develop Vision Transformer, SigLIP, MLP-Mixer, LiT and more.
Jupyter NotebookApache-2.0
Issues
- 0
Confusion on FlexiViT
#94 opened by zilunzhang - 0
Errors in notebooks
#93 opened by amrzv - 0
Clarification: SigLIP Image Transform
#91 opened by siddk - 0
Mixup Per Example?
#90 opened by dibyaghosh - 2
Load ViT with CLIPPO Weights
#88 opened by ahmadmustafaanis - 2
Reproduced result for flexivit
#77 opened by liguopeng0923 - 0
- 1
Negative rho values in GSAM training
#80 opened by ankitkv - 0
Text lowering issue
#79 opened by shkarupa-alex - 0
SigLIP and canonicalize
#78 opened by shkarupa-alex - 1
[BUG] in big_vision.models.proj.flexi.vit
#76 opened by KennyFu1997 - 2
Contrastive Input Pipeline
#75 opened by philippe-eecs - 7
Question about SigLIP
#58 opened by maximzubkov - 5
Where do I find the new ViT baseline param?
#1 opened by sayakpaul - 10
Any extra dataset prep needed?
#2 opened by sayakpaul - 2
- 11
- 2
- 6
Question regarding value range (-1,1)
#37 opened by LeoXinhaoLee - 2
- 1
- 2
Is there Pytorch version CLIPPO?
#57 opened by yukistavailable - 1
bfloat16 Training
#41 opened by LeoXinhaoLee - 1
question about FlexiViT
#30 opened by KimWu1994 - 1
questions about t-SNE visualization in FlexiViT
#43 opened by ShipengFu - 2
- 2
Memory Efficient Attention integration
#59 opened by maximzubkov - 1
- 3
- 0
Gradient accumulation
#54 opened by Mahesh-Ram1 - 5
- 2
requirements issue
#49 opened by sneccc - 2
Add other custom optimizers
#32 opened by buttercutter - 2
About RL fine-tuning code release
#40 opened by byminji - 2
AttributeError pp_img in lit notebook
#18 opened by amrzv - 5
Colorization uvim model not working
#19 opened by muhammad-ahmed-ghani - 1
Accuracy of vit-b-16 training
#7 opened by lucasliunju - 2
TPU utilization could be improved further?
#3 opened by sayakpaul