fix: Readme for mmdetection-3 and fix dino with expections (#494)

* fix: Readme for mmdetection-3 and fix dino with expections

* Fix

* Fixes

* Fix

* Fix

* Fix

label_studio_ml/examples/grounding_dino/dino.py

@@ -231,7 +231,8 @@ def multiple_tasks(self, tasks):
                     ls_host=LABEL_STUDIO_HOST,
                     task_id=task.get('id')
                 )
-            except:
+            except Exception as e:
+                logger.error(f"Error getting local path: {e}")
                 img_path = raw_img_path
 
             image_paths.append(img_path)

label_studio_ml/examples/mmdetection-3/README.md

@@ -1,3 +1,12 @@
+# Object detection with bounding boxes using MMDetection
+
+https://mmdetection.readthedocs.io/en/latest/
+
+This example demonstrates how to use the MMDetection model with Label Studio to annotate images with bounding boxes. 
+The model is based on the YOLOv3 architecture with MobileNetV2 backbone and trained on the COCO dataset. 
+
+![screenshot.png](screenshot.png)
+
 # Quick usage
 
 For quick usage run docker-compose in your working directory:
@@ -45,43 +54,91 @@ In this example, you can combine multiple labels into one Label Studio label. Fo
 
 # Run without docker
 
-> These steps provided by @raash1d [in this issue](https://github.com/heartexlabs/label-studio-ml-backend/issues/167#issuecomment-1495061050). Note: the patch from the comment is already applied, except hardcoding of label_config into kwargs.
-
-**It's highly recomended to use docker, it allows to avoid lots of problems!**
+**It's highly recommended to use docker, it allows to avoid lots of problems!**
 
 1. Clone the Label Studio ML Backend repository in your directory of choice
+
 ```
 git clone https://github.com/heartexlabs/label-studio-ml-backend
 cd label-studio-ml-backend/label_studio_ml/examples/mmdetection-3
 ```
 
 2. Create a virtual environment using venv and install all dependencies using pip
+
 ```
 python -m venv ml-backend
 source ml-backend/bin/activate
-pip install -r requirements.txt
 ```
 
-3. Install and Download mmdet related dependencies in the virtual environment
-```
-mim install mmengine
+3. Install all dependencies
+
+```bash
+
+# Set environment variables
+export PYTORCH="1.13.0"
+export CUDA="11.6"
+export CUDNN="8"
+
+# To fix GPG key error when running apt-get update
+apt-key adv --fetch-keys https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/3bf863cc.pub
+apt-key adv --fetch-keys https://developer.download.nvidia.com/compute/machine-learning/repos/ubuntu1804/x86_64/7fa2af80.pub
+
+# Update the base OS
+apt-get update
+apt install -y git
+
+# Install system dependencies for opencv-python
+apt-get install -y libgl1 libglib2.0-0
+
+# Upgrade pip
+pip install -U pip
+
+# Install base requirements
+pip install -r requirements-base.txt
+
+# Install custom requirements
+pip install -r requirements.txt
+
+# Mim install commands
+pip install mim
+mim install mmengine==0.10.3
+mim install mmdet==3.3.0
 mim download mmdet --config yolov3_mobilenetv2_8xb24-320-300e_coco --dest .
-mim install mmcv==2.0.0rc3
 ```
 
-4. Export required variables
+4. Set the environment variables
+
+```bash
+# set the log level for the model server
+export LOG_LEVEL=DEBUG
+# specify the model directory (likely you don't need to change this)
+export MODEL_DIR=
+# mmdetection model parameters
+export CHECKPOINT_FILE=yolov3_mobilenetv2_320_300e_coco_20210719_215349-d18dff72.pth
+export CONFIG_FILE=yolov3_mobilenetv2_8xb24-320-300e_coco.py
+export DEVICE=cpu
+export SCORE_THRESHOLD=0.5
+
+# Specify the Label Studio URL and API key to access
+# uploaded, local storage and cloud storage files.
+# Do not use 'localhost' as it does not work within Docker containers.
+# Use prefix 'http://' or 'https://' for the URL always.
+# Determine the actual IP using 'ifconfig' (Linux/Mac) or 'ipconfig' (Windows).
+export LABEL_STUDIO_URL=<your-ls-hostname>
+export LABEL_STUDIO_API_KEY=<your-ls-api-key>
+```
+
+4. Run the following command to start your ML backend
+
 ```
-export checkpoint_file=yolov3_mobilenetv2_320_300e_coco_20210719_215349-d18dff72.pth
-export config_file=yolov3_mobilenetv2_8xb24-320-300e_coco.py
+label-studio-ml start --root-dir .. mmdetection-3
 ```
 
-5. Run the following command to start your ML backend
+or using gunicorn for a Python web app 
+
 ```
-label-studio-ml start --root-dir .. mmdetection-3 --kwargs hostname=http://<ip.of.your.label.studio.server>:8080 access_token=<access_token> [score_threshold=0.5] [--check]
+gunicorn --preload --bind :9090 --workers 1 --threads 1 --timeout 0 _wsgi:app
 ```
 
 * Use this guide to find out your access token: https://labelstud.io/guide/api.html
-* If you're running label studio and the ML backend on the same machine, then the IP of your label studio server could also be "localhost". Try it, I haven't tested this.
-* `score_threshold` and `--check` are optional.
-* You can use and increased value of `score_threshold` parameter when you see a lot of unwanted detections or lower its value if you don't see any detections.
-* The `--check` parameter allows you to check if all requirements to run the backend have been fulfilled
+* You can use and increased value of `SCORE_THRESHOLD` parameter when you see a lot of unwanted detections or lower its value if you don't see any detections.