|
|
|
|
|
|
|
|
|
|
|
model:
|
|
arch: coco
|
|
load_finetuned: True
|
|
|
|
pretrained: "https://storage.googleapis.com/sfr-vision-language-research/LAVIS/models/BLIP2/blip2_pretrained.pth"
|
|
finetuned: "https://storage.googleapis.com/sfr-vision-language-research/LAVIS/models/BLIP2/blip2_finetune_coco.pth"
|
|
|
|
|
|
image_size: 364
|
|
drop_path_rate: 0
|
|
use_grad_checkpoint: True
|
|
vit_precision: "fp32"
|
|
freeze_vit: False
|
|
|
|
|
|
num_query_token: 32
|
|
|
|
|
|
preprocess:
|
|
vis_processor:
|
|
train:
|
|
name: "blip_image_train"
|
|
image_size: 364
|
|
eval:
|
|
name: "blip_image_eval"
|
|
image_size: 364
|
|
text_processor:
|
|
train:
|
|
name: "blip_caption"
|
|
eval:
|
|
name: "blip_caption"
|
|
|