Deploying your AI models into a mobile device now it is easy thanks to the app HuggingSpot. Download your AI models from HuggingFace repositories and run them seamlessly on your device. Performance is automatically optimised depending on your device, whether it's TPU, GPU, or CPU. Save data, computation, and reduce your carbon footprint. Gain efficiency and privacy all in one place with HuggingSpot.

Key Features:
* Deploy HuggingFace AI models on the Edge.
* Off the grid inferences. Ensure privacy with every use. Once the AI model is in your phone, no share data is sent to the Cloud
* Config AI parameters on the cloud, watch the results on the device.
* Local adjustment of thresholds
* Save data, computation, and reduce your carbon footprint.
* Unleash efficiency as we squeeze the best out of your mobile device. Each device has different hardware specifications, HuggingSpot finds the best configuration for the best performance.


Requirements:
* Models must be converted to TFLite
* Supported official HugginfaceTransformers architectures:
* MobileNetV2ForImageClassification
* MobileViTForSemanticSegmentation
* Supported Non-official architectures:
* SSDForObjectDetection
* Yolov8ForObjectDetection
* Try it out with our models: https://huggingface.co/SpotLab

Coming soon features
* AI model inferences not only over the camera feed, but also over pictures taken or imported.
* AI inferences on images, video, audio, text, conversational LLMs.