［ninfer v100X2］v100 achieves new glories! qwen3.8 27b nvfp4 2000t/s prefill, 115t/s decode
［Ninfer V100X2］v100 achieves glory again! qwen3.8 27b nvfp4 peak 236t/s decoding, 2000t/s prefill
Dual-card v100, the strongest inference engine in history has arrived! First in the world, second in China.
[Ninfer V100X2] Peak 68t/s! Dual V100 16g running qwen3.8 27b q4 quantization 180k context, achieving 68t/s decoding?
The strongest 4-card V100 inference engine in history! Number one in the world, number two in China!
A graphics card that is 10,000 times better than the 5090
《全网首发》使用显卡演奏全首千本樱
终极千问部署方案=ninfer+kvmem 首发！
【Save Money Tips】A 500-yuan NVIDIA V100 16G can run vllm: qwen3.8 27b, with speeds up to 200 tokens/s
GPT-SoVITs voice changer!??
"Debut" review of the 192GB memory AI Max+495 mini PC FEVM FAEX2, running Deepseek V4 Flash on a single unit, the price is bound to get flamed! 1 liter with 160W TDP is really pushing it!
Local deployment of Qwen3.8-Flash-Next-Q4_K_M with 12+64, supporting 1M context
[Open Source under GPL3.0] Kotlin & torchscript refactored mobile GPT-SoViTs Runner, supports OpenAI API, supports original GSV model conversion
Running a 27B large model on 16GB VRAM? Qwen3.8-27B local test: 40T/s, this time it's truly legendary!
AI Server with 8x 5090 32G, 8x 5090 GPUs—Can You Handle It?
The most cost-effective plan for deploying DeepSeek v4 Flash - ShuMeng AI 6-card V100 32G, single concurrent 25token/s, 3 concurrent 15, openclaw large model deployment
A 4x V100 32GB (128GB total VRAM) AI workstation for just ¥1500 – besides Supermicro X10Q, who else can do it?
amd yes! Using the 7900xtx to push qwen3.8-27b to 97t/s! 256K context with vision!
Memory slaughter! A brief test of single-card local deployment of Qwen3.8-Flash-Next-NVFP4: 110 for single shot, 270 for three shots, up to 11000 for prefill
Using only 5GB VRAM! Deploying Qwen3.6-35B-A3B + Visual Model Locally on a Laptop with RTX 3060 6GB
