openflamingo

/

OpenFlamingo-4B-vitl-rpj3b-langinstruct

like 2

When I run the given model init code:

from open_flamingo import create_model_and_transforms

model, image_processor, tokenizer = create_model_and_transforms(
    clip_vision_encoder_path="ViT-L-14",
    clip_vision_encoder_pretrained="openai",
    lang_encoder_path="togethercomputer/RedPajama-INCITE-Instruct-3B-v1",
    tokenizer_path="togethercomputer/RedPajama-INCITE-Instruct-3B-v1",
    cross_attn_every_n_layers=2
)

I get the error:
AttributeError: 'GPTNeoXForCausalLM' object has no attribute 'transformer'
Full error paste here: https://gist.github.com/JiahuiKChen/b5d46bb9fe8ac10a238447524b5bbd5f

I cannot load this model or openflamingo/OpenFlamingo-4B-vitl-rpj3b (same error for both)