home.social

#mixtureofexperts — Public Fediverse posts

Live and recent posts from across the Fediverse tagged #mixtureofexperts, aggregated by home.social.

  1. Nemotron 3 Super pushes the frontier with 40 M supervised & alignment samples, leveraging a Mamba‑Transformer backbone and Mixture‑of‑Experts scaling. The model shows stronger agent reasoning, RL‑based fine‑tuning, and tighter AI alignment. Dive into the details to see how this LLM reshapes open‑source AI. #Nemotron3 #MixtureOfExperts #AIAlignment #SupervisedFineTuning

    🔗 aidailypost.com/news/nemotron-

  2. Nemotron 3 Super pushes the frontier with 40 M supervised & alignment samples, leveraging a Mamba‑Transformer backbone and Mixture‑of‑Experts scaling. The model shows stronger agent reasoning, RL‑based fine‑tuning, and tighter AI alignment. Dive into the details to see how this LLM reshapes open‑source AI. #Nemotron3 #MixtureOfExperts #AIAlignment #SupervisedFineTuning

    🔗 aidailypost.com/news/nemotron-

  3. Nemotron 3 Super pushes the frontier with 40 M supervised & alignment samples, leveraging a Mamba‑Transformer backbone and Mixture‑of‑Experts scaling. The model shows stronger agent reasoning, RL‑based fine‑tuning, and tighter AI alignment. Dive into the details to see how this LLM reshapes open‑source AI. #Nemotron3 #MixtureOfExperts #AIAlignment #SupervisedFineTuning

    🔗 aidailypost.com/news/nemotron-

  4. Nemotron 3 Super pushes the frontier with 40 M supervised & alignment samples, leveraging a Mamba‑Transformer backbone and Mixture‑of‑Experts scaling. The model shows stronger agent reasoning, RL‑based fine‑tuning, and tighter AI alignment. Dive into the details to see how this LLM reshapes open‑source AI. #Nemotron3 #MixtureOfExperts #AIAlignment #SupervisedFineTuning

    🔗 aidailypost.com/news/nemotron-

  5. Alibaba just released the Qwen‑3.5‑Medium model as open‑source, delivering Sonnet 4.5‑level performance on a single GPU. It uses a Mixture‑of‑Experts architecture and a new “Thinking Mode” to boost AI inference efficiency while staying lightweight. Dive into the details and see how this could reshape open‑source LLM development. #Qwen3_5 #OpenSourceLLM #MixtureOfExperts #ModelEfficiency

    🔗 aidailypost.com/news/alibaba-o

  6. Alibaba just released the Qwen‑3.5‑Medium model as open‑source, delivering Sonnet 4.5‑level performance on a single GPU. It uses a Mixture‑of‑Experts architecture and a new “Thinking Mode” to boost AI inference efficiency while staying lightweight. Dive into the details and see how this could reshape open‑source LLM development. #Qwen3_5 #OpenSourceLLM #MixtureOfExperts #ModelEfficiency

    🔗 aidailypost.com/news/alibaba-o

  7. Alibaba just released the Qwen‑3.5‑Medium model as open‑source, delivering Sonnet 4.5‑level performance on a single GPU. It uses a Mixture‑of‑Experts architecture and a new “Thinking Mode” to boost AI inference efficiency while staying lightweight. Dive into the details and see how this could reshape open‑source LLM development. #Qwen3_5 #OpenSourceLLM #MixtureOfExperts #ModelEfficiency

    🔗 aidailypost.com/news/alibaba-o

  8. NVIDIA’s new co‑design with Sarvam AI slashes time‑to‑first‑token to under a second for LLM inference. By marrying Mixture‑of‑Experts models with GPU acceleration, they boost throughput while trimming latency. This hardware‑software synergy could reshape how we deploy large language models at scale. Read more to see the numbers and tech behind the breakthrough. #NVIDIA #SarvamAI #MixtureOfExperts #TTFT

    🔗 aidailypost.com/news/nvidia-co

  9. NVIDIA’s new co‑design with Sarvam AI slashes time‑to‑first‑token to under a second for LLM inference. By marrying Mixture‑of‑Experts models with GPU acceleration, they boost throughput while trimming latency. This hardware‑software synergy could reshape how we deploy large language models at scale. Read more to see the numbers and tech behind the breakthrough. #NVIDIA #SarvamAI #MixtureOfExperts #TTFT

    🔗 aidailypost.com/news/nvidia-co

  10. NVIDIA’s new co‑design with Sarvam AI slashes time‑to‑first‑token to under a second for LLM inference. By marrying Mixture‑of‑Experts models with GPU acceleration, they boost throughput while trimming latency. This hardware‑software synergy could reshape how we deploy large language models at scale. Read more to see the numbers and tech behind the breakthrough. #NVIDIA #SarvamAI #MixtureOfExperts #TTFT

    🔗 aidailypost.com/news/nvidia-co

  11. Alibaba's new Qwen 3.5 397B-A17 outperforms even larger rivals by using multi-token prediction and a sparse mixture-of-experts architecture. It cuts inference cost while keeping top-tier performance, hinting at a new era for multimodal AI. Curious how 397 billion parameters can be cheaper? Read the full story. #Qwen3_5 #AlibabaAI #MixtureOfExperts #MultiTokenPrediction

    🔗 aidailypost.com/news/alibabas-

  12. Alibaba's new Qwen 3.5 397B-A17 outperforms even larger rivals by using multi-token prediction and a sparse mixture-of-experts architecture. It cuts inference cost while keeping top-tier performance, hinting at a new era for multimodal AI. Curious how 397 billion parameters can be cheaper? Read the full story. #Qwen3_5 #AlibabaAI #MixtureOfExperts #MultiTokenPrediction

    🔗 aidailypost.com/news/alibabas-

  13. Alibaba's new Qwen 3.5 397B-A17 outperforms even larger rivals by using multi-token prediction and a sparse mixture-of-experts architecture. It cuts inference cost while keeping top-tier performance, hinting at a new era for multimodal AI. Curious how 397 billion parameters can be cheaper? Read the full story. #Qwen3_5 #AlibabaAI #MixtureOfExperts #MultiTokenPrediction

    🔗 aidailypost.com/news/alibabas-

  14. MiniMax's new M2.5 model slashes costs to 1/20 of Claude Opus while handling 30% of HQ tasks. Built on a Mixture‑of‑Experts sparse architecture, it delivers strong code‑generation and LLM performance—all open‑source. Discover how this AI agent could boost productivity in your projects. #MiniMaxM2_5 #MixtureOfExperts #OpenSourceAI #AIProductivity

    🔗 aidailypost.com/news/minimaxs-

  15. MiniMax's new M2.5 model slashes costs to 1/20 of Claude Opus while handling 30% of HQ tasks. Built on a Mixture‑of‑Experts sparse architecture, it delivers strong code‑generation and LLM performance—all open‑source. Discover how this AI agent could boost productivity in your projects. #MiniMaxM2_5 #MixtureOfExperts #OpenSourceAI #AIProductivity

    🔗 aidailypost.com/news/minimaxs-

  16. MiniMax's new M2.5 model slashes costs to 1/20 of Claude Opus while handling 30% of HQ tasks. Built on a Mixture‑of‑Experts sparse architecture, it delivers strong code‑generation and LLM performance—all open‑source. Discover how this AI agent could boost productivity in your projects. #MiniMaxM2_5 #MixtureOfExperts #OpenSourceAI #AIProductivity

    🔗 aidailypost.com/news/minimaxs-

  17. MoLE (Mixture of Lookup Experts) hứa hẹn giúp chạy mô hình lớn trên máy本地 bằng cách dồn tải vào SSD với tốc độ cao. Dù vậy, việc huấn luyện MoLE tốn kém và công ty khó áp dụng ngay. Đề xuất: tinh chỉnh mô hình đã huấn luyện sẵn (pre-trained) thành MoLE để tận dụng lợi thế lưu trữ và hiệu năng. Cần nghiên cứu cách giảm số expert mỗi token mà vẫn giữ chất lượng. Thử nghiệm trên model nhỏ như Qwen3 30B trước. #MoLE #LocalLLM #AI #FineTuning #MixtureOfExperts #AIResearch #MoLE #LLM #FineTuning #AIRe

  18. Blackwell Ultra promises a major AI speed boost, while Nvidia’s Vera Rubin platform aims to slash token costs for LLM inference with Mixture‑of‑Experts. Both moves could reshape large‑model deployment. Curious how this will affect your workloads? Read the full breakdown. #BlackwellUltra #VeraRubin #MixtureOfExperts #TokenCost

    🔗 aidailypost.com/news/blackwell

  19. Nvidia just unveiled Nemotron 3, its new Mamba-Transformer LLM built for massive token throughput and Mixture-of-Experts scaling. Early adopters like Accenture, Oracle Cloud Infrastructure and Zoom are already testing agentic-AI capabilities. Curious how this could reshape open-source AI? Read on! #Nvidia #Nemotron3 #MambaTransformer #MixtureOfExperts

    🔗 aidailypost.com/news/nvidia-la

  20. Nvidia unveils an AI server running mixture-of-experts models up to 1,000x faster with 72 high-speed GPUs. As AI shifts to real-time, large-scale deployment, competitors like AMD and Chinese AI firms challenge its lead. Nvidia aims to stay ahead in hardware, scalability, and operational efficiency.

    #Nvidia #AIHardware #MixtureOfExperts #GPUs #HighPerformanceComputing

    Read Full Article:- techi.com/nvidia-boost-moonsho

  21. Nvidia unveils an AI server running mixture-of-experts models up to 1,000x faster with 72 high-speed GPUs. As AI shifts to real-time, large-scale deployment, competitors like AMD and Chinese AI firms challenge its lead. Nvidia aims to stay ahead in hardware, scalability, and operational efficiency.

    Read Full Article:- techi.com/nvidia-boost-moonsho

  22. Nvidia unveils an AI server running mixture-of-experts models up to 1,000x faster with 72 high-speed GPUs. As AI shifts to real-time, large-scale deployment, competitors like AMD and Chinese AI firms challenge its lead. Nvidia aims to stay ahead in hardware, scalability, and operational efficiency.

    #Nvidia #AIHardware #MixtureOfExperts #GPUs #HighPerformanceComputing

    Read Full Article:- techi.com/nvidia-boost-moonsho