Crazy ML (@crazyMLguy) header
Crazy ML (@crazyMLguy) profile picture

Machine Learning & Backend ๐Ÿ˜Ž Founder of Humanity backed by Investors from Google, SF and Hong Kong.

Mountains
Joined June 2023
๐— ๐—ฒ๐—ถ๐˜๐˜‚๐—ฎ๐—ป ๐—ข๐—ฝ๐—ฒ๐—ป-๐—ฆ๐—ผ๐˜‚๐—ฟ๐—ฐ๐—ฒ๐˜€ ๐—Ÿ๐—ผ๐—ป๐—ด๐—–๐—ฎ๐˜-๐—ฉ๐—ถ๐—ฑ๐—ฒ๐—ผ-๐—”๐˜ƒ๐—ฎ๐˜๐—ฎ๐—ฟ ๐Ÿญ.๐Ÿฒ: ๐—ฆ๐˜๐—ฎ๐—ฏ๐—น๐—ฒ ๐Ÿญ๐Ÿฌ+ ๐— ๐—ถ๐—ป๐˜‚๐˜๐—ฒ ๐—ง๐—ฎ๐—น๐—ธ๐—ถ๐—ป๐—ด ๐—”๐˜ƒ๐—ฎ๐˜๐—ฎ๐—ฟ๐˜€ ๐˜„๐—ถ๐˜๐—ต ๐—ฅ๐—Ÿ + ๐——๐—ถ๐˜€๐˜๐—ถ๐—น๐—น๐—ฎ๐˜๐—ถ๐—ผ๐—ป One photo + audio โ†’ natural lip-synced video that holds identity and motion past 10 minutes. Whatโ€™s new under the hood: โ€ข Refined post-training pipeline: RL (GRPO-style), improved CFG for DMD-based few-step distillation, long-video rollouts, and on-policy distillation. โ€ข Two LoRA adapters (DMD + OPSD) mixed across an 8-step distilled schedule early steps for structure, later steps for detail. โ€ข Reduced autoregressive error accumulation so identity, body structure, and hand-object interactions stay stable far longer. โ€ข 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-longcโ€ฆ Project page: longcatavatar.com/longcat-avโ€ฆ #AI #VideoGeneration #OpenSource #LongCat #Meituan
1
๐—•๐—ฎ๐—ถ๐—ฑ๐˜‚ ๐—–๐—˜๐—ข ๐—ฅ๐—ผ๐—ฏ๐—ถ๐—ป ๐—Ÿ๐—ถ ๐—ฎ๐˜ ๐—”๐—ฃ๐—˜๐—–: ๐—ง๐—ต๐—ฒ ๐—ข๐—ป๐—ฒ ๐—ช๐—ผ๐—ฟ๐—ฑ ๐—ณ๐—ผ๐—ฟ ๐—”๐—œ ๐—ถ๐—ป ๐Ÿฎ๐Ÿฌ๐Ÿฎ๐Ÿฒ ๐—œ๐˜€ โ€œ๐—จ๐—ฟ๐—ด๐—ฒ๐—ป๐—ฐ๐˜†โ€ Tech is moving fast. Investment and compute spending are exploding. Token use is skyrocketing. But most companies still arenโ€™t seeing real revenue or profit growth from AI. The foundation layer is a red ocean. The real opportunity is in the application layer the โ€œsuper appโ€ that actually gets people using AI and making money. Whoever finds it first wins the next round. If applications donโ€™t take off, the bubble risk is real. #AI #Baidu #APEC #AIApplications
1
18
๐—Ÿ๐—ฒ๐—ถ ๐—๐˜‚๐—ป ๐—ฎ๐˜ ๐—”๐—ฃ๐—˜๐—–: ๐—”๐—œโ€™๐˜€ ๐—ฅ๐—ฒ๐—ฎ๐—น ๐—•๐—ฟ๐—ฒ๐—ฎ๐—ธ๐˜๐—ต๐—ฟ๐—ผ๐˜‚๐—ด๐—ต ๐—ช๐—ถ๐—น๐—น ๐—•๐—ฒ ๐—ถ๐—ป ๐˜๐—ต๐—ฒ ๐—ฃ๐—ต๐˜†๐˜€๐—ถ๐—ฐ๐—ฎ๐—น ๐—ช๐—ผ๐—ฟ๐—น๐—ฑ Not in chat. In phones, cars, and factories. Xiaomiโ€™s 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
๐—ง๐—ต๐—ฒ ๐—ช๐—ผ๐—ฟ๐—น๐—ฑโ€™๐˜€ ๐—™๐—ฎ๐˜€๐˜๐—ฒ๐˜€๐˜ ๐—ฆ๐˜‚๐—ฝ๐—ฒ๐—ฟ๐—ฐ๐—ผ๐—บ๐—ฝ๐˜‚๐˜๐—ฒ๐—ฟ ๐—ฅ๐˜‚๐—ป๐˜€ ๐—ผ๐—ป ๐—–๐—ฃ๐—จ๐˜€ ๐—ก๐—ผ๐˜ ๐—š๐—ฃ๐—จ๐˜€. Chinaโ€™s 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: โ€ข Simulating a silicon crystal containing 200 million atoms. โ€ข Running large-scale quantum-level material simulations. โ€ข 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 isnโ€™t just about faster AI chips. Itโ€™s 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 12โ€“38% 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-nโ€ฆ AMD: github.com/sgl-project/sgl-aโ€ฆ Read: lmsys.org/blog/2026-10-18-sgโ€ฆ #LLM #Inference #SGLang #GPU #AI
3
3
17
808
BridgeKV: Faster LLM Inference Without Losing Large-Model Quality Thatโ€™s the idea behind BridgeKV. It translates the large modelโ€™s 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 modelโ€™s 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
8
58
1,836
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/openroutโ€ฆ
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/lingโ€ฆ GitHub: github.com/robbyant/lingbot-โ€ฆ Models: huggingface.co/collections/rโ€ฆ Research: arxiv.org/abs/2611.12806 #Robotics #EmbodiedAI #AI #FoundationModels #TactileSensing
2
1
7
319
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