On-device
Apple's MobileCLIP-S0 encodes an image in 1.5 ms on an iPhone 12 Pro Max
Apple timed its image-text models on an iPhone and released four variants, the weights and the reinforced DataCompDR dataset.
BUPT measures 22 LLMs on four Android phones at about 200 ms per token
A BUPT team ran 22 models from 0.5B to 7B on four Android phones with llama.cpp and found 7B models take about 4 GB and 200 ms per token.
Apple Intelligence pairs a 3-billion-parameter on-device model with a server model
Apple reports 0.6 ms per prompt token and 30 tokens per second on an iPhone 15 Pro for a model compressed to an average of 3.7 bits per weight.
PowerInfer-2 runs a 47B model on a OnePlus 12 at 11.68 tokens per second
Shanghai Jiao Tong University researchers report a 47B model decoding at 11.68 tokens per second on a OnePlus 12, with weights streamed from flash.
BUPT proposes one 9.2B model in the OS that all apps call through adapters
A BUPT team proposes that the phone OS ship one 9.2B multimodal model all apps call through small adapters, and matched app models on 85% of 50 datasets.
ExecuTorch alpha runs Llama 2 7B on iPhone 15 Pro and Galaxy phones
PyTorch's edge runtime brought 4-bit Llama 2 7B to iPhone and Galaxy handsets, added early Llama 3 8B support and leaned on Apple, Arm and Qualcomm.
Apple releases OpenELM at 270M to 3B, with parameters spread unevenly across layers
Apple published four models from 270M to 3B parameters with the full training framework and code to run them through MLX on Apple silicon.
Microsoft runs Phi-3-mini offline on an iPhone 14 at over 12 tokens per second
The 3.8-billion-parameter model takes about 1.8 GB at 4-bit and scores 69 percent on MMLU, which Microsoft compares to Mixtral 8x7B and GPT-3.5.
Octopus v3 picks an action from an image and a query in under 1B parameters
A technical report describes a sub-billion multimodal agent that turns an image and a text query into a function call, shown through ten worked examples.
Octopus v2 is a 2B model that calls Android APIs with one token per function
A 2B Gemma fine-tune gives every Android API its own token, and the authors report 99.524% accuracy and 0.38 seconds per call, ahead of GPT-4.
One shared on-device LLM keeps a context per app and switches in 0.27 seconds
Peking University, BUPT and Tsinghua researchers compress and swap each app conversation context in 16-token chunks, switching up to 20 times faster.
Google ships an experimental MediaPipe LLM Inference API for web, Android and iOS
The experimental API runs Gemma 2B, Phi 2, Falcon 1B and Stable LM 3B fully on device, with int8 and int4 weights.