|
from collections import OrderedDict |
|
|
|
|
|
|
|
MODEL_ID_TO_NAME = OrderedDict( |
|
[ |
|
( |
|
"llava-lvis4v-lrv+lvis4v-lrv-resize-naive-clip-vit-l-14-336px-no-align-2-epochs-llama2pure+13b+stage-finetune+x7", |
|
"Prism-CLIP 13B", |
|
), |
|
( |
|
"llava-lvis4v-lrv+lvis4v-lrv-resize-naive-clip-vit-l-14-336px-no-align-2-epochs-llama2pure+7b+stage-finetune+x7", |
|
"Prism-CLIP 7B", |
|
), |
|
( |
|
"resize-naive-clip-vit-l-14-336px-no-align-llama2pure+13b+stage-finetune+x7", |
|
"Prism-CLIP 13B (Controlled)", |
|
), |
|
( |
|
"resize-naive-clip-vit-l-14-336px-no-align-llama2pure+7b+stage-finetune+x7", |
|
"Prism-CLIP 7B (Controlled)", |
|
), |
|
( |
|
"resize-naive-clip-vit-l-14-336px-no-align+13b+stage-finetune+x7", |
|
"Prism-CLIP 13B (Controlled) - Chat", |
|
), |
|
( |
|
"resize-naive-clip-vit-l-14-336px-no-align+7b+stage-finetune+x7", |
|
"Prism-CLIP 7B (Controlled) - Chat", |
|
), |
|
("llava-v1.5-7b", "LLaVA 1.5: 7B"), |
|
("llava-v1.5-13b", "LLaVA 1.5: 13B"), |
|
] |
|
) |
|
|
|
INTERACTION_MODES_MAP = OrderedDict( |
|
[ |
|
("Chat", "chat"), |
|
("Captioning", "captioning"), |
|
("Bounding Box Prediction", "bbox_pred"), |
|
("Visual Question Answering", "vqa"), |
|
("True/False Visual Question Answering", "true_false"), |
|
] |
|
) |
|
|
