Tech Meridian ← ENTITY INDEX
PROMY MERIDIAN RU

COMPANY · ENTITY #6222

Metal

Related event timeline, sources and context from the news index.

EVENT TIMELINE

4

CODING · 1 SOURCE · InfoQ AI, ML & Data Engineering

Google launches LiteRT Tensor API for building and optimizing on-device ML in C++

Google introduced the LiteRT Tensor API, a lightweight C++ API that lets developers build computation graphs operation-by-operation for on-device execution across CPU, GPU, and NPU backends, with features like direct weight ingestion (including .safetensors), zero-copy buffer interoperability, and compatibility with AI coding agents. Google demonstrated the API with a SAM 2 multi-object segmentation and tracking app and reported cross-platform benchmarks (e.g., about 1.5x faster than Apple MLX on a MacBook Pro in one test) and listed planned enhancements such as NPU-friendly ops and automatic full-graph quantization.

7.0

COMPANIES · 1 SOURCE · InfoQ AI, ML & Data Engineering

Google open-sources ML Drift GPU engine for on-device AI inference

Google released ML Drift, an Apache-2.0 open-source, cross-platform GPU compute engine for on-device AI and ML inference that abstracts multiple low-level APIs (OpenGL ES, OpenCL, Metal, WebGPU). ML Drift powers LiteRT, adds unified shaders via tensor virtualization, extensible custom ops, 5D tensor support, and hardware-specific optimizations for Arm, Intel, and Qualcomm GPUs; Google says it replaces the legacy TensorFlow Lite GPU delegate and reports lower memory overhead and app-level latency improvements in Chrome, YouTube Shorts, Photos, Adobe apps, and Snapchat.

8.0

CODING · 1 SOURCE · Hugging Face

PyTorch 2.15 brings FP8, faster LLM decoding and new Metal MPS kernels for Macs

PyTorch 2.15 (Hugging Face contributors) improves the MPS backend on Apple Silicon: adds torch.float8_e4m3fn (FP8) support with F.scaled_mm, faster single‑row linear and grouped matmul paths that speed Mixture‑of‑Experts decoding 5–6.5× (e.g., Qwen3.6‑35B‑A3B from 9.4→61.3 tok/s on an M3 Ultra), an ~8× faster softmax for sampling, and many new Metal kernels that cut driver memory for some video decoders (Wan2.2) from ~270 GiB to ~8 GiB. The release also raises platform requirements to macOS 15+ and Python 3.11+, and previews further MPSGraph→Metal migrations in 2.16.

7.0

MODELS · 1 SOURCE · Hugging Face

Transformers adds GGUF (llama.cpp quant) support on main

Hugging Face's transformers (main branch) can now load and run GGUF quantized checkpoints using llama.cpp/ggml kernels, aiming for efficient local inference (initially optimized for Apple Silicon and Qwen3.5 models). The integration also enables serving GGUF checkpoints via transformers-serve and falls back to dequantization if compatible kernels are unavailable.

7.0