๐ ๐ฒ๐ถ๐๐๐ฎ๐ป ๐ข๐ฝ๐ฒ๐ป-๐ฆ๐ผ๐๐ฟ๐ฐ๐ฒ๐ ๐๐ผ๐ป๐ด๐๐ฎ๐-๐ฉ๐ถ๐ฑ๐ฒ๐ผ-๐๐๐ฎ๐๐ฎ๐ฟ ๐ญ.๐ฒ: ๐ฆ๐๐ฎ๐ฏ๐น๐ฒ ๐ญ๐ฌ+ ๐ ๐ถ๐ป๐๐๐ฒ ๐ง๐ฎ๐น๐ธ๐ถ๐ป๐ด ๐๐๐ฎ๐๐ฎ๐ฟ๐ ๐๐ถ๐๐ต ๐ฅ๐ + ๐๐ถ๐๐๐ถ๐น๐น๐ฎ๐๐ถ๐ผ๐ป
One photo + audio โ natural lip-synced video that holds identity and motion past 10 minutes.
Whatโs new under the hood:
โข Refined post-training pipeline: RL (GRPO-style), improved CFG for DMD-based few-step distillation, long-video rollouts, and on-policy distillation.
โข Two LoRA adapters (DMD + OPSD) mixed across an 8-step distilled schedule early steps for structure, later steps for detail.
โข Reduced autoregressive error accumulation so identity, body structure, and hand-object interactions stay stable far longer.
โข Inference optimizations: optional weight-only INT8 DiT (~15 GB vs ~30 GB BF16), FP8 path, Torch compile, context parallelism, FlashAttention-2/3.
Open weights (MIT), same base architecture as 1.5, ranks #2 in their human preference arena (Elo 1068.1).
Weights: huggingface.co/meituan-longcโฆ
Project page: longcatavatar.com/longcat-avโฆ
#AI #VideoGeneration #OpenSource #LongCat #Meituan
1
๐๐ฎ๐ถ๐ฑ๐ ๐๐๐ข ๐ฅ๐ผ๐ฏ๐ถ๐ป ๐๐ถ ๐ฎ๐ ๐๐ฃ๐๐: ๐ง๐ต๐ฒ ๐ข๐ป๐ฒ ๐ช๐ผ๐ฟ๐ฑ ๐ณ๐ผ๐ฟ ๐๐ ๐ถ๐ป ๐ฎ๐ฌ๐ฎ๐ฒ ๐๐ โ๐จ๐ฟ๐ด๐ฒ๐ป๐ฐ๐โ
Tech is moving fast. Investment and compute spending are exploding. Token use is skyrocketing.
But most companies still arenโt seeing real revenue or profit growth from AI.
The foundation layer is a red ocean. The real opportunity is in the application layer the โsuper appโ that actually gets people using AI and making money.
Whoever finds it first wins the next round. If applications donโt take off, the bubble risk is real.
#AI #Baidu #APEC #AIApplications
1
18
๐๐ฒ๐ถ ๐๐๐ป ๐ฎ๐ ๐๐ฃ๐๐: ๐๐โ๐ ๐ฅ๐ฒ๐ฎ๐น ๐๐ฟ๐ฒ๐ฎ๐ธ๐๐ต๐ฟ๐ผ๐๐ด๐ต ๐ช๐ถ๐น๐น ๐๐ฒ ๐ถ๐ป ๐๐ต๐ฒ ๐ฃ๐ต๐๐๐ถ๐ฐ๐ฎ๐น ๐ช๐ผ๐ฟ๐น๐ฑ
Not in chat. In phones, cars, and factories.
Xiaomiโs factories already run fully automated with AI. Next step: industrial agents that can produce, spot problems, and improve on their own.
Their embodied robot just did an 8-hour factory shift and charged itself.
AI moving into the physical world is the real game-changer.
#AI #Xiaomi #EmbodiedAI #PhysicalAI #APEC
1
45
๐ง๐ต๐ฒ ๐ช๐ผ๐ฟ๐น๐ฑโ๐ ๐๐ฎ๐๐๐ฒ๐๐ ๐ฆ๐๐ฝ๐ฒ๐ฟ๐ฐ๐ผ๐บ๐ฝ๐๐๐ฒ๐ฟ ๐ฅ๐๐ป๐ ๐ผ๐ป ๐๐ฃ๐จ๐ ๐ก๐ผ๐ ๐๐ฃ๐จ๐.
Chinaโs LineShine is doing something unusual in the age of GPU-dominated AI.
It has secured the No. 1 spot in the TOP500 with 2.198 exaflops, powered by 13.78 million CPU cores.
But the more interesting achievement is what researchers are doing with it:
โข Simulating a silicon crystal containing 200 million atoms.
โข Running large-scale quantum-level material simulations.
โข Achieving 96.6% weak-scaling efficiency in a 100-million-atom study.
This could help scientists study materials and battery technologies at a scale that was previously out of reach.
The bigger story: the future of computing isnโt just about faster AI chips. Itโs also about solving scientific problems at an entirely different scale.
#Supercomputing #AI #HPC #Research
1
5
103
SGLang Low-Latency Mode: Making LLM Inference Faster
SGLang, NVIDIA and AMD are working on a new approach to reduce LLM inference latency.
Instead of relying only on standard kernels, they introduce specialized GPU kernels that reduce launch overhead, overlap computation with communication, and optimize small-batch decoding.
Key results:
Up to 1.38ร faster token generation on NVIDIA GB300.
Up to 1.56ร faster on AMD MI355X.
NVIDIA optimizations improve per-user decode speed by 12โ38% in the tested workloads.
The interesting part is the plugin architecture: specialized kernels can integrate with SGLang without putting every hardware-specific optimization into its core codebase.
Code:
NVIDIA: github.com/sgl-project/sgl-nโฆ
AMD: github.com/sgl-project/sgl-aโฆ
Read: lmsys.org/blog/2026-10-18-sgโฆ
#LLM #Inference #SGLang #GPU #AI
BridgeKV: Faster LLM Inference Without Losing Large-Model Quality
Thatโs the idea behind BridgeKV.
It translates the large modelโs KV cache into a format the smaller model can use, allowing the smaller model to handle decoding.
Key results:
Up to 3ร faster requests in a real A40 GPU test.
Retains much of the larger modelโs quality.
Works across six model pairs and five datasets.
A promising approach to making LLM serving faster and more efficient.
Paper: arxiv.org/abs/2611.11761
#LLM #Inference #AI #MachineLearning
Wow cool and can we able to call unrestricted and uncenscored models too like these ones because if so then it would be game changer.
We're teaming up to accelerate open-source Al.
OpenRouter is now integrated into Hugging Face directly. Same token, same endpoint, plus live model stats on the Hub and an inference widget we redesigned together.
The best of open, in one place. huggingface.co/blog/openroutโฆ
1
83
LingBot-Touch: Giving Robots a Sense of Touch.
Robbyant has introduced LingBot-Touch, a tactile foundation model designed to help robots understand physical contact across different sensor types.
Key highlights:
Trained on nearly 5,000 hours of tactile data and 1.33 billion frames.
Unifies magnetic, piezoresistive and optical tactile sensors.
Improves robot manipulation success by up to 26 percentage points across 26 real-world tasks.
Transfers to previously unseen tactile sensors without retraining.
The bigger idea: Robots need more than vision. Touch could be the missing piece for truly dexterous manipulation.
Website: technology.robbyant.com/lingโฆ
GitHub: github.com/robbyant/lingbot-โฆ
Models: huggingface.co/collections/rโฆ
Research: arxiv.org/abs/2611.12806
#Robotics #EmbodiedAI #AI #FoundationModels #TactileSensing
EVAThink: A Video AI Model That Knows When to Think and When to Answer
Researchers from East China Normal University and Huawei introduced EVAThink, a training method that helps video AI models decide when deeper reasoning is actually needed.
The idea is simple:
Enough visual evidence? Answer directly without wasting tokens.
Limited visual evidence? Think through the available clues before answering.
How does it learn?
Researchers create video pairs with the same question but different levels of visual evidence, then use two-stage reinforcement learning (GRPO) to teach the model when to use each response mode.
Results:
Trained using just 6.7K samples.
Achieved 65.6% average accuracy across six video benchmarks with 64 frames.
Outperformed both direct-only and think-only modes.
Reduced output length by 62% compared with always using reasoning.
Built on Qwen3-VL-8B-Instruct.
The bigger idea: Better video AI isn't about thinking more. It's about knowing when thinking is worth the extra computation.
Paper: arxiv.org/abs/2611.13907
1
12
420







