save tokenizer to each ckpt folder

Add usage example

Save prediction results

Add a script to create dummy pretrained models for testing

Add dataset usage example script

Add dataset script

Rename file

Use new FlaxVisionEncoderDecoderModel class

copy run_summarization_flax.py

remove dataset script

Update dataset script

Fix style

Update test_model.py

Change a parameter in decode(): deterministic -> train

remove unused scripts

final ver

revert

no residual for proj

fix output types

make more similar to pytorch gpt2 code

revert to original flax gpt2

update check against pytorch's version

Add a test against pytorch's GPT2

Change check to crossattention

Change Flax GPT2 with cross-attn layers to be the same as PyTorch's version

Change Flax GPT2 with cross-attn outputs to be the same as PyTorch's version

remove only_self_attn

update test_model.py

Fix project_encoder

Fix project_encoder

Fix project_encoder

Fix project_encoder

Fix project_encoder

Add project_encoder and related layers

Merge branch 'main' of https://huggingface.co/flax-community/vit-gpt2 into main

update test_model.py

Merge branch 'main' of https://huggingface.co/flax-community/vit-gpt2

Make a remark about a serious bug I made previously

update test_model.py

fix pixel_values being casted to int

fix decoder_position_ids in decode()

fix pad token id issue

fix attention_mask missing

fix FlaxViTGPT2LMModule return value

fix query split_heads

fix model args

update model script

update model script

update config script

update config script