Fuyu

This model was released on 2023-10-17 and added to Hugging Face Transformers on 2023-10-19.

Fuyu

Overview

The Fuyu model was created by ADEPT, and authored by Rohan Bavishi, Erich Elsen, Curtis Hawthorne, Maxwell Nye, Augustus Odena, Arushi Somani, Sağnak Taşırlar.

The authors introduced Fuyu-8B, a decoder-only multimodal model based on the classic transformers architecture, with query and key normalization. A linear encoder is added to create multimodal embeddings from image inputs.

By treating image tokens like text tokens and using a special image-newline character, the model knows when an image line ends. Image positional embeddings are removed. This avoids the need for different training phases for various image resolutions. With 8 billion parameters and licensed under CC-BY-NC, Fuyu-8B is notable for its ability to handle both text and images, its impressive context size of 16K, and its overall performance.

The Fuyu models were trained using bfloat16, but the original inference uses float16 The checkpoints uploaded on the hub use dtype = 'float16' which will be used by the AutoModel API to cast the checkpoints from torch.float32 to torch.float16.

The dtype of the online weights is mostly irrelevant, unless you are using dtype="auto" when initializing a model using model = AutoModelForCausalLM.from_pretrained("path", dtype = "auto"). The reason is that the model will first be downloaded ( using the dtype of the checkpoints online) then it will be cast to the default dtype of torch (becomes torch.float32). Users should specify the dtype they want, and if they don’t it will be torch.float32.

Finetuning the model in float16 is not recommended and known to produce nan, as such the model should be fine-tuned in bfloat16.

Tips:

• To convert the model, you need to clone the original repository using git clone https://github.com/persimmon-ai-labs/adept-inference, then get the checkpoints:

git clone https://github.com/persimmon-ai-labs/adept-inference
wget path/to/fuyu-8b-model-weights.tar
tar -xvf fuyu-8b-model-weights.tar
python src/transformers/models/fuyu/convert_fuyu_weights_to_hf.py  --input_dir /path/to/downloaded/fuyu/weights/ --output_dir /output/path \
    --pt_model_path /path/to/fuyu_8b_release/iter_0001251/mp_rank_00/model_optim_rng.pt
    --ada_lib_path /path/to/adept-inference

For the chat model:

wget https://axtkn4xl5cip.objectstorage.us-phoenix-1.oci.customer-oci.com/n/axtkn4xl5cip/b/adept-public-data/o/8b_chat_model_release.tar
tar -xvf 8b_base_model_release.tar

Then, model can be loaded via:

from transformers import FuyuConfig, FuyuForCausalLM
model_config = FuyuConfig()
model = FuyuForCausalLM(model_config).from_pretrained('/output/path')

Inputs need to be passed through a specific Processor to have the correct formats. A processor requires an image_processor and a tokenizer. Hence, inputs can be loaded via:

from PIL import Image
from transformers import AutoTokenizer
from transformers.models.fuyu.processing_fuyu import FuyuProcessor
from transformers.models.fuyu.image_processing_fuyu_fast import FuyuImageProcessorFast


tokenizer = AutoTokenizer.from_pretrained('adept-hf-collab/fuyu-8b')
image_processor = FuyuImageProcessorFast()


processor = FuyuProcessor(image_processor=image_processor, tokenizer=tokenizer)
text_prompt = "Generate a coco-style caption.\\n"

bus_image_url = "https://huggingface.co/datasets/hf-internal-testing/fixtures-captioning/resolve/main/bus.png"
bus_image_pil = Image.open(io.BytesIO(requests.get(bus_image_url).content))
inputs_to_model = processor(images=bus_image_pil, text=text_prompt)

This model was contributed by Molbap. The original code can be found here.

• Fuyu uses a sentencepiece based tokenizer, with a Unigram model. It supports bytefallback, which is only available in tokenizers==0.14.0 for the fast tokenizer. The LlamaTokenizer is used as it is a standard wrapper around sentencepiece.

• The authors suggest to use the following prompt for image captioning: f"Generate a coco-style caption.\\n"

FuyuConfig

[[autodoc]] FuyuConfig

FuyuModel

[[autodoc]] FuyuModel

FuyuForCausalLM

[[autodoc]] FuyuForCausalLM - forward

FuyuImageProcessor

[[autodoc]] FuyuImageProcessor - call

FuyuImageProcessor

[[autodoc]] FuyuImageProcessorFast - call

FuyuProcessor

[[autodoc]] FuyuProcessor - call