Chinese food-service chains adopt AI as a new 'operating partner' - Global Times
Chinese food-service chains adopt AI as a new 'operating partner' Global Times
Could not retrieve the full article text.
Read on Google News - Meituan AI →Google News - Meituan AI
https://news.google.com/rss/articles/CBMiYkFVX3lxTE5TVzNYTHVyZ2ZpWWhOa01NNEZMR1pzRjNadkJWelgzcnNZMElBVU1vbDdCYnh3eHhmVU9hQkJlSVRnT2tsTVhLaE1hY3NUQnZsX0JhNk1jQXAtYnNPd0Z1ZVdB?oc=5Sign in to highlight and annotate this article
Conversation starters
Daily AI Digest
Get the top 5 AI stories delivered to your inbox every morning.
More about
serviceglobal
The Navy brought a retired laser weapon back for a new drone fight
The U.S. Navy spent at least six months resurrecting a high-energy laser weapon that previously graced the bow of a warship for a new military exercise last year, the service recently revealed. The Navy’s Directed Energy Systems Integration Laboratory (DESIL)—the dedicated facility for evaluating laser weapons in a maritime environment located at Naval Base Ventura County in Point Mugu, California—“ramped up efforts to restore critical functions” to the service’s “one-of-a-kind” 150 kilowatt Solid State Laser Technology Maturation (SSL-TM) demonstrator starting in early March 2025, according to recently published ‘year in review’ bulletin from Naval Sea Systems Command (NAVSEA). Initiated in 2012 and officially known as the Laser Weapon System Demonstrator Mk 2 Mod 0, the SSL-TM demonstrat
Gemma 4 31B at 256K Full Context on a Single RTX 5090 — TurboQuant KV Cache Benchmark
Just got Gemma 4 31B running at full 256K context on a single RTX 5090 using TurboQuant KV cache compression. System Specs Component Spec GPU NVIDIA GeForce RTX 5090 (32GB VRAM) CPU AMD Ryzen 9 9950X3D (16-core) RAM 64GB DDR5 OS Windows 11 Setup Model : gemma-4-31B-it-UD-Q4_K_XL from Unsloth (17.46 GiB) Build : TheTom/llama-cpp-turboquant branch feature/turboquant-kv-cache , merged with latest upstream master for Gemma 4 support KV Cache : turbo3 (3-bit PolarQuant + Hadamard rotation, ~4.5x compression vs f16) Config : --n-gpu-layers 99 --no-mmap --flash-attn on --cache-type-k turbo3 --cache-type-v turbo3 Benchmark Results Test Speed (t/s) pp4096 3,362.71 pp16384 3,047.00 pp65536 2,077.96 pp131072 1,428.80 pp262144 899.55 tg128 61.51 VRAM usage at 262K : 27.7 GB / 32 GB (4.3 GB headroom) G
Knowledge Map
Connected Articles — Knowledge Graph
This article is connected to other articles through shared AI topics and tags.
More in Products
The Navy brought a retired laser weapon back for a new drone fight
The U.S. Navy spent at least six months resurrecting a high-energy laser weapon that previously graced the bow of a warship for a new military exercise last year, the service recently revealed. The Navy’s Directed Energy Systems Integration Laboratory (DESIL)—the dedicated facility for evaluating laser weapons in a maritime environment located at Naval Base Ventura County in Point Mugu, California—“ramped up efforts to restore critical functions” to the service’s “one-of-a-kind” 150 kilowatt Solid State Laser Technology Maturation (SSL-TM) demonstrator starting in early March 2025, according to recently published ‘year in review’ bulletin from Naval Sea Systems Command (NAVSEA). Initiated in 2012 and officially known as the Laser Weapon System Demonstrator Mk 2 Mod 0, the SSL-TM demonstrat
trunk/9557db4da58adc85b48136aad2e383f7ab6e456e: [xpu] Add deterministic matmul test to test_gemm.py (#179004)
Summary Adds test_matmul_deterministic_mode to test/xpu/test_gemm.py to verify that torch.matmul produces bit-identical results across repeated calls when deterministic algorithms are enabled via torch.use_deterministic_algorithms(True) . Motivation XPU implementation allows using deterministic mode for matmul, so we should test if it produce correct results. The two shapes ( 513 and 767 ) are chosen to cover non-power-of-two sizes that commonly expose non-determinism in GPU tiling strategies. This PR is a follow up task after conversation in intel/torch-xpu-ops#3201 Changes test/xpu/test_gemm.py Added DeterministicGuard to the torch.testing._internal.common_utils import block. Added test_matmul_deterministic_mode to TestBasicGEMM . Pull Request resolved: #179004 Approved by: https://githu
Discussion
Sign in to join the discussion
No comments yet — be the first to share your thoughts!