Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Bug in gather_cc #3

Closed
fuersta opened this issue Aug 2, 2021 · 1 comment
Closed

Bug in gather_cc #3

fuersta opened this issue Aug 2, 2021 · 1 comment

Comments

@fuersta
Copy link

fuersta commented Aug 2, 2021

Hi there,

first of all thanks for the code, i appreciate your effort!

I think there is a bug in gather_cc.py:
In line 86 there is a hardcoded 'val', which should probably be split.

@mitchellnw
Copy link
Contributor

Thank you so much for pointing this out!

Fix is made :) https://github.com/mlfoundations/open_clip/blob/main/src/data/gather_cc.py#L86

arampacha added a commit to arampacha/open_clip that referenced this issue May 22, 2022
* mean pooler
* max pooler
* cls pooler
* default to mean pooler
rom1504 added a commit that referenced this issue Nov 7, 2022
* Generalizable Text Transformer Usage: Adding HuggingFace Model

* sketch for wrapper class

* Building casual attention mask

* better way of getting attn_mask + remove useless params

* adding config arg

* add Mlp from timm

* Poolers with simple tests (#3)

* mean pooler
* max pooler
* cls pooler
* default to mean pooler

* simple pretrained_model test (#4)

* arch_dict: unify naming convention + defaults (#2)

* arch_dict: unify naming convention + defaults

* remove conflict

* migrate arch_dict to separate file

* remove timm Mlp

* get pooler in init from arch_dict

* remove get_pooler

* remove get_pooler typo

* dev integration 2 (#10)

* dev integration 2

* changes to hf_mdoel

* modle.py

* tokenizer

* data.py

* main.py

* params.py

* first debug run

* fix + temporary training script for testing

Co-authored-by: iejmac <iejmac@gpu-st-p4d-24xlarge-4.hpc-1click-sandbox.pcluster>

* add transformers to requirements

* add section in README

* dummy PretrainedConfig

* remove requirements

* try adding to test and train rqs

* strings

* rename config

* update README

* addressed TODO

* remove hf_run script

* add test for roberta-base

* fix synthetic dataset

Co-authored-by: Arto <aruthart@gmail.com>
Co-authored-by: arampacha <69807323+arampacha@users.noreply.github.com>
Co-authored-by: iejmac <iejmac@ip-172-31-34-110.ec2.internal>
Co-authored-by: iejmac <iejmac@gpu-st-p4d-24xlarge-4.hpc-1click-sandbox.pcluster>
Co-authored-by: iejmac <iejmac@gpu-st-p4d-24xlarge-1.hpc-1click-sandbox.pcluster>
Co-authored-by: Romain Beaumont <romain.rom1@gmail.com>
rom1504 pushed a commit that referenced this issue Dec 20, 2022
CoCa PR: fix eval (which uses encode_x instead of forward)
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants