ipadapter masks api support

modules/api/control.py

@@ -95,7 +95,7 @@ class APIControl():
 
     def prepare_ip_adapter(self, request):
         if hasattr(request, "ip_adapter") and request.ip_adapter:
-            args = { 'ip_adapter_names': [], 'ip_adapter_scales': [], 'ip_adapter_starts': [], 'ip_adapter_ends': [], 'ip_adapter_images': [] }
+            args = { 'ip_adapter_names': [], 'ip_adapter_scales': [], 'ip_adapter_starts': [], 'ip_adapter_ends': [], 'ip_adapter_images': [], 'ip_adapter_masks': [] }
             for ipadapter in request.ip_adapter:
                 if not ipadapter.images or len(ipadapter.images) == 0:
                     continue
@@ -104,6 +104,9 @@ class APIControl():
                 args['ip_adapter_starts'].append(ipadapter.start)
                 args['ip_adapter_ends'].append(ipadapter.end)
                 args['ip_adapter_images'].append([helpers.decode_base64_to_image(x) for x in ipadapter.images])
+                if ipadapter.masks:
+                    args['ip_adapter_masks'].append([helpers.decode_base64_to_image(x) for x in ipadapter.masks])
+
             del request.ip_adapter
             return args
         else:

modules/api/generate.py

@@ -67,7 +67,6 @@ class APIGenerate():
             p.ip_adapter_starts = []
             p.ip_adapter_ends = []
             p.ip_adapter_images = []
-            p.ip_adapter_masks = []
             for ipadapter in request.ip_adapter:
                 if not ipadapter.images or len(ipadapter.images) == 0:
                     continue
@@ -76,7 +75,9 @@ class APIGenerate():
                 p.ip_adapter_starts.append(ipadapter.start)
                 p.ip_adapter_ends.append(ipadapter.end)
                 p.ip_adapter_images.append([helpers.decode_base64_to_image(x) for x in ipadapter.images])
-                p.ip_adapter_masks.append([helpers.decode_base64_to_image(x) for x in ipadapter.masks])
+                p.ip_adapter_masks = []
+                if ipadapter.masks:
+                    p.ip_adapter_masks.append([helpers.decode_base64_to_image(x) for x in ipadapter.masks])
             del request.ip_adapter
 
     def post_text2img(self, txt2imgreq: models.ReqTxt2Img):

modules/ipadapter.py

@@ -45,7 +45,9 @@ def get_images(input_images):
     if not isinstance(input_images, list):
         input_images = [input_images]
     for image in input_images:
-        if isinstance(image, list):
+        if image is None:
+            continue
+        elif isinstance(image, list):
             output_images.append(get_images(image)) # recursive
         elif isinstance(image, Image.Image):
             output_images.append(image)
@@ -108,9 +110,11 @@ def apply(pipe, p: processing.StableDiffusionProcessing, adapter_names=[], adapt
     if hasattr(p, 'ip_adapter_images'):
         adapter_images = p.ip_adapter_images
     adapter_images = get_images(adapter_images)
-    if hasattr(p, 'ip_adapter_masks'):
+    if hasattr(p, 'ip_adapter_masks') and len(p.ip_adapter_masks) > 0:
         adapter_masks = p.ip_adapter_masks
-    adapter_masks = get_images(adapter_masks)
+        adapter_masks = get_images(adapter_masks)
+    else:
+        adapter_masks = []
     if len(adapter_masks) > 0:
         from diffusers.image_processor import IPAdapterMaskProcessor
         mask_processor = IPAdapterMaskProcessor()

requirements.txt

@@ -22,9 +22,7 @@ psutil
 pyyaml
 resize-right
 rich
-safetensors
 scipy
-tensordict==0.1.2
 toml
 torchdiffeq
 voluptuous
@@ -33,8 +31,10 @@ scikit-image
 fasteners
 dctorch
 pymatting
-peft
 orjson
+safetensors==0.4.3
+tensordict==0.1.2
+peft==0.10.0
 httpx==0.24.1
 compel==2.0.2
 torchsde==0.2.6
@@ -42,7 +42,7 @@ clip-interrogator==0.6.0
 antlr4-python3-runtime==4.9.3
 requests==2.31.0
 tqdm==4.66.2
-accelerate==0.29.1
+accelerate==0.29.3
 opencv-contrib-python-headless==4.9.0.80
 diffusers==0.27.2
 einops==0.4.1
@@ -54,11 +54,11 @@ numba==0.59.1
 pandas
 protobuf==3.20.3
 pytorch_lightning==1.9.4
-tokenizers==0.15.2
-transformers==4.39.3
+tokenizers==0.19.1
+transformers==4.40.0
 tomesd==0.1.3
 urllib3==1.26.18
 Pillow==10.3.0
 timm==0.9.16
-pydantic==1.10.13
+pydantic==1.10.15
 typing-extensions==4.11.0

scripts/ipadapter.py

@@ -93,4 +93,5 @@ class Script(scripts.Script):
             p.ip_adapter_ends = args[MAX_ADAPTERS*4:MAX_ADAPTERS*5][:units]
         if getattr(p, 'ip_adapter_masks', []) == []:
             p.ip_adapter_masks = args[MAX_ADAPTERS*5:MAX_ADAPTERS*6][:units]
+            p.ip_adapter_masks = [x for x in p.ip_adapter_masks if x]
         # ipadapter.apply(shared.sd_model, p, p.ip_adapter_names, p.ip_adapter_scales, p.ip_adapter_starts, p.ip_adapter_ends, p.ip_adapter_images) # called directly from processing.process_images_inner