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

Attribute Error in vit_jax_augreg #249

Open
TianzhangZhao opened this issue Nov 10, 2022 · 1 comment
Open

Attribute Error in vit_jax_augreg #249

TianzhangZhao opened this issue Nov 10, 2022 · 1 comment

Comments

@TianzhangZhao
Copy link

Hi,

I get an attribute error when I run the vit_jax_augreg.ipynb on colab, it states that
image

copybara-service bot pushed a commit that referenced this issue Nov 11, 2022
1. Flax introduced a new format for GroupNorm weights in checkpoints (google/flax#1721) which is fixed with the added `_fix_groupnorm()`.

2. It was reported in #249 that passing a list as argument does not work anymore. This has been fixed by converting the EagerTensor to a numpy array and adding the batch dimension via fancy indexing.

PiperOrigin-RevId: 487740672
copybara-service bot pushed a commit that referenced this issue Nov 11, 2022
1. Flax introduced a new format for GroupNorm weights in checkpoints (google/flax#1721) which is fixed with the added `_fix_groupnorm()`.

2. It was reported in #249 that passing a list as argument does not work anymore. This has been fixed by converting the EagerTensor to a numpy array and adding the batch dimension via fancy indexing.

PiperOrigin-RevId: 487740672
copybara-service bot pushed a commit that referenced this issue Nov 11, 2022
1. Flax introduced a new format for GroupNorm weights in checkpoints (google/flax#1721) which is fixed with the added `_fix_groupnorm()`.

2. It was reported in #249 that passing a list as argument does not work anymore. This has been fixed by converting the EagerTensor to a numpy array and adding the batch dimension via fancy indexing.

PiperOrigin-RevId: 487740672
@andsteing
Copy link
Collaborator

Hi,

thank you for the error report! I think this is due to the fact that at some point Flax/JAX stopped converting the lists to arrays when calling model.apply().

When running the Colab, I found another error due to some checkpoint inconsistencies introduced in Flax 0.3.6

Both are fixed in #250. I hope that solves the issues for you!

Best, Andreas

copybara-service bot pushed a commit that referenced this issue Nov 11, 2022
1. Flax introduced a new format for GroupNorm weights in checkpoints (google/flax#1721) which is fixed with the added `_fix_groupnorm()`.

2. It was reported in #249 that passing a list as argument does not work anymore. This has been fixed by converting the EagerTensor to a numpy array and adding the batch dimension via fancy indexing.

PiperOrigin-RevId: 487740672
copybara-service bot pushed a commit that referenced this issue Nov 11, 2022
1. Flax introduced a new format for GroupNorm weights in checkpoints (google/flax#1721) which is fixed with the added `_fix_groupnorm()`.

2. It was reported in #249 that passing a list as argument does not work anymore. This has been fixed by converting the EagerTensor to a numpy array and adding the batch dimension via fancy indexing.

PiperOrigin-RevId: 487760232
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