lucidrains CoCa-pytorch issues

lucidrains / CoCa-pytorch

Implementation of CoCa, Contrastive Captioners are Image-Text Foundation Models, in Pytorch

MIT License

1.03k stars 88 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

contrastive_labels - dist batch size isn't correct

#20 gillotte opened 4 months ago
0
Reproduce Problem

#19 Camellia-tx closed 10 months ago
2
pad_dim_to() is not implemented

#18 aspartate closed 10 months ago
3
register_buffer for masks and position encodings breaks DDP

#17 gshaikov-paige closed 12 months ago
14
Residual connection misssing?

#16 gshaikov-paige closed 12 months ago
0
Contribution to the projects Software Engineering: implementation of CI using Docker

#15 AnHoff opened 1 year ago
4
convolution encoder better result then vit

#14 Alexandr1111111 opened 1 year ago
13
LayerNorm after attentional_pooler

#13 gpucce closed 1 year ago
1
change in runtime

#12 etousi closed 1 year ago
0
Image tokens dimension mismatch with the layernorm bias dimension

#11 szxuhongye closed 1 year ago
12
Contrastive loss should be applied to L2-normed embeddings instead of layer normed?

#10 fedshyvana closed 1 year ago
2
How to train the model using my own dataset?

#9 keepcodeandsmile closed 2 years ago
1
attn_mask

#8 pldlgb opened 2 years ago
3
Do you plan to give the code for fine-tuning？

#7 yang178908 closed 8 months ago
4
Extractor in vit_pytorch will detach the tensor.

#6 techkang closed 2 years ago
2
Reproducing the results in the paper

#5 GKIBMNY closed 10 months ago
2
why train VIT visual encoder first?

#4 Flowerfan closed 2 years ago
1
Generating the caption of a given image

#3 claudiogreco opened 2 years ago
2
Any plan to share the pre-trained models?

#2 sbkim052 opened 2 years ago
24
Maybe don't need this rearrange

#1 CiaoHe closed 2 years ago
2