# Mobile Optimization

A compact language model based on the Llama architecture, supporting English and Portuguese, with 96 million parameters and a context length of 4096 tokens.

Multiple variants based on DeepSeek-R1-Distill-Qwen-1.5B, adapted for LiteRT framework and MediaPipe LLM inference API, deployable on Android platforms.

Llama 3.2 is a 1B and 3B parameter-scale multilingual pre-trained and instruction-tuned generative model from Meta, optimized for multilingual dialogue use cases and supporting 8 official languages.

SAM 2.1 Tiny is a lightweight image and video universal segmentation model introduced by Facebook AI Research (FAIR), supporting controllable visual segmentation based on prompts.

Llama 3.2 is a collection of multilingual large language models released by Meta, including pre-trained and instruction-tuned versions with 1B and 3B parameter scales, optimized for multilingual conversational use cases and supporting 8 official languages.

MobileCLIP-S1 is an efficient image-text model that achieves fast zero-shot image classification through multimodal reinforcement training.

Imp-v1.5-4B-Phi3 is a high-performance lightweight multimodal large model with only 4 billion parameters, built on the Phi-3 framework and SigLIP visual encoder.

A semantic segmentation model based on MobileViTv2 architecture with DeepLabV3 head, pretrained on PASCAL VOC dataset at 512x512 resolution

A plant disease identification model fine-tuned based on the MobileNetV2 architecture, achieving 95.41% accuracy on the PlantVillage dataset

EfficientNet is an efficient convolutional neural network that achieves high-performance image classification by uniformly scaling depth, width, and resolution

EfficientNet is a mobile-friendly pure convolutional model that uniformly scales depth/width/resolution dimensions through compound coefficients, trained on the ImageNet-1k dataset.

EfficientNet is a mobile-friendly pure convolutional model that uniformly scales depth, width, and resolution dimensions, trained on the ImageNet-1k dataset.

EfficientNet is a mobile-friendly pure convolutional neural network that achieves efficient scaling by uniformly adjusting depth/width/resolution dimensions through compound coefficients

EfficientNet is a mobile-friendly pure convolutional model that achieves excellent performance in image classification tasks by uniformly scaling depth/width/resolution dimensions with compound coefficients.

EfficientNet is a mobile-friendly pure convolutional neural network that achieves efficient scaling by uniformly adjusting depth/width/resolution dimensions through compound coefficients.

EfficientFormer-L3 is a lightweight vision Transformer model developed by Snap Research, optimized for mobile devices to achieve low latency while maintaining high performance.

MobileNet V1 is a lightweight convolutional neural network designed for mobile devices, balancing latency, model size, and accuracy in image classification tasks.

MobileNet V1 is a lightweight convolutional neural network designed for mobile and embedded vision applications, pre-trained on the ImageNet-1k dataset.

A lightweight semantic segmentation model based on MobileNetV2 architecture combined with DeepLabV3+ segmentation head, pre-trained on the PASCAL VOC dataset

A lightweight image classification model pre-trained on the ImageNet-1k dataset, specifically optimized for mobile devices

MobileNet V2 is a lightweight convolutional neural network designed for mobile devices, excelling in image classification tasks.

MobileNet V2 is a lightweight convolutional neural network designed for mobile devices, excelling in image classification tasks.

MobileNet V1 is a lightweight convolutional neural network designed for mobile and embedded vision applications, pretrained on the ImageNet-1k dataset.

A lightweight semantic segmentation model pre-trained on the PASCAL VOC dataset, combining MobileViT and DeepLabV3 architectures

A lightweight vision Transformer model combining MobileNetV2 and Transformer modules, suitable for mobile semantic segmentation tasks.

Lightweight vision Transformer model combining MobileNetV2 and Transformer modules, suitable for mobile semantic segmentation tasks

MobileViT is a lightweight, low-latency vision Transformer model that combines the strengths of CNNs and Transformers, making it suitable for mobile devices.

TensorFlow 2 and PyTorch-based Optical Character Recognition (OCR) model supporting text detection and recognition in document images

A Q&A model based on the MobileBERT architecture, fine-tuned for the SQuAD v2 dataset

MobileBERT is a lightweight variant of BERT, optimized for mobile devices, featuring efficient inference speed and a compact model size.

MobileBERT is a lightweight version of BERT_LARGE, fine-tuned on the SQuAD2.0 dataset for question answering systems.