Menu

#Tensor

13 posts

Feed·
12 of 13 posts
Boosting MoE Training Throughput with Advanced Fusion Kernels
🖼️
0Read More

Boosting MoE Training Throughput with Advanced Fusion Kernels

NVIDIA Technical Blog·Rachit Garg·5 days ago
#ZtXDYIe5
#developer#kernels#kernel#activation#tensor

Mixture-of-experts (MoE) models have quickly become a foundational component of modern, large-scale AI systems. They are widely adopted because they enable…

15s
Google Launches LiteRT Tensor API for Native On-Device ML Development
🖼️
0Read More

Google Launches LiteRT Tensor API for Native On-Device ML Development

#webpronews#news#google#model#litert

Google's new LiteRT Tensor API lets developers hand-build on-device models op by op in C++ and run the same graph on CPU, GPU or NPU, or export a .tflite file.…

15s
The Pixel 12’s Tensor G7 chip just leaked, and it already has a name
🖼️
0Read More

The Pixel 12’s Tensor G7 chip just leaked, and it already has a name

Phandroid·Tyler Lee·6 months ago
#nwgcYMBG
#mnmd#comments#google#pixel#tensor

The Tensor G7 powering the Pixel 12 has a codename, and Google is staying the course on custom silicon. Here's what the leak tells us.

15s