![Roboflow Inference banner](https://github.com/roboflow/inference/blob/main/banner.png?raw=true) ## 🎬 pip install inference [Roboflow](https://roboflow.com) Inference is the easiest way to use and deploy computer vision models. Inference supports running object detection, classification, instance segmentation, and even foundation models (like CLIP and SAM). You can [train and deploy your own custom model](https://github.com/roboflow/notebooks) or use one of the 50,000+ [fine-tuned models shared by the community](https://universe.roboflow.com). There are three primary `inference` interfaces: * A Python-native package (`pip install inference`) * A self-hosted inference server (`inference server start`) * A [fully-managed, auto-scaling API](https://docs.roboflow.com). ## πŸƒ Getting Started Get up and running with `inference` on your local machine in 3 minutes. ```sh pip install inference # or inference-gpu if you have CUDA ``` Setup [your Roboflow Private API Key](https://app.roboflow.com/settings/api) by exporting a `ROBOFLOW_API_KEY` environment variable or adding it to a `.env` file. ```sh export ROBOFLOW_API_KEY=your_key_here ``` Run [an open-source Rock, Paper, Scissors model](https://universe.roboflow.com/roboflow-58fyf/rock-paper-scissors-sxsw) on your webcam stream: ```python import inference inference.Stream( source="webcam", # or rtsp stream or camera id model="rock-paper-scissors-sxsw/11", # from Universe on_prediction=lambda predictions, image: ( print(predictions) # now hold up your hand: πŸͺ¨ πŸ“„ βœ‚οΈ ) ) ``` > [!NOTE] > Currently, the stream interface only supports object detection Now let's extend the example to use [Supervision](https://roboflow.com/supervision) to visualize the predictions and display them on screen with OpenCV: ```python import cv2 import inference import supervision as sv annotator = sv.BoxAnnotator() inference.Stream( source="webcam", # or rtsp stream or camera id model="rock-paper-scissors-sxsw/11", # from Universe output_channel_order="BGR", use_main_thread=True, # for opencv display on_prediction=lambda predictions, image: ( print(predictions), # now hold up your hand: πŸͺ¨ πŸ“„ βœ‚οΈ cv2.imshow( "Prediction", annotator.annotate( scene=image, detections=sv.Detections.from_roboflow(predictions) ) ), cv2.waitKey(1) ) ) ``` ## πŸ‘©β€πŸ« More Examples The [`/examples`](https://github.com/roboflow/inference/tree/main/examples/) directory contains code samples for working with and extending `inference` including using foundation models like CLIP, HTTP and UDP clients, and an insights dashboard, along with community examples (PRs welcome)! ## πŸŽ₯ Inference in action Check out Inference running on a video of a football game: https://github.com/roboflow/inference/assets/37276661/121ab5f4-5970-4e78-8052-4b40f2eec173 ## πŸ’» Why Inference? Inference provides a scalable method through which you can manage inferences for your vision projects. Inference is composed of: - Thousands of [pre-trained community models](https://universe.roboflow.com) that you can use as a starting point. - Foundation models like CLIP, SAM, and OCR. - A tight integration with [Supervision](https://roboflow.com/supervision). - An HTTP server, so you don’t have to reimplement things like image processing and prediction visualization on every project and you can scale your GPU infrastructure independently of your application code, and access your model from whatever language your app is written in. - Standardized APIs for computer vision tasks, so switching out the model weights and architecture can be done independently of your application code. - A model registry, so your code can be independent from your model weights & you don't have to re-build and re-deploy every time you want to iterate on your model weights. - Active Learning integrations, so you can collect more images of edge cases to improve your dataset & model the more it sees in the wild. - Seamless interoperability with [Roboflow](https://roboflow.com) for creating datasets, training & deploying custom models. And more! ### πŸ“Œ Use the Inference Server You can learn more about Roboflow Inference Docker Image build, pull and run in our [documentation](https://inference.roboflow.com/quickstart/docker/). - Run on x86 CPU: ```bash docker run --net=host roboflow/roboflow-inference-server-cpu:latest ``` - Run on NVIDIA GPU: ```bash docker run --network=host --gpus=all roboflow/roboflow-inference-server-gpu:latest ```
πŸ‘‰ more docker run options - Run on arm64 CPU: ```bash docker run -p 9001:9001 roboflow/roboflow-inference-server-arm-cpu:latest ``` - Run on NVIDIA GPU with TensorRT Runtime: ```bash docker run --network=host --gpus=all roboflow/roboflow-inference-server-trt:latest ``` - Run on NVIDIA Jetson with JetPack `4.x`: ```bash docker run --privileged --net=host --runtime=nvidia roboflow/roboflow-inference-server-jetson:latest ``` - Run on NVIDIA Jetson with JetPack `5.x`: ```bash docker run --privileged --net=host --runtime=nvidia roboflow/roboflow-inference-server-jetson-5.1.1:latest ```
### Extras: Some functionality requires extra dependencies. These can be installed by specifying the desired extras during installation of Roboflow Inference. | extra | description | |:-------|:-------------------------------------------------| | `clip` | Ability to use the core `CLIP` model (by OpenAI) | | `gaze` | Ability to use the core `Gaze` model | | `http` | Ability to run the http interface | | `sam` | Ability to run the core `Segment Anything` model (by Meta AI) | **_Note:_** Both CLIP and Segment Anything require pytorch to run. These are included in their respective dependencies however pytorch installs can be highly environment dependent. See the [official pytorch install page](https://pytorch.org/get-started/locally/) for instructions specific to your enviornment. Example install with CLIP dependencies: ```bash pip install "inference[clip]" ``` ## Inference Client To consume predictions from inference server in Python you can use the `inference-sdk` package. ```bash pip install inference-sdk ``` ```python from inference_sdk import InferenceHTTPClient image_url = "https://media.roboflow.com/inference/soccer.jpg" # Replace ROBOFLOW_API_KEY with your Roboflow API Key client = InferenceHTTPClient( api_url="http://localhost:9001", # or https://detect.roboflow.com for Hosted API api_key="ROBOFLOW_API_KEY" ) with client.use_model("soccer-players-5fuqs/1"): predictions = client.infer(image_url) print(predictions) ``` Visit our [documentation](https://inference.roboflow.com/) to discover capabilities of `inference-clients` library. ## Single Image Inference After installing `inference` via pip, you can run a simple inference on a single image (vs the video stream example above) by instantiating a `model` and using the `infer` method (don't forget to setup your `ROBOFLOW_API_KEY` environment variable or `.env` file): ```python from inference.models.utils import get_roboflow_model model = get_roboflow_model( model_id="soccer-players-5fuqs/1" ) # you can also infer on local images by passing a file path, # a PIL image, or a numpy array results = model.infer( image="https://media.roboflow.com/inference/soccer.jpg", confidence=0.5, iou_threshold=0.5 ) print(results) ``` ## Getting CLIP Embeddings You can run inference with [OpenAI's CLIP model](https://blog.roboflow.com/openai-clip) using: ```python from inference.models import Clip image_url = "https://media.roboflow.com/inference/soccer.jpg" model = Clip() embeddings = model.embed_image(image_url) print(embeddings) ``` ## Using SAM You can run inference with [Meta's Segment Anything model](https://blog.roboflow.com/segment-anything-breakdown/) using: ```python from inference.models import SegmentAnything image_url = "https://media.roboflow.com/inference/soccer.jpg" model = SegmentAnything() embeddings = model.embed_image(image_url) print(embeddings) ``` ## πŸ—οΈ inference Process To standardize the inference process throughout all our models, Roboflow Inference has a structure for processing inference requests. The specifics can be found on each model's respective page, but overall it works like this for most models: inference structure ## βœ… Supported Models ### Load from Roboflow You can use models hosted on Roboflow with the following architectures through Inference: - YOLOv5 Object Detection - YOLOv5 Instance Segmentation - YOLOv8 Object Detection - YOLOv8 Classification - YOLOv8 Segmentation - YOLACT Segmentation - ViT Classification ### Core Models Core Models are foundation models and models that have not been fine-tuned on a specific dataset. The following core models are supported: 1. CLIP 2. L2CS (Gaze Detection) 3. Segment Anything (SAM) ## πŸ“ License The Roboflow Inference code is distributed under an [Apache 2.0 license](https://github.com/roboflow/inference/blob/master/LICENSE.md). The models supported by Roboflow Inference have their own licenses. View the licenses for supported models below. | model | license | | :------------------------ | :-----------------------------------------------------------------------------------------------------------------------------------: | | `inference/models/clip` | [MIT](https://github.com/openai/CLIP/blob/main/LICENSE) | | `inference/models/gaze` | [MIT](https://github.com/Ahmednull/L2CS-Net/blob/main/LICENSE), [Apache 2.0](https://github.com/google/mediapipe/blob/master/LICENSE) | | `inference/models/sam` | [Apache 2.0](https://github.com/facebookresearch/segment-anything/blob/main/LICENSE) | | `inference/models/vit` | [Apache 2.0](https://github.com/roboflow/inference/main/inference/models/vit/LICENSE) | | `inference/models/yolact` | [MIT](https://github.com/dbolya/yolact/blob/master/README.md) | | `inference/models/yolov5` | [AGPL-3.0](https://github.com/ultralytics/yolov5/blob/master/LICENSE) | | `inference/models/yolov7` | [GPL-3.0](https://github.com/WongKinYiu/yolov7/blob/main/README.md) | | `inference/models/yolov8` | [AGPL-3.0](https://github.com/ultralytics/ultralytics/blob/master/LICENSE) | ## Inference CLI We've created a CLI tool with useful commands to make the `inference` usage easier. Check out [docs](./inference_cli/README.md). ## πŸš€ Enterprise With a Roboflow Inference Enterprise License, you can access additional Inference features, including: - Server cluster deployment - Device management - Active learning - YOLOv5 and YOLOv8 commercial license To learn more, [contact the Roboflow team](https://roboflow.com/sales). ## πŸ“š documentation Visit our [documentation](https://inference.roboflow.com) for usage examples and reference for Roboflow Inference. ## πŸ† contribution We would love your input to improve Roboflow Inference! Please see our [contributing guide](https://github.com/roboflow/inference/blob/master/CONTRIBUTING.md) to get started. Thank you to all of our contributors! πŸ™ ## πŸ’» explore more Roboflow open source projects | Project | Description | | :---------------------------------------------------------------- | :----------------------------------------------------------------------------------------------------------------------------------------------------- | | [supervision](https://roboflow.com/supervision) | General-purpose utilities for use in computer vision projects, from predictions filtering and display to object tracking to model evaluation. | | [Autodistill](https://github.com/autodistill/autodistill) | Automatically label images for use in training computer vision models. | | [Inference](https://github.com/roboflow/inference) (this project) | An easy-to-use, production-ready inference server for computer vision supporting deployment of many popular model architectures and fine-tuned models. | | [Notebooks](https://roboflow.com/notebooks) | Tutorials for computer vision tasks, from training state-of-the-art models to tracking objects to counting objects in a zone. | | [Collect](https://github.com/roboflow/roboflow-collect) | Automated, intelligent data collection powered by CLIP. |