runnable version

GroundingDINO

@@ -1 +0,0 @@
-Subproject commit 9dac4c605b40d87d82bfff5961467556f4d381a9

README.md

@@ -1,29 +1,18 @@
 # Segment Anything for Stable Diffusion WebUI
 
-**This branch is still developing. I discourage you to use it before I remove this warning from README.**
+This **developing** branch attempts to enable [GroundingDINO](https://github.com/IDEA-Research/GroundingDINO), achieving the goal of text prompt->object detection->segment anything->stable diffusion inpainting.
 
-This branch attemts to enable [GroundingDINO](https://github.com/IDEA-Research/GroundingDINO), achieving the goal of text prompt->object detection->segment anything->stable diffusion inpainting.
-
-Warning: You may have a lot trouble using this extension from this branch, due to several incompability of SAM, GroundingDINO, etc. Please follow these steps to bypass potential error.
-
-- If you see `ModuleNotFoundError: No module named 'groundingdino'` on your terminal, restart your webui may help.
-- You may need a long time to get GroundingDINO and its dependency build, because they have C++.
+Warning: You may need a long time to get GroundingDINO and its dependency build, because they have C++.
 
 To use this extension, simply run `git checkout GroundingDINO` on your terminal after `cd` to this extension directory. This branch will be merged to master branch someday, when I think it is stable enough.
 
-## Help Wanted
-
-If anyone know a good solution of installing `GroundingDINO`, please let me know and submit a pull request. My way of installment (in `install.py`) needs a restart, which is honestly not clean. I have not found another way of installing `GroundingDINO`
-
 ## Next Step:
 
-- Find a smooth way to install GroundingDINO without restart
 - point + detection, find a best way to satisfy need:
   - 0 point + N detection->box: BHW, mask: BCHW
   - ALL point + 1 detection->3 images
   - Preview detection + select the best
 
-
 ## How to Use
 
 TODO

install.py

@@ -8,10 +8,7 @@ with open(req_file) as file:
     for lib in file:
         lib = lib.strip()
         if not launch.is_installed(lib):
+            if lib == "groundingdino":
+                lib = "git+https://github.com/IDEA-Research/GroundingDINO"
             launch.run_pip(
                 f"install {lib}", f"sd-webui-segment-anything requirement: {lib}")
-
-# dino_dir = os.path.join(current_dir, "GroundingDINO")
-# launch.git_clone("https://github.com/IDEA-Research/GroundingDINO.git", dino_dir, "GroundingDINO")
-# launch.run(f"{launch.python} -m pip install -e {dino_dir}")
-launch.git_clone()

models/grounding-dino/GroundingDINO_SwinB.cfg.py

@@ -0,0 +1,43 @@
+batch_size = 1
+modelname = "groundingdino"
+backbone = "swin_B_384_22k"
+position_embedding = "sine"
+pe_temperatureH = 20
+pe_temperatureW = 20
+return_interm_indices = [1, 2, 3]
+backbone_freeze_keywords = None
+enc_layers = 6
+dec_layers = 6
+pre_norm = False
+dim_feedforward = 2048
+hidden_dim = 256
+dropout = 0.0
+nheads = 8
+num_queries = 900
+query_dim = 4
+num_patterns = 0
+num_feature_levels = 4
+enc_n_points = 4
+dec_n_points = 4
+two_stage_type = "standard"
+two_stage_bbox_embed_share = False
+two_stage_class_embed_share = False
+transformer_activation = "relu"
+dec_pred_bbox_embed_share = True
+dn_box_noise_scale = 1.0
+dn_label_noise_ratio = 0.5
+dn_label_coef = 1.0
+dn_bbox_coef = 1.0
+embed_init_tgt = True
+dn_labelbook_size = 2000
+max_text_len = 256
+text_encoder_type = "bert-base-uncased"
+use_text_enhancer = True
+use_fusion_layer = True
+use_checkpoint = True
+use_transformer_ckpt = True
+use_text_cross_attention = True
+text_dropout = 0.0
+fusion_dropout = 0.0
+fusion_droppath = 0.1
+sub_sentence_present = True

models/grounding-dino/GroundingDINO_SwinT_OGC.py

@@ -0,0 +1,43 @@
+batch_size = 1
+modelname = "groundingdino"
+backbone = "swin_T_224_1k"
+position_embedding = "sine"
+pe_temperatureH = 20
+pe_temperatureW = 20
+return_interm_indices = [1, 2, 3]
+backbone_freeze_keywords = None
+enc_layers = 6
+dec_layers = 6
+pre_norm = False
+dim_feedforward = 2048
+hidden_dim = 256
+dropout = 0.0
+nheads = 8
+num_queries = 900
+query_dim = 4
+num_patterns = 0
+num_feature_levels = 4
+enc_n_points = 4
+dec_n_points = 4
+two_stage_type = "standard"
+two_stage_bbox_embed_share = False
+two_stage_class_embed_share = False
+transformer_activation = "relu"
+dec_pred_bbox_embed_share = True
+dn_box_noise_scale = 1.0
+dn_label_noise_ratio = 0.5
+dn_label_coef = 1.0
+dn_bbox_coef = 1.0
+embed_init_tgt = True
+dn_labelbook_size = 2000
+max_text_len = 256
+text_encoder_type = "bert-base-uncased"
+use_text_enhancer = True
+use_fusion_layer = True
+use_checkpoint = True
+use_transformer_ckpt = True
+use_text_cross_attention = True
+text_dropout = 0.0
+fusion_dropout = 0.0
+fusion_droppath = 0.1
+sub_sentence_present = True

requirements.txt

@@ -1 +1,2 @@
 segment_anything
+groundingdino

scripts/dino.py

@@ -7,10 +7,10 @@ from modules import scripts, shared
 from modules.devices import device, torch_gc, cpu
 
 # Grounding DINO
-import GroundingDINO.groundingdino.datasets.transforms as T
-from GroundingDINO.groundingdino.models import build_model
-from GroundingDINO.groundingdino.util.slconfig import SLConfig
-from GroundingDINO.groundingdino.util.utils import clean_state_dict
+import groundingdino.datasets.transforms as T
+from groundingdino.models import build_model
+from groundingdino.util.slconfig import SLConfig
+from groundingdino.util.utils import clean_state_dict
 
 
 dino_model_cache = OrderedDict()
@@ -21,12 +21,12 @@ dino_model_info = {
     "repo_id": "ShilongLiu/GroundingDINO",
     "GroundingDINO_SwinT_OGC (694MB)": {
         "checkpoint": "groundingdino_swint_ogc.pth",
-        "config": os.path.join(scripts.basedir(), "GroundingDINO/groundingdino/config/GroundingDINO_SwinT_OGC.py"),
-        "url": "https://huggingface.co/ShilongLiu/GroundingDINO/resolve/main/groundingdino_swint_ogc.pth"
+        "config": os.path.join(dino_model_dir, "GroundingDINO_SwinT_OGC.py"),
+        "url": "https://huggingface.co/ShilongLiu/GroundingDINO/resolve/main/groundingdino_swint_ogc.pth",
     },
     "GroundingDINO_SwinB (938MB)": {
         "checkpoint": "groundingdino_swinb_cogcoor.pth",
-        "config": os.path.join(scripts.basedir(), "GroundingDINO/groundingdino/config/GroundingDINO_SwinB.cfg.py"),
+        "config": os.path.join(dino_model_dir, "GroundingDINO_SwinB.cfg.py"),
         "url": "https://huggingface.co/ShilongLiu/GroundingDINO/resolve/main/groundingdino_swinb_cogcoor.pth"
     },
 }