Video wird geladen...

Video konnte nicht geladen werden

Beim Laden dieses Videos ist ein Problem aufgetreten. Dies könnte an einem vorübergehenden Netzwerkproblem liegen oder das Video ist möglicherweise nicht verfügbar.

We open-sourced QeRL — Quantization-enhanced Reinforcement Learning ! 🧠 4-bit quantized RL training 💪 Train a 32B LLM on a single H100 GPU ⚙️ 1.7× faster overall training 🎯 Accuracy on par with bfloat16-level accuracy 🔥 Supports NVFP4 quantization format Moreover, we show that quantization helps exploration in RL... show more

Yukang Chen

1,416 subscribers

69,720 Aufrufe • vor 7 Monaten •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

0 Kommentare

Keine Kommentare verfügbar

Kommentare vom Original-Post werden hier angezeigt

Ähnliche Videos

Long video generation is a systems problem. Introducing LongLive-2.0 from NVIDIA Research: an end-to-end NVFP4 training and inference system for long video generation. Low-precision deployment often relies on post-training quantization, creating a gap between how models are trained and how they run. LongLive-2.0 aligns NVFP4-aware training, distillation, and W4A4 inference, maintaining strong benchmark quality while improving speed and memory efficiency.

Long video generation is a systems problem. Introducing LongLive-2.0 from NVIDIA Research: an end-to-end NVFP4 training and inference system for long video generation. Low-precision deployment often relies on post-training quantization, creating a gap between how models are trained and how they run. LongLive-2.0 aligns NVFP4-aware training, distillation, and W4A4 inference, maintaining strong benchmark quality while improving speed and memory efficiency.

NVIDIA AI

60,087 Aufrufe • vor 14 Tagen

LTX-2.3 in 4.5s? A 5s 1080p video gen in a blink? FastVideo makes video gen feel almost interactive. - On ONE GPU - NVFP4 quantization + fused Blackwell kernels. - Yeah… a B200 Try this madness yourself

LTX-2.3 in 4.5s? A 5s 1080p video gen in a blink? FastVideo makes video gen feel almost interactive. - On ONE GPU - NVFP4 quantization + fused Blackwell kernels. - Yeah… a B200 Try this madness yourself

Wildminder

21,462 Aufrufe • vor 2 Monaten

DeepSeek R1 671B running on 2 M2 Ultras faster than reading speed. Getting close to open-source O1, at home, on consumer hardware. With mlx.distributed and mlx-lm, 3-bit quantization (~4 bpw)

DeepSeek R1 671B running on 2 M2 Ultras faster than reading speed. Getting close to open-source O1, at home, on consumer hardware. With mlx.distributed and mlx-lm, 3-bit quantization (~4 bpw)

Awni Hannun

865,941 Aufrufe • vor 1 Jahr

Our work, "A Primer on SO(3) Action Representations in Deep Reinforcement Learning," was accepted to #ICLR2026! We provide a systematic study of action representation choices in RL, showing that they fundamentally impact training stability and performance. #Robotics #AI #RL

Our work, "A Primer on SO(3) Action Representations in Deep Reinforcement Learning," was accepted to #ICLR2026! We provide a systematic study of action representation choices in RL, showing that they fundamentally impact training stability and performance. #Robotics #AI #RL

Learning Systems and Robotics Lab (is hiring!)

49,596 Aufrufe • vor 3 Monaten

🚀 The 4-bit era has arrived! Meet #SVDQuant, our new W4A4 quantization paradigm for diffusion models. Now, 12B FLUX can run on a 16GB 4090 laptop without offloading—with 3x speedups over W4A16 models (like NF4) while maintaining top-tier image quality. #AI #Quantization. 1/7

🚀 The 4-bit era has arrived! Meet #SVDQuant, our new W4A4 quantization paradigm for diffusion models. Now, 12B FLUX can run on a 16GB 4090 laptop without offloading—with 3x speedups over W4A16 models (like NF4) while maintaining top-tier image quality. #AI #Quantization. 1/7

Muyang Li

48,890 Aufrufe • vor 1 Jahr

🤯Real-time video generation just got HUGE. Introducing Helios: A 14B parameter model running at 19.5 FPS on a single H100. Cheaper, Faster, and Stronger than 1.3B models, generating minute-long videos without the usual tricks (even no quantization).

🤯Real-time video generation just got HUGE. Introducing Helios: A 14B parameter model running at 19.5 FPS on a single H100. Cheaper, Faster, and Stronger than 1.3B models, generating minute-long videos without the usual tricks (even no quantization).

Bin Lin

27,663 Aufrufe • vor 3 Monaten

We asked Sholto Douglas from Anthropic about the costs of RL (Reinforcement Learning) runs. "In Dario Amodei's essay, he said that RL runs cost only $1M back in December." "RL is a more naively parallelizable and scalable than pre-training." "With pre-training, you need everything in one big data center ideally. For RL, in theory, you could scale all over the world."

We asked Sholto Douglas from Anthropic about the costs of RL (Reinforcement Learning) runs. "In Dario Amodei's essay, he said that RL runs cost only $1M back in December." "RL is a more naively parallelizable and scalable than pre-training." "With pre-training, you need everything in one big data center ideally. For RL, in theory, you could scale all over the world."

TBPN

76,634 Aufrufe • vor 1 Jahr

🚨 New Paper Training an LLM to speak low-resource language (EACL workshop, 2026) Tulu is spoken by 2M+ people in coastal Karnataka and LLMs basically can't speak it. We got to 85% grammar accuracy without fine-tuning anything or collecting a single new training example.

🚨 New Paper Training an LLM to speak low-resource language (EACL workshop, 2026) Tulu is spoken by 2M+ people in coastal Karnataka and LLMs basically can't speak it. We got to 85% grammar accuracy without fine-tuning anything or collecting a single new training example.

Lossfunk

120,485 Aufrufe • vor 2 Monaten

LLMs can take gigabytes of memory to store, which limits what can be run on consumer hardware. But quantization can dramatically compress models, making a wider selection of models available to developers. You can often reduce model size by 4x or more while maintaining reasonable performance. In our new short course Quantization Fundamentals taught by Hugging Face's @younesbelkada and Marc Sun, you'll: - Learn how to quantize nearly any open source model - Use int8 and bfloat16 (Brain float 16) data types to load and run LLMs using PyTorch and the Hugging Face Transformers library - Dive into the technical details of linear quantization to map 32-bit floats to 8-bit integers As models get bigger and bigger, quantization becomes more important for making models practical and accessible. Please check out the course here:

LLMs can take gigabytes of memory to store, which limits what can be run on consumer hardware. But quantization can dramatically compress models, making a wider selection of models available to developers. You can often reduce model size by 4x or more while maintaining reasonable performance. In our new short course Quantization Fundamentals taught by Hugging Face's @younesbelkada and Marc Sun, you'll: - Learn how to quantize nearly any open source model - Use int8 and bfloat16 (Brain float 16) data types to load and run LLMs using PyTorch and the Hugging Face Transformers library - Dive into the technical details of linear quantization to map 32-bit floats to 8-bit integers As models get bigger and bigger, quantization becomes more important for making models practical and accessible. Please check out the course here:

Andrew Ng

288,261 Aufrufe • vor 2 Jahren

Reasoning LLMs generate very long chains-of-thought, so even small quantization errors add up. With AWQ, Qwen3-4B drops 71.0 → 68.2 on MMLU-Pro (~4% relative loss). 😬 ParoQuant fixes this! It keeps only the critical rotation pairs and fuses everything into a single kernel. Recovers most of the lost reasoning accuracy with minimal overhead — so 4-bit models stay strong at reasoning. 💪💪

Reasoning LLMs generate very long chains-of-thought, so even small quantization errors add up. With AWQ, Qwen3-4B drops 71.0 → 68.2 on MMLU-Pro (~4% relative loss). 😬 ParoQuant fixes this! It keeps only the critical rotation pairs and fuses everything into a single kernel. Recovers most of the lost reasoning accuracy with minimal overhead — so 4-bit models stay strong at reasoning. 💪💪

Zhijian Liu

170,758 Aufrufe • vor 3 Monaten

ParoQuant just got a big upgrade 🚀 ✅ Supports the new Qwen3.5 models ⚡ Now runs on MLX (fast local inference on Apple Silicon) 🧠 Preserves reasoning quality with 4-bit quantization We also built an agent demo running locally on my 4-year-old M2 Max. Can't wait to upgrade to an M5 Max and see what kind of magic we can do. ✨

ParoQuant just got a big upgrade 🚀 ✅ Supports the new Qwen3.5 models ⚡ Now runs on MLX (fast local inference on Apple Silicon) 🧠 Preserves reasoning quality with 4-bit quantization We also built an agent demo running locally on my 4-year-old M2 Max. Can't wait to upgrade to an M5 Max and see what kind of magic we can do. ✨

Zhijian Liu

49,004 Aufrufe • vor 3 Monaten

What if you could train AI agents on a laptop as easily as on a GPU cluster? Researchers from UIUC's U Lab, led by Prof. Jiaxuan You, just open-sourced OpenTinker. It's a new "Reinforcement-Learning-as-a-Service" (RLaaS) system that decouples the complex training pipeline into simple, distributed services with friendly APIs. The result? It breaks down the major engineering barriers to RL, outperforming traditional frameworks in accessibility and ease of deployment, finally making agent training viable for more developers and teams. Project: Code: U Lab: Our report: 📬 #PapersAccepted by Jiqizhixin

What if you could train AI agents on a laptop as easily as on a GPU cluster? Researchers from UIUC's U Lab, led by Prof. Jiaxuan You, just open-sourced OpenTinker. It's a new "Reinforcement-Learning-as-a-Service" (RLaaS) system that decouples the complex training pipeline into simple, distributed services with friendly APIs. The result? It breaks down the major engineering barriers to RL, outperforming traditional frameworks in accessibility and ease of deployment, finally making agent training viable for more developers and teams. Project: Code: U Lab: Our report: 📬 #PapersAccepted by Jiqizhixin

机器之心 JIQIZHIXIN

15,862 Aufrufe • vor 4 Monaten

We teamed up with NVIDIA and Matthew Berman to teach you how to do Reinforcement Learning! Learn about: - RL environments, reward functions & reward hacking - Training OpenAI gpt-oss to automatically solve 2048 - Local Windows training with NVIDIA_AI_PC RTX GPUs - How RLVR (verifiable rewards) works - How to interpret RL metrics like KL Divergence Full video tutorial:

We teamed up with NVIDIA and Matthew Berman to teach you how to do Reinforcement Learning! Learn about: - RL environments, reward functions & reward hacking - Training OpenAI gpt-oss to automatically solve 2048 - Local Windows training with NVIDIA_AI_PC RTX GPUs - How RLVR (verifiable rewards) works - How to interpret RL metrics like KL Divergence Full video tutorial:

Unsloth AI

54,026 Aufrufe • vor 5 Monaten

🤔Want a principled way to RL your diffusion model? Check Data-regularized Reinforcement Learning (DDRL)! Post-train NVIDIA #Cosmos World Foundation models with a million GPU hours! 🤯 Novel formulation ➡️ Theoretically integrates SFT into RL ➡️ Robust to Reward Hacking 🛑 Details: #DDRL #Diffusion #RL #NVIDIA #Cosmos

🤔Want a principled way to RL your diffusion model? Check Data-regularized Reinforcement Learning (DDRL)! Post-train NVIDIA #Cosmos World Foundation models with a million GPU hours! 🤯 Novel formulation ➡️ Theoretically integrates SFT into RL ➡️ Robust to Reward Hacking 🛑 Details: #DDRL #Diffusion #RL #NVIDIA #Cosmos

Haotian Ye

77,422 Aufrufe • vor 6 Monaten

(1/5) FP4 hardware is here, but 4-bit attention still kills model quality, blocking true end-to-end FP4 serving. To fix that, we propose Attn-QAT, the first systematic study of quantization-aware training for attention. The result: FP4 attention quality is comparable to BF16 attention with 1.1x–1.5x higher throughput than SageAttention3 on an RTX 5090 and 1.39x speedup over FlashAttention-4 on a B200. Blog: Code: Checkpoints:

(1/5) FP4 hardware is here, but 4-bit attention still kills model quality, blocking true end-to-end FP4 serving. To fix that, we propose Attn-QAT, the first systematic study of quantization-aware training for attention. The result: FP4 attention quality is comparable to BF16 attention with 1.1x–1.5x higher throughput than SageAttention3 on an RTX 5090 and 1.39x speedup over FlashAttention-4 on a B200. Blog: Code: Checkpoints:

Hao AI Lab

37,175 Aufrufe • vor 1 Monat

1/ Introducing RL Swarm’s new backend: GenRL. A modular reinforcement learning library built for distributed, fault-tolerant training - now powering RL Swarm from the ground up. 🧵

1/ Introducing RL Swarm’s new backend: GenRL. A modular reinforcement learning library built for distributed, fault-tolerant training - now powering RL Swarm from the ground up. 🧵

gensyn

82,118 Aufrufe • vor 11 Monaten

OpenAI shows how gpt-oss can autonomously beat 2048 using reinforcement learning (RL). Training was done locally with Unsloth on NVIDIA DGX Spark. You can also do it free on Colab. 🦥 OpenAI DevDay notebook:

OpenAI shows how gpt-oss can autonomously beat 2048 using reinforcement learning (RL). Training was done locally with Unsloth on NVIDIA DGX Spark. You can also do it free on Colab. 🦥 OpenAI DevDay notebook:

Unsloth AI

98,526 Aufrufe • vor 7 Monaten

NeurIPS 2025 Paper: LLMs are Reinforcement Learners 🤯! Surprisingly, we show that LLMs can solve RL tasks without any external component! We introduce Prompted Policy Search (ProPS), an RL method based only LLMs and in-context learning. [Paper]

NeurIPS 2025 Paper: LLMs are Reinforcement Learners 🤯! Surprisingly, we show that LLMs can solve RL tasks without any external component! We introduce Prompted Policy Search (ProPS), an RL method based only LLMs and in-context learning. [Paper]

Heni Ben Amor

51,172 Aufrufe • vor 6 Monaten

◈@Exla_ai speeds up AI on edge devices (e.g. NVIDIA Jetsons) with advanced quantization. Reduce memory use by up to 80% and boost inference speed by 3x–20x— all with just a few lines of code. Congrats on the launch, Viraat & Pranav!

◈@Exla_ai speeds up AI on edge devices (e.g. NVIDIA Jetsons) with advanced quantization. Reduce memory use by up to 80% and boost inference speed by 3x–20x— all with just a few lines of code. Congrats on the launch, Viraat & Pranav!

Y Combinator

19,937 Aufrufe • vor 1 Jahr

We are releasing Star Elastic - turn ONE reasoning LLM into MANY sizes with a single post-training run. 360× cheaper than pretraining a family of models. 7× better than SOTA compression. Split reasoning capability. Plus elastic budget control that beats the accuracy-latency frontier. Paper: HF models: Thread 👇

We are releasing Star Elastic - turn ONE reasoning LLM into MANY sizes with a single post-training run. 360× cheaper than pretraining a family of models. 7× better than SOTA compression. Split reasoning capability. Plus elastic budget control that beats the accuracy-latency frontier. Paper: HF models: Thread 👇

Pavlo Molchanov

21,313 Aufrufe • vor 23 Tagen