📄 文章列表
23 篇文章
cn_ai
浙大00后博士生再交卷!把十几种物理要素装进一个世界模型|华为昇腾原生
cn_ai
百度开源文心5.5 Flash Base,多基准测试刷新同尺寸模型 SOTA
AlphaSignal
Baidu Drops ERNIE 5.5 Flash, a Raw 280B Open Checkpoint Beating DeepSeek
AlphaSignal
Institute of Foundation Models Opens Full Training Stack Behind K2 Horizon 360
cn_ai
英伟达团队基于 MiniMax H3 开源视频增强方案,最高支持 4K 重生成
VentureBeat
GLM-5.3 is here with advanced cyber capabilities — and reportedly already found a 'serious vulnerability' in Cursor
AlphaSignal
PyTorch 2.13 Ships 12x Faster Attention and 4x Memory Savings for LLM Training | AlphaSignal
TDS
Tokenminning: How to Get More from Your Chatbot for Less
AlphaSignal
Z.ai's ZCode Ships a Desktop Agent Built Around GLM-5.2's 1M-Token Window | AlphaSignal
AlphaSignal
Top Agentic Performance, Low Cost: GLM-5.2, designed for coding and long-running agentic jobs, now the top open model
AlphaSignal
Claude Opus 4.8 advertises its uncertainty: Dynamic workflows manage Anthropic subagents
AlphaSignal
NVIDIA Shrinks GLM-5.2 Memory by 1.8x With NVFP4 Without Losing Accuracy | AlphaSignal
AlphaSignal
Z.ai's GLM-5.2 Hits 1M Context at 10x Cheaper Than Claude | AlphaSignal
AlphaSignal
SGLang Breaks 12,000 Tokens per Second on DeepSeek V4 Pro | AlphaSignal
AlphaSignal
NVIDIA's PixelDiT Beats FLUX at 1024px Without a VAE Bottleneck | AlphaSignal
AlphaSignal
Meta's TLX Block Attention Beats Flash Attention v2 by 2.3x on Blackwell GPUs | AlphaSignal
TDS
Positional Embeddings in Transformers: A Math Guide to RoPE & ALiBi
TDS
Using Vision Language Models to Process Millions of Documents
TDS
RoPE, Clearly Explained
TDS
TDS Newsletter: January Must-Reads on Data Platforms, Infinite Context, and More
1
2