From d97b9476783fed56d58c622a85210f023057501a Mon Sep 17 00:00:00 2001 From: Avinash Madasu Date: Wed, 28 Aug 2024 11:39:27 +0530 Subject: [PATCH] Add files via upload --- CLIP-InterpreT/src/topic_seg_labels.py | 128 ++++++++++++------------- 1 file changed, 64 insertions(+), 64 deletions(-) diff --git a/CLIP-InterpreT/src/topic_seg_labels.py b/CLIP-InterpreT/src/topic_seg_labels.py index 0f4fd6f..f00901a 100644 --- a/CLIP-InterpreT/src/topic_seg_labels.py +++ b/CLIP-InterpreT/src/topic_seg_labels.py @@ -259,85 +259,85 @@ def get_ViT_B_32_openai_layer11_head_anno(): ###################################################################################################### def get_ViT_L_14_laion2b_s32b_b82k_layer20_head_anno(): - layer_head_anno = ["L20.H0 (Locations)", - "L20.H1 (Locations)", - "L20.H2 (Locations)", - "L20.H3 (Location)", - "L20.H4 (Numbers)", - "L20.H5 (Locations)", - "L20.H6 (Environments)", - "L20.H7 (Descriptions)", - "L20.H8 (Locations)", - "L20.H9 (Locations)", - "L20.H10 (Objects)", - "L20.H11 (Photography styles)", - "L20.H12 (Visual)", - "L20.H13 (Photography)", - "L20.H14 (Descriptive)", - "L20.H15 (Diverse)", + layer_head_anno = ["L20.H0 (Location)", + "L20.H1 (Location)", + "L20.H2 (Vehicles)", + "L20.H3 (Patterns)", + "L20.H4 (Pair)", + "L20.H5 (Location)", + "L20.H6 (Location)", + "L20.H7 (Age)", + "L20.H8 (Location)", + "L20.H9 (Location)", + "L20.H10 (Object)", + "L20.H11 (Location)", + "L20.H12 (Effect)", + "L20.H13 (Perspective)", + "L20.H14 (Performer)", + "L20.H15 (Location)", "Aggregate", ] return layer_head_anno def get_ViT_L_14_laion2b_s32b_b82k_layer21_head_anno(): - layer_head_anno = ["L21.H0 (Colors)", - "L21.H1 (Locations)", - "L21.H2 (People)", - "L21.H3 (Locations)", - "L21.H4 (Diverse)", - "L21.H5 (Descriptive)", - "L21.H6 (Environments)", - "L21.H7 (Objects)", - "L21.H8 (Settings)", - "L21.H9 (Colors)", - "L21.H10 (Portraits)", - "L21.H11 (Locations)", - "L21.H12 (Themes)", - "L21.H13 (Locations)", - "L21.H14 (Locations)", - "L21.H15 (Scenes)", + layer_head_anno = ["L21.H0 (Monochrome)", + "L21.H1 (Region)", + "L21.H2 (Subject)", + "L21.H3 (Timepiece)", + "L21.H4 (Animals)", + "L21.H5 (Expression)", + "L21.H6 (Location.)", + "L21.H7 (Items)", + "L21.H8 (Description)", + "L21.H9 (Colors)", + "L21.H10 (Facial expressions)", + "L21.H11 (Location)", + "L21.H12 (Effect)", + "L21.H13 (Descriptive)", + "L21.H14 (Location)", + "L21.H15 (Setting)", "Aggregate", ] return layer_head_anno def get_ViT_L_14_laion2b_s32b_b82k_layer22_head_anno(): - layer_head_anno = ["L22.H0 (Patterns)", - "L22.H1 (Geometric shapes)", - "L22.H2 (Locations)", - "L22.H3 (Objects)", - "L22.H4 (Subjects)", - "L22.H5 (Description)", - "L22.H6 (Animals)", - "L22.H7 (Seasons)", - "L22.H8 (Letters)", - "L22.H9 (Activities)", - "L22.H10 (Colors)", - "L22.H11 (Colors)", - "L22.H12 (Locations)", - "L22.H13 (Locations)", - "L22.H14 (Colors)", - "L22.H15 (Emotions)", + layer_head_anno = ["L22.H0 (Technique)", + "L22.H1 (Shape)", + "L22.H2 (Nature)", + "L22.H3 (Bodyparts)", + "L22.H4 (Pets)", + "L22.H5 (Color)", + "L22.H6 (Animals)", + "L22.H7 (Nature)", + "L22.H8 (Letters)", + "L22.H9 (Subject)", + "L22.H10 (Color)", + "L22.H11 (Color)", + "L22.H12 (Location)", + "L22.H13 (Location)", + "L22.H14 (Colors)", + "L22.H15 (Observations)", "Aggregate", ] return layer_head_anno def get_ViT_L_14_laion2b_s32b_b82k_layer23_head_anno(): - layer_head_anno = ["L23.H0 (Materials)", - "L23.H1 (Descriptions)", - "L23.H2 (Animals)", - "L23.H3 (Settings)", - "L23.H4 (People)", - "L23.H5 (Nature)", - "L23.H6 (Locations)", - "L23.H7 (Objects)", - "L23.H8 (Colors)", - "L23.H9 (Varied)", - "L23.H10 (Numbers)", - "L23.H11 (Letters)", - "L23.H12 (Patterns)", - "L23.H13 (Photography)", - "L23.H14 (Modes of transportation)", - "L23.H15 (Emotions)", + layer_head_anno = ["L23.H0 (Patterns)", + "L23.H1 (Setting)", + "L23.H2 (Bird)", + "L23.H3 (Location)", + "L23.H4 (Relationship)", + "L23.H5 (Description)", + "L23.H6 (Location)", + "L23.H7 (Objects)", + "L23.H8 (Colorpalette)", + "L23.H9 (Architecture)", + "L23.H10 (Counting)", + "L23.H11 (Letters)", + "L23.H12 (Patterns)", + "L23.H13 (Location)", + "L23.H14 (Animals)", + "L23.H15 (Music)", "Aggregate", ] return layer_head_anno