Raghav/image embedding #3429

Raghavyadav17 · 2025-06-23T05:20:15Z

Hi @dkalinowski , @dtrawins
Please review this PR

What’s Included

convert_clip.py: Converts HuggingFace CLIP vision model to OpenVINO IR (.xml/.bin)
pre.py: Preprocessing handler for converting images into model-ready tensors
post.py: Postprocessing handler to normalize embeddings
config.json: OVMS configuration for model + mediapipe graph
grph_pipeline.pbtxt: Mediapipe graph defining the inference pipeline

… model is openai/clip 32

…stprocessing, and Mediapipe graph

ovms_env/bin/Activate.ps1

demos/python_demos/clip_image_classification/model_conversion/Image_embeddings/convert_clip.py

demos/python_demos/image_embeddings/servable/graph_pipeline.pbtxt

Dockerfile.ubuntu

demos/python_demos/image_embeddings/servable/config_model.json

dkalinowski · 2025-08-18T09:16:49Z

Dockerfile.ubuntu

@@ -396,8 +396,8 @@ RUN apt-get update ; \
            curl -L -O https://github.com/intel/linux-npu-driver/releases/download/v1.16.0/intel-driver-compiler-npu_1.16.0.20250328-14132024782_ubuntu24.04_amd64.deb ; \
            curl -L -O https://github.com/intel/linux-npu-driver/releases/download/v1.16.0/intel-fw-npu_1.16.0.20250328-14132024782_ubuntu24.04_amd64.deb   ; \
            curl -L -O https://github.com/intel/linux-npu-driver/releases/download/v1.16.0/intel-level-zero-npu_1.16.0.20250328-14132024782_ubuntu24.04_amd64.deb ; \
-            curl -L -O https://github.com/oneapi-src/level-zero/releases/download/v1.20.2/level-zero_1.20.2+u24.04_amd64.deb ; \ 
-        fi ; \      
+            curl -L -O https://github.com/oneapi-src/level-zero/releases/download/v1.20.2/level-zero_1.20.2+u24.04_amd64.deb ; \


please don't edit Dockerfile.ubuntu in this pull request (this makes conflicts and is really not required)

dkalinowski · 2025-08-18T09:17:54Z

demos/python_demos/image_embeddings/grpc_cli.py

+end_time = datetime.datetime.now()
+duration = (end_time - start_time).total_seconds() * 1000
+processing_times.append(int(duration))
+print(f"Detection:\n{results.as_numpy('embedding')}\n")


Suggested change

print(f"Detection:\n{results.as_numpy('embedding')}\n")

print(f"Embeddings:\n{results.as_numpy('embedding')}\n")

dkalinowski · 2025-08-18T09:19:44Z

demos/python_demos/image_embeddings/grpc_cli.py

+
+processing_times = []
+start_time = datetime.datetime.now()
+results = client.infer("python_model", [image_input])


Suggested change

results = client.infer("python_model", [image_input])

results = client.infer(args['model'], [image_input])

please add --model CLI param. with this, user will not need to modify client code, but simply select model with --model dino_graph for example. you can keep some default for easier of use. im sure current python_model will not work, since you serve:

clip_graph

dino_graph

laion_graph

@dkalinowski sure i will do this

dkalinowski · 2025-08-18T09:21:37Z

demos/python_demos/image_embeddings/servable/post.py

+    def initialize(self, kwargs: dict):
+        self.node_name = kwargs.get("node_name", "")
+        if "clip" in self.node_name.lower():
+            self.mode = "clip"


is self.mode really needed?

dkalinowski · 2025-08-18T09:21:40Z

demos/python_demos/image_embeddings/servable/pre.py

+
+        if "clip" in self.node_name.lower():
+            self.processor=CLIPProcessor.from_pretrained("openai/clip-vit-base-patch32")
+            self.mode="clip"


is self.mode really needed?

…ts.txt and updated postprocessing script

…ctor db

Copilot

Pull Request Overview

Adds a complete image embeddings demo system for OpenVINO Model Server (OVMS) with support for multiple AI models (CLIP, DINO, LAION) and vector similarity search capabilities.

Implementation of MediaPipe graph pipelines for CLIP, DINO, and LAION image embedding models
Python preprocessing and postprocessing handlers for image transformation and embedding normalization
Command-line tools for database building and image search functionality
Interactive Streamlit web application for visual similarity search

Reviewed Changes

Copilot reviewed 14 out of 122 changed files in this pull request and generated 12 comments.

Show a summary per file

File	Description
streamlit_app.py	Complete Streamlit web interface for image similarity search with model selection
pre.py	Image preprocessing handler supporting multiple model types via transformers
post.py	Embedding normalization postprocessing handler
graph_*.pbtxt	MediaPipe pipeline configurations for CLIP, DINO, and LAION models
config_model.json	OVMS configuration defining model paths and graph endpoints
search_images.py	Command-line image search utility with performance metrics
requirements.txt	Python dependencies with version constraints
model_conversion/*.py	Scripts for converting HuggingFace models to OpenVINO format
grpc_cli.py	Database building tool with interactive model selection

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

Copilot · 2025-08-29T12:15:15Z

demos/python_demos/image_embeddings/servable/pre.py

@@ -0,0 +1,33 @@
+from pyovms import Tensor
+from transformers import CLIPProcessor,AutoImageProcessor


Missing space after comma in import statement. Should be CLIPProcessor, AutoImageProcessor.

Suggested change

from transformers import CLIPProcessor,AutoImageProcessor

from transformers import CLIPProcessor, AutoImageProcessor

Copilot · 2025-08-29T12:15:15Z

demos/python_demos/image_embeddings/servable/pre.py

+
+class OvmsPythonModel:
+    def initialize(self, kwargs:dict):
+        self.node_name=kwargs.get("node_name","")


Missing spaces around assignment operator and after comma. Should be self.node_name = kwargs.get(\"node_name\", \"\").

Suggested change

self.node_name=kwargs.get("node_name","")

self.node_name = kwargs.get("node_name", "")

Copilot · 2025-08-29T12:15:15Z

demos/python_demos/image_embeddings/servable/pre.py

+        self.node_name=kwargs.get("node_name","")
+
+        if "clip" in self.node_name.lower():
+            self.processor=CLIPProcessor.from_pretrained("openai/clip-vit-base-patch32")
+            self.mode="clip"
+
+        elif "dino" in self.node_name.lower():
+            self.processor=AutoImageProcessor.from_pretrained("facebook/dinov2-base")
+            self.mode="dino"
+
+        elif "laion" in self.node_name.lower():
+            self.processor=CLIPProcessor.from_pretrained("laion/CLIP-ViT-B-32-laion2B-s34B-b79K")
+            self.mode="laion"


Missing spaces around assignment operators throughout the method. Should be self.processor = ... and self.mode = ....

Suggested change

self.node_name=kwargs.get("node_name","")

if "clip" in self.node_name.lower():

self.processor=CLIPProcessor.from_pretrained("openai/clip-vit-base-patch32")

self.mode="clip"

elif "dino" in self.node_name.lower():

self.processor=AutoImageProcessor.from_pretrained("facebook/dinov2-base")

self.mode="dino"

elif "laion" in self.node_name.lower():

self.processor=CLIPProcessor.from_pretrained("laion/CLIP-ViT-B-32-laion2B-s34B-b79K")

self.mode="laion"

self.node_name = kwargs.get("node_name", "")

if "clip" in self.node_name.lower():

self.processor = CLIPProcessor.from_pretrained("openai/clip-vit-base-patch32")

self.mode = "clip"

elif "dino" in self.node_name.lower():

self.processor = AutoImageProcessor.from_pretrained("facebook/dinov2-base")

self.mode = "dino"

elif "laion" in self.node_name.lower():

self.processor = CLIPProcessor.from_pretrained("laion/CLIP-ViT-B-32-laion2B-s34B-b79K")

self.mode = "laion"

Copilot · 2025-08-29T12:15:16Z

demos/python_demos/image_embeddings/servable/pre.py

+        self.node_name=kwargs.get("node_name","")
+
+        if "clip" in self.node_name.lower():
+            self.processor=CLIPProcessor.from_pretrained("openai/clip-vit-base-patch32")
+            self.mode="clip"
+
+        elif "dino" in self.node_name.lower():
+            self.processor=AutoImageProcessor.from_pretrained("facebook/dinov2-base")
+            self.mode="dino"
+
+        elif "laion" in self.node_name.lower():
+            self.processor=CLIPProcessor.from_pretrained("laion/CLIP-ViT-B-32-laion2B-s34B-b79K")
+            self.mode="laion"


Missing spaces around assignment operators throughout the method. Should be self.processor = ... and self.mode = ....

Suggested change

self.node_name=kwargs.get("node_name","")

if "clip" in self.node_name.lower():

self.processor=CLIPProcessor.from_pretrained("openai/clip-vit-base-patch32")

self.mode="clip"

elif "dino" in self.node_name.lower():

self.processor=AutoImageProcessor.from_pretrained("facebook/dinov2-base")

self.mode="dino"

elif "laion" in self.node_name.lower():

self.processor=CLIPProcessor.from_pretrained("laion/CLIP-ViT-B-32-laion2B-s34B-b79K")

self.mode="laion"

self.node_name = kwargs.get("node_name", "")

if "clip" in self.node_name.lower():

self.processor = CLIPProcessor.from_pretrained("openai/clip-vit-base-patch32")

self.mode = "clip"

elif "dino" in self.node_name.lower():

self.processor = AutoImageProcessor.from_pretrained("facebook/dinov2-base")

self.mode = "dino"

elif "laion" in self.node_name.lower():

self.processor = CLIPProcessor.from_pretrained("laion/CLIP-ViT-B-32-laion2B-s34B-b79K")

self.mode = "laion"

Copilot · 2025-08-29T12:15:16Z

demos/python_demos/image_embeddings/servable/pre.py

+        self.node_name=kwargs.get("node_name","")
+
+        if "clip" in self.node_name.lower():
+            self.processor=CLIPProcessor.from_pretrained("openai/clip-vit-base-patch32")
+            self.mode="clip"
+
+        elif "dino" in self.node_name.lower():
+            self.processor=AutoImageProcessor.from_pretrained("facebook/dinov2-base")
+            self.mode="dino"
+
+        elif "laion" in self.node_name.lower():
+            self.processor=CLIPProcessor.from_pretrained("laion/CLIP-ViT-B-32-laion2B-s34B-b79K")
+            self.mode="laion"


Missing spaces around assignment operators throughout the method. Should be self.processor = ... and self.mode = ....

Suggested change

self.node_name=kwargs.get("node_name","")

if "clip" in self.node_name.lower():

self.processor=CLIPProcessor.from_pretrained("openai/clip-vit-base-patch32")

self.mode="clip"

elif "dino" in self.node_name.lower():

self.processor=AutoImageProcessor.from_pretrained("facebook/dinov2-base")

self.mode="dino"

elif "laion" in self.node_name.lower():

self.processor=CLIPProcessor.from_pretrained("laion/CLIP-ViT-B-32-laion2B-s34B-b79K")

self.mode="laion"

self.node_name = kwargs.get("node_name", "")

if "clip" in self.node_name.lower():

self.processor = CLIPProcessor.from_pretrained("openai/clip-vit-base-patch32")

self.mode = "clip"

elif "dino" in self.node_name.lower():

self.processor = AutoImageProcessor.from_pretrained("facebook/dinov2-base")

self.mode = "dino"

elif "laion" in self.node_name.lower():

self.processor = CLIPProcessor.from_pretrained("laion/CLIP-ViT-B-32-laion2B-s34B-b79K")

self.mode = "laion"

Copilot · 2025-08-29T12:15:17Z

demos/python_demos/image_embeddings/model_conversion/dino_conversion.py

+model_id="facebook/dinov2-base"
+print(f"Downloading pretrained model {model_id}...")
+
+model=AutoModel.from_pretrained(model_id)
+processor=AutoImageProcessor.from_pretrained(model_id)
+
+image=Image.new("RGB",(224,224))
+inputs=processor(images=image,return_tensors="pt")["pixel_values"]
+
+print("Converting models...")
+ov_model=ov.convert_model(model,example_input=inputs)
+ov.save_model(ov_model,"dino_image_encoder.xml")
+print("Model saved!")
+
+mod_path="saved_mod/dino/1"
+os.makedirs(mod_path,exist_ok=True)


Missing spaces around assignment operators and after commas throughout the file. Should follow PEP 8 spacing conventions.

Suggested change

model_id="facebook/dinov2-base"

print(f"Downloading pretrained model {model_id}...")

model=AutoModel.from_pretrained(model_id)

processor=AutoImageProcessor.from_pretrained(model_id)

image=Image.new("RGB",(224,224))

inputs=processor(images=image,return_tensors="pt")["pixel_values"]

print("Converting models...")

ov_model=ov.convert_model(model,example_input=inputs)

ov.save_model(ov_model,"dino_image_encoder.xml")

print("Model saved!")

mod_path="saved_mod/dino/1"

os.makedirs(mod_path,exist_ok=True)

model_id = "facebook/dinov2-base"

print(f"Downloading pretrained model {model_id}...")

model = AutoModel.from_pretrained(model_id)

processor = AutoImageProcessor.from_pretrained(model_id)

image = Image.new("RGB", (224, 224))

inputs = processor(images=image, return_tensors="pt")["pixel_values"]

print("Converting models...")

ov_model = ov.convert_model(model, example_input=inputs)

ov.save_model(ov_model, "dino_image_encoder.xml")

print("Model saved!")

mod_path = "saved_mod/dino/1"

os.makedirs(mod_path, exist_ok=True)

Copilot · 2025-08-29T12:15:17Z

demos/python_demos/image_embeddings/model_conversion/dino_conversion.py

+model_id="facebook/dinov2-base"
+print(f"Downloading pretrained model {model_id}...")
+
+model=AutoModel.from_pretrained(model_id)
+processor=AutoImageProcessor.from_pretrained(model_id)
+
+image=Image.new("RGB",(224,224))
+inputs=processor(images=image,return_tensors="pt")["pixel_values"]
+
+print("Converting models...")
+ov_model=ov.convert_model(model,example_input=inputs)
+ov.save_model(ov_model,"dino_image_encoder.xml")
+print("Model saved!")
+
+mod_path="saved_mod/dino/1"
+os.makedirs(mod_path,exist_ok=True)


Missing spaces around assignment operators and after commas throughout the file. Should follow PEP 8 spacing conventions.

Suggested change

model_id="facebook/dinov2-base"

print(f"Downloading pretrained model {model_id}...")

model=AutoModel.from_pretrained(model_id)

processor=AutoImageProcessor.from_pretrained(model_id)

image=Image.new("RGB",(224,224))

inputs=processor(images=image,return_tensors="pt")["pixel_values"]

print("Converting models...")

ov_model=ov.convert_model(model,example_input=inputs)

ov.save_model(ov_model,"dino_image_encoder.xml")

print("Model saved!")

mod_path="saved_mod/dino/1"

os.makedirs(mod_path,exist_ok=True)

model_id = "facebook/dinov2-base"

print(f"Downloading pretrained model {model_id}...")

model = AutoModel.from_pretrained(model_id)

processor = AutoImageProcessor.from_pretrained(model_id)

image = Image.new("RGB", (224, 224))

inputs = processor(images=image, return_tensors="pt")["pixel_values"]

print("Converting models...")

ov_model = ov.convert_model(model, example_input=inputs)

ov.save_model(ov_model, "dino_image_encoder.xml")

print("Model saved!")

mod_path = "saved_mod/dino/1"

os.makedirs(mod_path, exist_ok=True)

Copilot · 2025-08-29T12:15:18Z

demos/python_demos/image_embeddings/model_conversion/dino_conversion.py

+model_id="facebook/dinov2-base"
+print(f"Downloading pretrained model {model_id}...")
+
+model=AutoModel.from_pretrained(model_id)
+processor=AutoImageProcessor.from_pretrained(model_id)
+
+image=Image.new("RGB",(224,224))
+inputs=processor(images=image,return_tensors="pt")["pixel_values"]
+
+print("Converting models...")
+ov_model=ov.convert_model(model,example_input=inputs)
+ov.save_model(ov_model,"dino_image_encoder.xml")
+print("Model saved!")
+
+mod_path="saved_mod/dino/1"
+os.makedirs(mod_path,exist_ok=True)


Missing spaces around assignment operators and after commas throughout the file. Should follow PEP 8 spacing conventions.

Suggested change

model_id="facebook/dinov2-base"

print(f"Downloading pretrained model {model_id}...")

model=AutoModel.from_pretrained(model_id)

processor=AutoImageProcessor.from_pretrained(model_id)

image=Image.new("RGB",(224,224))

inputs=processor(images=image,return_tensors="pt")["pixel_values"]

print("Converting models...")

ov_model=ov.convert_model(model,example_input=inputs)

ov.save_model(ov_model,"dino_image_encoder.xml")

print("Model saved!")

mod_path="saved_mod/dino/1"

os.makedirs(mod_path,exist_ok=True)

model_id = "facebook/dinov2-base"

print(f"Downloading pretrained model {model_id}...")

model = AutoModel.from_pretrained(model_id)

processor = AutoImageProcessor.from_pretrained(model_id)

image = Image.new("RGB", (224, 224))

inputs = processor(images=image, return_tensors="pt")["pixel_values"]

print("Converting models...")

ov_model = ov.convert_model(model, example_input=inputs)

ov.save_model(ov_model, "dino_image_encoder.xml")

print("Model saved!")

mod_path = "saved_mod/dino/1"

os.makedirs(mod_path, exist_ok=True)

Copilot · 2025-08-29T12:15:18Z

demos/python_demos/image_embeddings/model_conversion/dino_conversion.py

+model_id="facebook/dinov2-base"
+print(f"Downloading pretrained model {model_id}...")
+
+model=AutoModel.from_pretrained(model_id)
+processor=AutoImageProcessor.from_pretrained(model_id)
+
+image=Image.new("RGB",(224,224))
+inputs=processor(images=image,return_tensors="pt")["pixel_values"]
+
+print("Converting models...")
+ov_model=ov.convert_model(model,example_input=inputs)
+ov.save_model(ov_model,"dino_image_encoder.xml")
+print("Model saved!")
+
+mod_path="saved_mod/dino/1"
+os.makedirs(mod_path,exist_ok=True)


Missing spaces around assignment operators and after commas throughout the file. Should follow PEP 8 spacing conventions.

Suggested change

model_id="facebook/dinov2-base"

print(f"Downloading pretrained model {model_id}...")

model=AutoModel.from_pretrained(model_id)

processor=AutoImageProcessor.from_pretrained(model_id)

image=Image.new("RGB",(224,224))

inputs=processor(images=image,return_tensors="pt")["pixel_values"]

print("Converting models...")

ov_model=ov.convert_model(model,example_input=inputs)

ov.save_model(ov_model,"dino_image_encoder.xml")

print("Model saved!")

mod_path="saved_mod/dino/1"

os.makedirs(mod_path,exist_ok=True)

model_id = "facebook/dinov2-base"

print(f"Downloading pretrained model {model_id}...")

model = AutoModel.from_pretrained(model_id)

processor = AutoImageProcessor.from_pretrained(model_id)

image = Image.new("RGB", (224, 224))

inputs = processor(images=image, return_tensors="pt")["pixel_values"]

print("Converting models...")

ov_model = ov.convert_model(model, example_input=inputs)

ov.save_model(ov_model, "dino_image_encoder.xml")

print("Model saved!")

mod_path = "saved_mod/dino/1"

os.makedirs(mod_path, exist_ok=True)

Copilot · 2025-08-29T12:15:18Z

demos/python_demos/image_embeddings/model_conversion/dino_conversion.py

+model_id="facebook/dinov2-base"
+print(f"Downloading pretrained model {model_id}...")
+
+model=AutoModel.from_pretrained(model_id)
+processor=AutoImageProcessor.from_pretrained(model_id)
+
+image=Image.new("RGB",(224,224))
+inputs=processor(images=image,return_tensors="pt")["pixel_values"]
+
+print("Converting models...")
+ov_model=ov.convert_model(model,example_input=inputs)
+ov.save_model(ov_model,"dino_image_encoder.xml")
+print("Model saved!")
+
+mod_path="saved_mod/dino/1"
+os.makedirs(mod_path,exist_ok=True)


Missing spaces around assignment operators and after commas throughout the file. Should follow PEP 8 spacing conventions.

Suggested change

model_id="facebook/dinov2-base"

print(f"Downloading pretrained model {model_id}...")

model=AutoModel.from_pretrained(model_id)

processor=AutoImageProcessor.from_pretrained(model_id)

image=Image.new("RGB",(224,224))

inputs=processor(images=image,return_tensors="pt")["pixel_values"]

print("Converting models...")

ov_model=ov.convert_model(model,example_input=inputs)

ov.save_model(ov_model,"dino_image_encoder.xml")

print("Model saved!")

mod_path="saved_mod/dino/1"

os.makedirs(mod_path,exist_ok=True)

model_id = "facebook/dinov2-base"

print(f"Downloading pretrained model {model_id}...")

model = AutoModel.from_pretrained(model_id)

processor = AutoImageProcessor.from_pretrained(model_id)

image = Image.new("RGB", (224, 224))

inputs = processor(images=image, return_tensors="pt")["pixel_values"]

print("Converting models...")

ov_model = ov.convert_model(model, example_input=inputs)

ov.save_model(ov_model, "dino_image_encoder.xml")

print("Model saved!")

mod_path = "saved_mod/dino/1"

os.makedirs(mod_path, exist_ok=True)

… meeting

Raghavyadav17 · 2025-08-30T05:20:45Z

@dkalinowski please review the final pushed code

demos/python_demos/image_embeddings/Readme.md

Co-authored-by: Damian Kalinowski <[email protected]>

Raghavyadav17 · 2025-09-01T13:29:55Z

@dkalinowski both the changes are applied

Raghavyadav17 added 3 commits June 9, 2025 06:15

Created a model download + Openvino IR format conversion script, used…

df317c5

… model is openai/clip 32

Add image embedding pipeline with model conversion, preprocessing, po…

730d0d1

…stprocessing, and Mediapipe graph

Fix: Use numpy<2.0.0 in Dockerfile for compatibility

39e7126

dkalinowski reviewed Jun 23, 2025

View reviewed changes

ovms_env/bin/Activate.ps1 Outdated Show resolved Hide resolved

dkalinowski reviewed Jun 23, 2025

View reviewed changes

demos/python_demos/clip_image_classification/model_conversion/Image_embeddings/convert_clip.py Outdated Show resolved Hide resolved

dkalinowski reviewed Jun 23, 2025

View reviewed changes

demos/python_demos/image_embeddings/servable/graph_pipeline.pbtxt Outdated Show resolved Hide resolved

dkalinowski reviewed Jun 23, 2025

View reviewed changes

demos/python_demos/image_embeddings/servable/graph_pipeline.pbtxt Outdated Show resolved Hide resolved

dkalinowski reviewed Jun 23, 2025

View reviewed changes

demos/python_demos/image_embeddings/servable/graph_pipeline.pbtxt Outdated Show resolved Hide resolved

Raghavyadav17 added 3 commits June 24, 2025 05:26

Cleanup: Removed venvs and config/make.save files per mentor feedback

9c3be96

Removed stray convert_clip.py from old PR (not needed)

ad2151f

Fix: updated MediaPipe graph based on mentor feedback

9c61416

mzegla added the GSoC Contributions that are part of Google Summer of Code projects label Jun 25, 2025

dkalinowski reviewed Jun 26, 2025

View reviewed changes

Dockerfile.ubuntu Outdated Show resolved Hide resolved

dkalinowski reviewed Jun 26, 2025

View reviewed changes

demos/python_demos/image_embeddings/servable/config_model.json Outdated Show resolved Hide resolved

Raghavyadav17 added 7 commits July 2, 2025 07:24

Updated files corrected under Damian's supervision in the meeting

df9413d

Restore Dockerfile.ubuntu after accidental deletion

30bc6a9

fully functional grpc client + requirements.txt

38ad1f3

Conversion scripts for dino and laion models to openvino IR

f92dd9c

Updated pre.py

11c17b5

Updated post.py

d2519fa

updated .json file for all models and their respective mediapipe

012583a

dkalinowski reviewed Aug 18, 2025

View reviewed changes

Raghavyadav17 added 4 commits August 27, 2025 03:44

Final grpc script, updated clip conversion script, updated requiremen…

a2d7460

…ts.txt and updated postprocessing script

Demo images folder which contains all the images required to build ve…

b835e8a

…ctor db

mediapipe for clip model

709e1f7

mediapipe for dino

e119826

Raghavyadav17 added 4 commits August 27, 2025 03:55

mediapipe for laion

a2fd941

demo image

b0368b2

search logic

629fa64

demo frontend

12bb4a6

dkalinowski requested a review from Copilot August 29, 2025 12:12

Copilot AI reviewed Aug 29, 2025

View reviewed changes

Raghavyadav17 added 6 commits August 30, 2025 00:55

config json for the project

57b2bc9

readme for the project

8367268

Final grpc script with cli argument + loading bar as discussed in the…

04f4a53

… meeting

demo img for testing

259e4ac

final dino conversion script based on copilot suggestions

24196cb

final pre.py based on copilt suggestion

cfc063a

Raghavyadav17 added 2 commits August 30, 2025 01:40

fixed dockerfile

d1f24b4

Merge branch 'main' into raghav/image-embedding

1a5129b

dkalinowski reviewed Sep 1, 2025

View reviewed changes

demos/python_demos/image_embeddings/Readme.md Outdated Show resolved Hide resolved

dkalinowski reviewed Sep 1, 2025

View reviewed changes

demos/python_demos/image_embeddings/Readme.md Outdated Show resolved Hide resolved

Raghavyadav17 and others added 3 commits September 1, 2025 09:22

Merge branch 'main' into raghav/image-embedding

00c8fb2

Update demos/python_demos/image_embeddings/Readme.md

6e0f4d0

Co-authored-by: Damian Kalinowski <[email protected]>

Update demos/python_demos/image_embeddings/Readme.md

212485d

Co-authored-by: Damian Kalinowski <[email protected]>

	print(f"Detection:\n{results.as_numpy('embedding')}\n")
	print(f"Embeddings:\n{results.as_numpy('embedding')}\n")

	results = client.infer("python_model", [image_input])
	results = client.infer(args['model'], [image_input])

		@@ -0,0 +1,33 @@
		from pyovms import Tensor
		from transformers import CLIPProcessor,AutoImageProcessor

	from transformers import CLIPProcessor,AutoImageProcessor
	from transformers import CLIPProcessor, AutoImageProcessor

	self.node_name=kwargs.get("node_name","")
	self.node_name = kwargs.get("node_name", "")

Raghav/image embedding #3429

Are you sure you want to change the base?

Raghav/image embedding #3429

Uh oh!

Conversation

Raghavyadav17 commented Jun 23, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

What’s Included

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

dkalinowski Aug 18, 2025

Choose a reason for hiding this comment

Uh oh!

dkalinowski Aug 18, 2025

Choose a reason for hiding this comment

Uh oh!

dkalinowski Aug 18, 2025

Choose a reason for hiding this comment

Uh oh!

Raghavyadav17 Aug 27, 2025

Choose a reason for hiding this comment

Uh oh!

dkalinowski Aug 18, 2025

Choose a reason for hiding this comment

Uh oh!

dkalinowski Aug 18, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull Request Overview

Reviewed Changes

Uh oh!

Copilot AI Aug 29, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Aug 29, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Aug 29, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Aug 29, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Aug 29, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Aug 29, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Aug 29, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Aug 29, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Aug 29, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Aug 29, 2025

Choose a reason for hiding this comment

Uh oh!

Raghavyadav17 commented Aug 30, 2025

Uh oh!

Uh oh!

Uh oh!

Raghavyadav17 commented Sep 1, 2025

Uh oh!

Uh oh!

Raghavyadav17 commented Jun 23, 2025 •

edited

Loading