metadata
license: mit
language:
- en
library_name: transformers
tags:
- vision
- VISION-ENCODER-DECODER-MODEL
pipeline_tag: image-to-text
ADD HEAD
Mistral
VISION-ENCODER-DECODER-MODEL
print('Add Vision...')
# ADD HEAD
# Combine pre-trained encoder and pre-trained decoder to form a Seq2Seq model
Vmodel = VisionEncoderDecoderModel.from_encoder_decoder_pretrained(
"google/vit-base-patch16-224-in21k", "LeroyDyer/Mixtral_AI_Tiny"
)
_Encoder_ImageProcessor = Vmodel.encoder
_Decoder_ImageTokenizer = Vmodel.decoder
_VisionEncoderDecoderModel = Vmodel
# Add Pad tokems
LM_MODEL.VisionEncoderDecoder = _VisionEncoderDecoderModel
# Add Sub Components
LM_MODEL.Encoder_ImageProcessor = _Encoder_ImageProcessor
LM_MODEL.Decoder_ImageTokenizer = _Decoder_ImageTokenizer
LM_MODEL