Video yükleniyor...

Video Yüklenemedi

Ana Sayfaya Dön

🚀 Introducing Next-Concept Modeling (NCM): our new paradigm for fast, parallel, coarse-to-fine text generation. Compared to strong diffusion language models (DLMs) at similar parameter size, NCM achieves: ⚡ Up to 31× faster generation on downstream tasks like summarization and machine translation with comparable or better quality. 🏎️ 10–21× faster...

KumbongHermann's profile picture

Hermann Kumbong

2,171 subscribers

219,267 görüntüleme • 1 gün önce •via X (Twitter)

EğitimBilim & Teknoloji
LIVE

Anya Rossi• Live Now

Private livecam show

31 Yorum

Hermann Kumbong profil fotoğrafı
Hermann Kumbong1 gün önce

Quality at the Fraction of the Cost: NCM pushes the non-autoregressive quality–efficiency frontier. We evaluate NCM across 3 groups of tasks and compare against strong diffusion language model baselines at similar parameter scales: 🏎️ Unconditional generation: TinyStories, LM1B, and OpenWebText. ⚡ Conditional generation: XSum summarization and WMT14 machine translation. 🧩 Constraint-based reasoning: Sudoku across three difficulty levels. We summarize some of our key findings in the figure below.

Hermann Kumbong profil fotoğrafı
Hermann Kumbong1 gün önce

🧩 How does NCM work? NCM follows a simple two-stage approach: 1️⃣ Learn: A multiscale VQ-VAE learns a hierarchy of discrete text representations, from coarse concepts to fine details. Prior work (VAR) already does this in images but getting it to work well with text requires many substantial modeling changes. 2️⃣ Generate: A transformer predicts the hierarchy level by level, generating all codes within each level in parallel.

Hermann Kumbong profil fotoğrafı
Hermann Kumbong1 gün önce

✨ Coarse-to-Fine Generation What does NCM's generation trajectory actually look like? To understand how text emerges across the hierarchy, we decode the intermediate latent representations into text at each generation step. NCM progressively adds finer details to the text as it traverses its hierarchy

Hermann Kumbong profil fotoğrafı
Hermann Kumbong1 gün önce

⚡ Why is NCM so efficient? NCM's inference efficiency comes from two key advantages: 1️⃣ Fewer sampling steps: NCM achieves high generation quality in far fewer steps. 2️⃣ Cheaper steps: Unlike diffusion models (e.g Duo, MDLM etc) that process the full sequence at every step, NCM processes shorter representations at early levels, reducing the average computation per step. Fewer steps + cheaper steps = substantially faster generation! 🚀

Hermann Kumbong profil fotoğrafı
Hermann Kumbong1 gün önce

📉 Strong Generation Quality with Fewer Training Tokens Across a wide range of tasks & datasets, NCM achieves strong quality while training on 3x-5x fewer tokens compared to its diffusion language baselines. One possible explanation for this is that NCM learns a single pre-set text ordering while many of the diffusion baselines have to learn way more different orderings of the text.

Hermann Kumbong profil fotoğrafı
Hermann Kumbong1 gün önce

📈 Scaling We explore two directions for scaling NCM: 🧠 Model Scaling: Increasing NCM's model size consistently improves both generation quality and diversity, even at a fixed training-token budget. ⚡ Test-Time Scaling: NCM's fast generation allows us to sample more candidates while still being cheaper or at the same cost of diffusion baselines. We use this to further improving generation quality.

Hermann Kumbong profil fotoğrafı
Hermann Kumbong1 gün önce

🧩Coarse-to-Fine Reasoning Across different difficulty levels on sudoku, NCM outperforms both AR and diffusion baselines.

Hermann Kumbong profil fotoğrafı
Hermann Kumbong1 gün önce

🎁 Bonus NCM cranking a sudoku puzzle! It iteratively refines its solutions arriving at the correct solved board at the end.

Hermann Kumbong profil fotoğrafı
Hermann Kumbong1 gün önce

🚀 NCM is open source! We're excited to share the paper, code, checkpoints, and datasets with the community. Paper: Blog 1: Blog 2: Code: Checkpoints/Datasets: alphaXiv:

Hermann Kumbong profil fotoğrafı
Hermann Kumbong1 gün önce

Thanks to @modal and @LambdaAPI for compute support towards this project.

Qizheng Zhang @ COLM 2026 profil fotoğrafı
Qizheng Zhang @ COLM 20261 gün önce

@IddahMlauzi @HazyResearch @Azaliamirh @simran_s_arora congrats on the release 🔥🔥🔥

Justin Deschenaux profil fotoğrafı
Justin Deschenaux1 gün önce

@IddahMlauzi @HazyResearch @Azaliamirh @simran_s_arora Looks cool! Just curious, did you get a chance to try it on tinyGSM?

Hermann Kumbong profil fotoğrafı
Hermann Kumbong1 gün önce

@IddahMlauzi @HazyResearch @Azaliamirh @simran_s_arora Thank you!! Not yet, but we will soon!

Justin Deschenaux profil fotoğrafı
Justin Deschenaux1 gün önce

@IddahMlauzi @HazyResearch @Azaliamirh @simran_s_arora Great, let me know what comes out!

Hermann Kumbong profil fotoğrafı
Hermann Kumbong1 gün önce

@IddahMlauzi @HazyResearch @Azaliamirh @simran_s_arora Will do

Eric Nguyen profil fotoğrafı
Eric Nguyen1 gün önce

@IddahMlauzi @HazyResearch @Azaliamirh @simran_s_arora This is really cool research, congrats Hermann! ⚡️🔥

Yohei profil fotoğrafı
Yohei23 saat önce

@IddahMlauzi @HazyResearch @Azaliamirh @simran_s_arora this is so weird I love it

elio profil fotoğrafı
elio23 saat önce

@IddahMlauzi @HazyResearch @Azaliamirh @simran_s_arora incredible work! this is the future

David W. Romero profil fotoğrafı
David W. Romero22 saat önce

@IddahMlauzi @HazyResearch @Azaliamirh @simran_s_arora Looks very cool! Congratulations @KumbongHermann !!!!

Hermann Kumbong profil fotoğrafı
Hermann Kumbong22 saat önce

@IddahMlauzi @HazyResearch @Azaliamirh @simran_s_arora Thank you so much David!! I learned a lot of this from you 🙏🙏

David W. Romero profil fotoğrafı
David W. Romero22 saat önce

@IddahMlauzi @HazyResearch @Azaliamirh @simran_s_arora Haha I wish I could take credit! ;) This is truly awesome! Super proud of you!!!

太阳 profil fotoğrafı
太阳23 saat önce

@IddahMlauzi @HazyResearch @Azaliamirh @simran_s_arora 31 倍加速还不掉质量，这 coarse-to-fine 的思路是真漂亮😂

zappy profil fotoğrafı
zappy23 saat önce

@IddahMlauzi @HazyResearch @Azaliamirh @simran_s_arora Incredible release!

太阳 profil fotoğrafı
太阳23 saat önce

@IddahMlauzi @HazyResearch @Azaliamirh @simran_s_arora 先粗后细 31 倍速，这加速我是服气的⚡

Austin DeWolfe profil fotoğrafı
Austin DeWolfe23 saat önce

@IddahMlauzi @HazyResearch @Azaliamirh @simran_s_arora I have been building something similar for looped AR models. This line of work is definitely the future of the field.

Dylan profil fotoğrafı
Dylan23 saat önce

@IddahMlauzi @HazyResearch @Azaliamirh @simran_s_arora really cool!! might take a stab and try NCM on other puzzle tasks.

Hermann Kumbong profil fotoğrafı
Hermann Kumbong23 saat önce

@IddahMlauzi @HazyResearch @Azaliamirh @simran_s_arora Sounds great! Let me know if you need help setting up

Charles Ding profil fotoğrafı
Charles Ding22 saat önce

@IddahMlauzi @HazyResearch @Azaliamirh @simran_s_arora Thanks for the huge effort Hermann! One of the coolest projects to be a part of

Jean-Christophe Plantier profil fotoğrafı
Jean-Christophe Plantier22 saat önce

@IddahMlauzi @HazyResearch @Azaliamirh @simran_s_arora Real full in bio 🔥

Amir Zohrenejad profil fotoğrafı
Amir Zohrenejad21 saat önce

@IddahMlauzi @HazyResearch @Azaliamirh @simran_s_arora Very cool. Congrata @KumbongHermann

Hermann Kumbong profil fotoğrafı
Hermann Kumbong21 saat önce

@IddahMlauzi @HazyResearch @Azaliamirh @simran_s_arora Thank you!

Benzer Videolar

Introducing HRM-Text. An ultra-lean 1B-parameter reasoning language model designed to deliver strong general performance with a fraction of the data, compute, and infrastructure. Trained on just 40B structured tokens, HRM-Text achieves competitive performance while using ~1/1000 of the training data of comparable models. The kicker? The full model trains in roughly one day on a $1,000 budget. This opens the door to a new generation of AI that is powerful, accessible, and radically easier to adapt. Theories and research concepts once deemed too expensive to test are officially back in the game. Sapient Intelligence invites you to help us shape a new paradigm for general intelligence.
5:31

Introducing HRM-Text. An ultra-lean 1B-parameter reasoning language model designed to deliver strong general performance with a fraction of the data, compute, and infrastructure. Trained on just 40B structured tokens, HRM-Text achieves competitive performance while using ~1/1000 of the training data of comparable models. The kicker? The full model trains in roughly one day on a $1,000 budget. This opens the door to a new generation of AI that is powerful, accessible, and radically easier to adapt. Theories and research concepts once deemed too expensive to test are officially back in the game. Sapient Intelligence invites you to help us shape a new paradigm for general intelligence.

Sapient Intelligence

518,646 görüntüleme • 4 ay önce

Text-to-3D Generation with Bidirectional Diffusion using both 2D and 3D priors paper page: Most 3D generation research focuses on up-projecting 2D foundation models into the 3D space, either by minimizing 2D Score Distillation Sampling (SDS) loss or fine-tuning on multi-view datasets. Without explicit 3D priors, these methods often lead to geometric anomalies and multi-view inconsistency. Recently, researchers have attempted to improve the genuineness of 3D objects by directly training on 3D datasets, albeit at the cost of low-quality texture generation due to the limited texture diversity in 3D datasets. To harness the advantages of both approaches, we propose Bidirectional Diffusion(BiDiff), a unified framework that incorporates both a 3D and a 2D diffusion process, to preserve both 3D fidelity and 2D texture richness, respectively. Moreover, as a simple combination may yield inconsistent generation results, we further bridge them with novel bidirectional guidance. In addition, our method can be used as an initialization of optimization-based models to further improve the quality of 3D model and efficiency of optimization, reducing the generation process from 3.4 hours to 20 minutes. Experimental results have shown that our model achieves high-quality, diverse, and scalable 3D generation.
0:44

Text-to-3D Generation with Bidirectional Diffusion using both 2D and 3D priors paper page: Most 3D generation research focuses on up-projecting 2D foundation models into the 3D space, either by minimizing 2D Score Distillation Sampling (SDS) loss or fine-tuning on multi-view datasets. Without explicit 3D priors, these methods often lead to geometric anomalies and multi-view inconsistency. Recently, researchers have attempted to improve the genuineness of 3D objects by directly training on 3D datasets, albeit at the cost of low-quality texture generation due to the limited texture diversity in 3D datasets. To harness the advantages of both approaches, we propose Bidirectional Diffusion(BiDiff), a unified framework that incorporates both a 3D and a 2D diffusion process, to preserve both 3D fidelity and 2D texture richness, respectively. Moreover, as a simple combination may yield inconsistent generation results, we further bridge them with novel bidirectional guidance. In addition, our method can be used as an initialization of optimization-based models to further improve the quality of 3D model and efficiency of optimization, reducing the generation process from 3.4 hours to 20 minutes. Experimental results have shown that our model achieves high-quality, diverse, and scalable 3D generation.

AK

38,853 görüntüleme • 2 yıl önce

Presto! Distilling Steps and Layers for Accelerating Music Generation Despite advances in diffusion-based text-to-music (TTM) methods, efficient, high-quality generation remains a challenge. We introduce Presto!, an approach to inference acceleration for score-based diffusion transformers via reducing both sampling steps and cost per step. To reduce steps, we develop a new score-based distribution matching distillation (DMD) method for the EDM-family of diffusion models, the first GAN-based distillation method for TTM. To reduce the cost per step, we develop a simple, but powerful improvement to a recent layer distillation method that improves learning via better preserving hidden state variance. Finally, we combine our step and layer distillation methods together for a dual-faceted approach. We evaluate our step and layer distillation methods independently and show each yield best-in-class performance. Our combined distillation method can generate high-quality outputs with improved diversity, accelerating our base model by 10-18x (230/435ms latency for 32 second mono/stereo 44.1kHz, 15x faster than comparable SOTA) -- the fastest high-quality TTM to our knowledge.
1:20

Presto! Distilling Steps and Layers for Accelerating Music Generation Despite advances in diffusion-based text-to-music (TTM) methods, efficient, high-quality generation remains a challenge. We introduce Presto!, an approach to inference acceleration for score-based diffusion transformers via reducing both sampling steps and cost per step. To reduce steps, we develop a new score-based distribution matching distillation (DMD) method for the EDM-family of diffusion models, the first GAN-based distillation method for TTM. To reduce the cost per step, we develop a simple, but powerful improvement to a recent layer distillation method that improves learning via better preserving hidden state variance. Finally, we combine our step and layer distillation methods together for a dual-faceted approach. We evaluate our step and layer distillation methods independently and show each yield best-in-class performance. Our combined distillation method can generate high-quality outputs with improved diversity, accelerating our base model by 10-18x (230/435ms latency for 32 second mono/stereo 44.1kHz, 15x faster than comparable SOTA) -- the fastest high-quality TTM to our knowledge.

AK

30,446 görüntüleme • 2 yıl önce

GPT-5 on Sudoku-Bench 🧩 Since releasing Sudoku-Bench in May 2025, when no LLM could solve a classic 9x9 puzzle, we've been evaluating the latest generation of models. GPT-5 now leads our leaderboard with 33% puzzles solved--approximately 2x the previous leader--and is the first LLM we've tested to solve a 9x9 Sudoku variant. However, with 67% of the much harder puzzles remaining unsolved, Sudoku-Bench continues to present significant challenges for AI reasoning. Modern Sudoku variants require models to first understand novel rulesets through meta-reasoning, then maintain global consistency across long reasoning chains. Our experiments with GRPO fine-tuning on Qwen2.5-7b and "Thought Cloning" (training on expert human reasoning from Cracking the Cryptic) show that current approaches still struggle with the spatial reasoning and creative "break-in" points that human solvers use naturally. We believe new approaches are required to solve our benchmark. These results highlight persistent gaps between computational problem-solving and human-like reasoning, particularly in tasks requiring integrated mathematical logic, spatial awareness, and creative insight. Read more about our update here: 🔗 Blogpost →
0:28

GPT-5 on Sudoku-Bench 🧩 Since releasing Sudoku-Bench in May 2025, when no LLM could solve a classic 9x9 puzzle, we've been evaluating the latest generation of models. GPT-5 now leads our leaderboard with 33% puzzles solved--approximately 2x the previous leader--and is the first LLM we've tested to solve a 9x9 Sudoku variant. However, with 67% of the much harder puzzles remaining unsolved, Sudoku-Bench continues to present significant challenges for AI reasoning. Modern Sudoku variants require models to first understand novel rulesets through meta-reasoning, then maintain global consistency across long reasoning chains. Our experiments with GRPO fine-tuning on Qwen2.5-7b and "Thought Cloning" (training on expert human reasoning from Cracking the Cryptic) show that current approaches still struggle with the spatial reasoning and creative "break-in" points that human solvers use naturally. We believe new approaches are required to solve our benchmark. These results highlight persistent gaps between computational problem-solving and human-like reasoning, particularly in tasks requiring integrated mathematical logic, spatial awareness, and creative insight. Read more about our update here: 🔗 Blogpost →

Sakana AI

154,970 görüntüleme • 11 ay önce

Wrapping up the year and coinciding with #NeurIPS2024, today at Meta FAIR we’re releasing a collection of nine new open source AI research artifacts across our work in developing agents, robustness & safety and new architectures. More in the video from Joelle Pineau. All of this work is part of FAIR’s continued work towards the goal of achieving advanced machine intelligence A few highlights from what we’re releasing today: • Meta Motivo: A first-of-its-kind behavioral foundation model that controls the movements of a virtual embodied humanoid agent to perform complex tasks. • Meta Video Seal: a state-of-the art comprehensive framework for neural video watermarking. • Meta Explore Theory-of-Mind: A program-guided adversarial data generation for theory of mind reasoning. • Meta Large Concept Models: A fundamentally different training paradigm for language modeling that decouples reasoning from language representation. And much more! We’re excited to share this work with the research community and look forward to seeing how it inspires new innovation across the field. Details and access to everything released by FAIR today ➡️
1:34

Wrapping up the year and coinciding with #NeurIPS2024, today at Meta FAIR we’re releasing a collection of nine new open source AI research artifacts across our work in developing agents, robustness & safety and new architectures. More in the video from Joelle Pineau. All of this work is part of FAIR’s continued work towards the goal of achieving advanced machine intelligence A few highlights from what we’re releasing today: • Meta Motivo: A first-of-its-kind behavioral foundation model that controls the movements of a virtual embodied humanoid agent to perform complex tasks. • Meta Video Seal: a state-of-the art comprehensive framework for neural video watermarking. • Meta Explore Theory-of-Mind: A program-guided adversarial data generation for theory of mind reasoning. • Meta Large Concept Models: A fundamentally different training paradigm for language modeling that decouples reasoning from language representation. And much more! We’re excited to share this work with the research community and look forward to seeing how it inspires new innovation across the field. Details and access to everything released by FAIR today ➡️

AI at Meta

156,123 görüntüleme • 1 yıl önce

We’re excited to announce the release and open-source of HunyuanImage 3.0 — the largest and most powerful open-source text-to-image model to date, with over 80 billion total parameters, of which 13 billion are activated per token during inference.The effect is completely comparable to the industry’s flagship closed-source model.🚀🚀🚀 HunyuanImage 3.0 originates from our internally developed native multimodal large language model, with fine-tuning and post-training focused on text-to-image generation. This unique foundation gives the model a powerful set of capabilities: ✅Reason with world knowledge ✅Understand complex, thousand-word prompts ✅Generate precise text within images Different from traditional DiT architecture image generation models, HunyuanImage 3.0’s MoE architecture uses a Transfusion-based approach to deeply couple Diffusion and LLM training for a single, powerful system. Built on Hunyuan-A13B, HunyuanImage 3.0 was trained on a massive dataset: 5 billion image-text pairs, video frames, interleaved image-text data, and 6 trillion tokens of text corpora. This hybrid training across multimodal generation, understanding, and LLM capabilities allows the model to seamlessly integrate multiple tasks. Whether you're an illustrator, designer, or creator, this is built to slash your workflow from hours to minutes. HunyuanImage 3.0 can generate intricate text, detailed comics, expressive emojis, and lively, engaging illustrations for educational content. The current release focuses solely on text-to-image generation and future updates will include image-to-image, image editing, multi-turn interaction, and more. 👉🏻Try it now: 🔗GitHub: 🤗Hugging Face:
2:09

We’re excited to announce the release and open-source of HunyuanImage 3.0 — the largest and most powerful open-source text-to-image model to date, with over 80 billion total parameters, of which 13 billion are activated per token during inference.The effect is completely comparable to the industry’s flagship closed-source model.🚀🚀🚀 HunyuanImage 3.0 originates from our internally developed native multimodal large language model, with fine-tuning and post-training focused on text-to-image generation. This unique foundation gives the model a powerful set of capabilities: ✅Reason with world knowledge ✅Understand complex, thousand-word prompts ✅Generate precise text within images Different from traditional DiT architecture image generation models, HunyuanImage 3.0’s MoE architecture uses a Transfusion-based approach to deeply couple Diffusion and LLM training for a single, powerful system. Built on Hunyuan-A13B, HunyuanImage 3.0 was trained on a massive dataset: 5 billion image-text pairs, video frames, interleaved image-text data, and 6 trillion tokens of text corpora. This hybrid training across multimodal generation, understanding, and LLM capabilities allows the model to seamlessly integrate multiple tasks. Whether you're an illustrator, designer, or creator, this is built to slash your workflow from hours to minutes. HunyuanImage 3.0 can generate intricate text, detailed comics, expressive emojis, and lively, engaging illustrations for educational content. The current release focuses solely on text-to-image generation and future updates will include image-to-image, image editing, multi-turn interaction, and more. 👉🏻Try it now: 🔗GitHub: 🤗Hugging Face:

Tencent Hy

413,272 görüntüleme • 1 yıl önce

We've officially released and open-sourced HunyuanImage 2.1, our latest text-to-image model. The new model delivers on our commitment to balancing performance and quality. With native 2K image generation, HunyuanImage 2.1 is an advanced open-source text-to-image model.🎨 ✨ New in 2.1: 🔹Advanced Semantics: Supports ultra-long and complex prompts of up to 1000 tokens, and precisely controls the generation of multiple subjects in a single image. 🔹Precise Chinese and English Text Rendering with seamless image–text integration: The model naturally integrates text into images, making it suitable for a wide range of applications such as product covers, illustrations, and poster design to meet the needs of various fields. 🔹Rich Styles and High Aesthetic: Capable of generating images in various styles—including photorealistic portraits, comics, and vinyl figures—it delivers outstanding visual appeal and artistic quality. 🔹High-Quality Generation: Efficiently produces ultra-high-definition (2K) images in the same time other models take to generate a 1K image. HunyuanImage 2.1 uses two text encoders: a multimodal large language model (MLLM) to improve the model's image and text alignment capabilities, and a multi-language character-aware encoder to improve text rendering capabilities. The model is a single- and double-stream diffusion transformer with 17B parameters. We've also open-sourced the weights of the the accelerated version with meanflow which reduces inference steps from 100 to just 8, and PromptEnhancer, the first industrial-grade rewriting model that enhances your prompts for more nuanced and expressive image generation. Now, creators turn complex ideas—like posters with slogans or multi-panel comics—into visuals faster than ever. We’re just getting started. Stay tuned for our native multimodal image generation model coming soon. 🌐Website: 🔗Github: 🤗Hugging Face: ✨Hugging Face Demo:
1:39

We've officially released and open-sourced HunyuanImage 2.1, our latest text-to-image model. The new model delivers on our commitment to balancing performance and quality. With native 2K image generation, HunyuanImage 2.1 is an advanced open-source text-to-image model.🎨 ✨ New in 2.1: 🔹Advanced Semantics: Supports ultra-long and complex prompts of up to 1000 tokens, and precisely controls the generation of multiple subjects in a single image. 🔹Precise Chinese and English Text Rendering with seamless image–text integration: The model naturally integrates text into images, making it suitable for a wide range of applications such as product covers, illustrations, and poster design to meet the needs of various fields. 🔹Rich Styles and High Aesthetic: Capable of generating images in various styles—including photorealistic portraits, comics, and vinyl figures—it delivers outstanding visual appeal and artistic quality. 🔹High-Quality Generation: Efficiently produces ultra-high-definition (2K) images in the same time other models take to generate a 1K image. HunyuanImage 2.1 uses two text encoders: a multimodal large language model (MLLM) to improve the model's image and text alignment capabilities, and a multi-language character-aware encoder to improve text rendering capabilities. The model is a single- and double-stream diffusion transformer with 17B parameters. We've also open-sourced the weights of the the accelerated version with meanflow which reduces inference steps from 100 to just 8, and PromptEnhancer, the first industrial-grade rewriting model that enhances your prompts for more nuanced and expressive image generation. Now, creators turn complex ideas—like posters with slogans or multi-panel comics—into visuals faster than ever. We’re just getting started. Stay tuned for our native multimodal image generation model coming soon. 🌐Website: 🔗Github: 🤗Hugging Face: ✨Hugging Face Demo:

Tencent Hy

89,430 görüntüleme • 1 yıl önce

🚀Exciting News! Tencent Hunyuan 3D AI Creation Engine v2.5 is here, revolutionizing 3D generation with ultra-high-definition modeling! 1⃣Massive Leap in Precision: Upgraded geometric model with 10B parameters (up from 1B) and 10x more mesh faces, delivering 1024-resolution geometric details—smoother surfaces, sharper edges, and richer visuals! 2⃣High-quality Textures & PBR: Supports high-quality textures and fine-grained bump mapping for lifelike surfaces. Industry-first multi-view PBR generation for realistic lighting and reflections. 3⃣Animation Boost: Optimized skeletal rigging with automatic bone binding and skin weight assignment, streamlining 3D animation workflows. 4⃣Flexible Workflows: New pipeline templates for text-to-3D and image-to-3D mesh reduction, plus multi-view 3D generation, empowering game development, animation, and more with customizable 3D assets. 5⃣Accessible Creation: Generate high-quality 3D models from text or images, slashing complexity. Free daily quota doubled to 20 for the AI Engine, with the API now live on Tencent Cloud! ☀️From games to AR/VR, Hunyuan 3D is redefining 3D AIGC, making creation faster, smarter, and more inclusive. Try it now! Tencent Hunyuan 3D AI Creation Engine: Tencent Hunyuan 3D Open Source:
1:30

🚀Exciting News! Tencent Hunyuan 3D AI Creation Engine v2.5 is here, revolutionizing 3D generation with ultra-high-definition modeling! 1⃣Massive Leap in Precision: Upgraded geometric model with 10B parameters (up from 1B) and 10x more mesh faces, delivering 1024-resolution geometric details—smoother surfaces, sharper edges, and richer visuals! 2⃣High-quality Textures & PBR: Supports high-quality textures and fine-grained bump mapping for lifelike surfaces. Industry-first multi-view PBR generation for realistic lighting and reflections. 3⃣Animation Boost: Optimized skeletal rigging with automatic bone binding and skin weight assignment, streamlining 3D animation workflows. 4⃣Flexible Workflows: New pipeline templates for text-to-3D and image-to-3D mesh reduction, plus multi-view 3D generation, empowering game development, animation, and more with customizable 3D assets. 5⃣Accessible Creation: Generate high-quality 3D models from text or images, slashing complexity. Free daily quota doubled to 20 for the AI Engine, with the API now live on Tencent Cloud! ☀️From games to AR/VR, Hunyuan 3D is redefining 3D AIGC, making creation faster, smarter, and more inclusive. Try it now! Tencent Hunyuan 3D AI Creation Engine: Tencent Hunyuan 3D Open Source:

Tencent Hy

58,225 görüntüleme • 1 yıl önce

Introducing Contrastive Language Model (CLM): an ultra-fast System One Model trained with a contrastive learning objective that connects states and actions. CLM-8B is pre-trained on internet-scale data and delivers up to 9× faster inference than Jev ⚡ while achieving comparable performance across computer-use, gaming, and tool-calling tasks. With lightweight fine-tuning, CLM-8B sets a new SOTA on challenging agentic coding benchmarks, such as DeepSWE (81.6%) and Terminal-Bench 2.1 (87.6%). In contrast, Jev fails to serve as an effective verifier for these long-horizon tasks. We also build an efficient training and serving infra for CLMs by disaggregating states and actions, allowing their embeddings to be cached and reused independently. This substantially reduces inference latency in settings where the state evolves continuously while the action set remains fixed. Finally, we establish scaling laws for CLMs and show that the test contrastive loss decreases predictably as a power law in training compute, model size, and dataset size. 📄 Blog: 💻 Code: 🗣️ Discord: 🤗 Data & Models: More details on CLM’s architecture, data recipe, and scaling laws in the thread below 🧵
1:20

Introducing Contrastive Language Model (CLM): an ultra-fast System One Model trained with a contrastive learning objective that connects states and actions. CLM-8B is pre-trained on internet-scale data and delivers up to 9× faster inference than Jev ⚡ while achieving comparable performance across computer-use, gaming, and tool-calling tasks. With lightweight fine-tuning, CLM-8B sets a new SOTA on challenging agentic coding benchmarks, such as DeepSWE (81.6%) and Terminal-Bench 2.1 (87.6%). In contrast, Jev fails to serve as an effective verifier for these long-horizon tasks. We also build an efficient training and serving infra for CLMs by disaggregating states and actions, allowing their embeddings to be cached and reused independently. This substantially reduces inference latency in settings where the state evolves continuously while the action set remains fixed. Finally, we establish scaling laws for CLMs and show that the test contrastive loss decreases predictably as a power law in training compute, model size, and dataset size. 📄 Blog: 💻 Code: 🗣️ Discord: 🤗 Data & Models: More details on CLM’s architecture, data recipe, and scaling laws in the thread below 🧵

Jacky Kwok

1,061,600 görüntüleme • 21 gün önce

New Course: Reinforcement Fine-Tuning LLMs with GRPO! Learn to use reinforcement learning to improve your LLM performance in this short course, built in collaboration with Predibase by Rubrik, and taught by Travis Addair, its Co-Founder and CTO, and Arnav Garg, its Senior Engineer and Machine Learning Lead. Reasoning models have been one of the most important developments in LLMs. Reinforcement Fine-Tuning (RFT) uses rewards to encourage LLMs to find solutions to multi-step reasoning tasks such as solving math problems and debugging code - without needing pre-existing training examples like in traditional supervised fine-tuning. Group Relative Policy Optimization (GRPO) is a reinforcement fine-tuning algorithm gaining rapid adoption. Developed by the DeepSeek team and used to train the R1 reasoning model, GRPO uses reward functions that you can write in Python to assign rewards to model responses. It’s beneficial for tasks with verifiable outcomes and can work well even with fewer than 100 training examples. It can also significantly improve the reasoning ability of smaller LLMs, making applications faster and more cost effective. In this course, you’ll take a technical deep dive into RFT with GRPO. You’ll learn to build reward functions that you can use in the GRPO training process to guide an LLM toward better performance on multi-step reasoning tasks. In detail, you’ll: - Learn when reinforcement fine-tuning is a better fit than supervised fine-tuning, especially for tasks involving multi-step reasoning or limited labeled data. - Understand how GRPO uses programmable reward functions as a more scalable alternative to the human feedback required for other reinforcement learning algorithms, such as RLHF and DPO. - Frame the Wordle game as a reinforcement fine-tuning problem and see how an LLM can learn to plan, analyze feedback, and improve its strategy over time. - Design reward functions that power the reinforcement fine-tuning process. - Learn techniques for evaluating more subjective tasks, such as rating the quality of a text summary, using an LLM as a judge. - Understand why reward hacking happens and how to avoid it by adding penalty functions to discourage undesirable behaviors. - Learn the four key components of the loss calculation in the GRPO algorithm: token probability distribution ratios, advantages, clipping, and KL-divergence. - Launch reinforcement fine-tuning jobs using Predibase’s hosted training services. By the end of this course, you’ll be able to build and fine-tune LLMs using reinforcement learning to improve reasoning without relying on large labeled datasets or subjective human feedback. Please sign up here:
2:41

New Course: Reinforcement Fine-Tuning LLMs with GRPO! Learn to use reinforcement learning to improve your LLM performance in this short course, built in collaboration with Predibase by Rubrik, and taught by Travis Addair, its Co-Founder and CTO, and Arnav Garg, its Senior Engineer and Machine Learning Lead. Reasoning models have been one of the most important developments in LLMs. Reinforcement Fine-Tuning (RFT) uses rewards to encourage LLMs to find solutions to multi-step reasoning tasks such as solving math problems and debugging code - without needing pre-existing training examples like in traditional supervised fine-tuning. Group Relative Policy Optimization (GRPO) is a reinforcement fine-tuning algorithm gaining rapid adoption. Developed by the DeepSeek team and used to train the R1 reasoning model, GRPO uses reward functions that you can write in Python to assign rewards to model responses. It’s beneficial for tasks with verifiable outcomes and can work well even with fewer than 100 training examples. It can also significantly improve the reasoning ability of smaller LLMs, making applications faster and more cost effective. In this course, you’ll take a technical deep dive into RFT with GRPO. You’ll learn to build reward functions that you can use in the GRPO training process to guide an LLM toward better performance on multi-step reasoning tasks. In detail, you’ll: - Learn when reinforcement fine-tuning is a better fit than supervised fine-tuning, especially for tasks involving multi-step reasoning or limited labeled data. - Understand how GRPO uses programmable reward functions as a more scalable alternative to the human feedback required for other reinforcement learning algorithms, such as RLHF and DPO. - Frame the Wordle game as a reinforcement fine-tuning problem and see how an LLM can learn to plan, analyze feedback, and improve its strategy over time. - Design reward functions that power the reinforcement fine-tuning process. - Learn techniques for evaluating more subjective tasks, such as rating the quality of a text summary, using an LLM as a judge. - Understand why reward hacking happens and how to avoid it by adding penalty functions to discourage undesirable behaviors. - Learn the four key components of the loss calculation in the GRPO algorithm: token probability distribution ratios, advantages, clipping, and KL-divergence. - Launch reinforcement fine-tuning jobs using Predibase’s hosted training services. By the end of this course, you’ll be able to build and fine-tune LLMs using reinforcement learning to improve reasoning without relying on large labeled datasets or subjective human feedback. Please sign up here:

Andrew Ng

86,697 görüntüleme • 1 yıl önce

We benchmarked leading multimodal foundation models (GPT-4o, Claude 3.5 Sonnet, Gemini, Llama, etc.) on standard computer vision tasks—from segmentation to surface normal estimation—using standard datasets like COCO and ImageNet. These models have made remarkable progress; however, it is unclear exactly where they stand in terms of understanding vision in detail. Especially when it comes to tasks beyond question-answering. How well do they understand an object's segments or geometry? Our analyses yield an assessment that is quantitatively and qualitatively detailed and is compatible with evaluations developed in the field of computer vision over the past decades. Observed trends: 🔹 The foundation models consistently underperform task-specific SOTA models across all tasks. However, they are respectable generalists, which is remarkable as they are presumably trained primarily on image-text-based tasks. 🔹 They perform semantic tasks notably better than geometric ones. 🔹 GPT-4o performs the best among non-reasoning models, getting the top position in 4 out of 6 tasks. 🔹 Reasoning models, e.g., o3, show improvements in geometric tasks. 🔹 The 'image generation' models, e.g., GPT-40 Image Generation, which have been natively trained multimodally, exhibit quirks. E.g., hallucinated objects, misalignment between the input and output, etc. 🔹 While the prompting techniques affect performance, better models exhibit less sensitivity to variations in prompts. We control for the variance introduced by the prompting methods in our experiments. 🌐 Detailed analyses, visualizations: ⌨️ code: 🧵 1/n
2:36

We benchmarked leading multimodal foundation models (GPT-4o, Claude 3.5 Sonnet, Gemini, Llama, etc.) on standard computer vision tasks—from segmentation to surface normal estimation—using standard datasets like COCO and ImageNet. These models have made remarkable progress; however, it is unclear exactly where they stand in terms of understanding vision in detail. Especially when it comes to tasks beyond question-answering. How well do they understand an object's segments or geometry? Our analyses yield an assessment that is quantitatively and qualitatively detailed and is compatible with evaluations developed in the field of computer vision over the past decades. Observed trends: 🔹 The foundation models consistently underperform task-specific SOTA models across all tasks. However, they are respectable generalists, which is remarkable as they are presumably trained primarily on image-text-based tasks. 🔹 They perform semantic tasks notably better than geometric ones. 🔹 GPT-4o performs the best among non-reasoning models, getting the top position in 4 out of 6 tasks. 🔹 Reasoning models, e.g., o3, show improvements in geometric tasks. 🔹 The 'image generation' models, e.g., GPT-40 Image Generation, which have been natively trained multimodally, exhibit quirks. E.g., hallucinated objects, misalignment between the input and output, etc. 🔹 While the prompting techniques affect performance, better models exhibit less sensitivity to variations in prompts. We control for the variance introduced by the prompting methods in our experiments. 🌐 Detailed analyses, visualizations: ⌨️ code: 🧵 1/n

Amir Zamir

73,435 görüntüleme • 1 yıl önce

🚨🚨Can agents earn money, run a business, or even self-organize a society in the physical social world? 🤖🤖 Can agents learn continually to survive and thrive in embodied environments, like how human babies grow? 👶 Super excited to introduce SimWorld, an open-ended simulator of LLM agents in infinite, realistic embodied worlds. SimWorld features 3 key designs: 1⃣Open-ended realistic world simulation - built on Unreal Engine 5, with accurate physical social dynamics - 100+ built-in environments (city, island, wilderness ...) - language-controllable procedural generation - text-to-3D asset generation 2⃣Native interface for LLM/VLM agents - Gym-like agent-environment interaction APIs - plug in any LLMs/VLMs (GPTs, Gemini, Qwen ...) - rich multi-modal perception - open-vocabulary natural-language action outputs 3⃣Diverse physical and social reasoning scenarios - long-horizon embodied reasoning - multi-agent collaboration / competition - easily customizable for any reasoning tasks SimWorld is fully open-sourced, with a hope to become a foundational infrastructure for real-world agent research across disciplines: robotics, economy, public health, education, etc. Project website + more details in the thread👇 ...1/
1:10

🚨🚨Can agents earn money, run a business, or even self-organize a society in the physical social world? 🤖🤖 Can agents learn continually to survive and thrive in embodied environments, like how human babies grow? 👶 Super excited to introduce SimWorld, an open-ended simulator of LLM agents in infinite, realistic embodied worlds. SimWorld features 3 key designs: 1⃣Open-ended realistic world simulation - built on Unreal Engine 5, with accurate physical social dynamics - 100+ built-in environments (city, island, wilderness ...) - language-controllable procedural generation - text-to-3D asset generation 2⃣Native interface for LLM/VLM agents - Gym-like agent-environment interaction APIs - plug in any LLMs/VLMs (GPTs, Gemini, Qwen ...) - rich multi-modal perception - open-vocabulary natural-language action outputs 3⃣Diverse physical and social reasoning scenarios - long-horizon embodied reasoning - multi-agent collaboration / competition - easily customizable for any reasoning tasks SimWorld is fully open-sourced, with a hope to become a foundational infrastructure for real-world agent research across disciplines: robotics, economy, public health, education, etc. Project website + more details in the thread👇 ...1/

Lianhui Qin@COLM2026

65,331 görüntüleme • 10 ay önce

StyleDrop: Text-to-Image Generation in Any Style introduce StyleDrop, a method that enables the synthesis of images that faithfully follow a specific style using a text-to-image model. The proposed method is extremely versatile and captures nuances and details of a user-provided style, such as color schemes, shading, design patterns, and local and global effects. It efficiently learns a new style by fine-tuning very few trainable parameters (less than 1% of total model parameters) and improving the quality via iterative training with either human or automated feedback. Better yet, StyleDrop is able to deliver impressive results even when the user supplies only a single image that specifies the desired style. An extensive study shows that, for the task of style tuning text-to-image models, StyleDrop implemented on Muse convincingly outperforms other methods, including DreamBooth and textual inversion on Imagen or Stable Diffusion. paper page:
1:56

StyleDrop: Text-to-Image Generation in Any Style introduce StyleDrop, a method that enables the synthesis of images that faithfully follow a specific style using a text-to-image model. The proposed method is extremely versatile and captures nuances and details of a user-provided style, such as color schemes, shading, design patterns, and local and global effects. It efficiently learns a new style by fine-tuning very few trainable parameters (less than 1% of total model parameters) and improving the quality via iterative training with either human or automated feedback. Better yet, StyleDrop is able to deliver impressive results even when the user supplies only a single image that specifies the desired style. An extensive study shows that, for the task of style tuning text-to-image models, StyleDrop implemented on Muse convincingly outperforms other methods, including DreamBooth and textual inversion on Imagen or Stable Diffusion. paper page:

AK

56,408 görüntüleme • 3 yıl önce

Anthropic's Sholto Douglas describes the concept of a software-only singularity: "It's one where the models are far better at digital tasks than they are at physical ones. And so we see rapid change in the digital world with models, and relatively little change in the physical world. So information and software changes dramatically, and this ends up having some pretty weird effects." "It means that maybe like the drivers of what have been the last couple decades of progress in the economy turn around." "And I think we'll see that flow on into the physical world but at a delay. So you get much better at doing chip design. You get much better at training AI models. AI models get a lot faster. Chips get a lot better. The general economy gets a lot more efficient because the sort of information and message parsing that is much of the rest of the economy ends up becoming much more efficient." "But at the same time you don't yet have robots providing limitless physical abundance. Science probably progresses really fast up to the degree that you need interaction with labs or larger particle colliders or something like this. And then you go, 'Okay well I need to build the robots.'"
3:13

Anthropic's Sholto Douglas describes the concept of a software-only singularity: "It's one where the models are far better at digital tasks than they are at physical ones. And so we see rapid change in the digital world with models, and relatively little change in the physical world. So information and software changes dramatically, and this ends up having some pretty weird effects." "It means that maybe like the drivers of what have been the last couple decades of progress in the economy turn around." "And I think we'll see that flow on into the physical world but at a delay. So you get much better at doing chip design. You get much better at training AI models. AI models get a lot faster. Chips get a lot better. The general economy gets a lot more efficient because the sort of information and message parsing that is much of the rest of the economy ends up becoming much more efficient." "But at the same time you don't yet have robots providing limitless physical abundance. Science probably progresses really fast up to the degree that you need interaction with labs or larger particle colliders or something like this. And then you go, 'Okay well I need to build the robots.'"

TBPN

208,090 görüntüleme • 8 ay önce

For the first time in human history, we are teaching a Foundation Model to master the diverse tasks of medicinal chemists, biologists, and computational scientists all in one place. In our latest collaboration with Liquid AI, we are moving away from fragmented, specialized tools toward a single, super-intelligent model. What surprised me most? This model isn't just performing at reasonable levels—it has started outperforming specialist models across physics-based tasks, imaging, and longitudinal data. Why this changes everything: -Synergy over Specialization: Fine-tuning on specific tasks has unlocked unexpected capabilities in synergetic areas, opening a new frontier in multimodal AI research. -Zero-Shot Potential: We are building a model that can perform out-of-scope tasks, moving us closer to an "AI deity" for drug discovery. -Quality First: The goal isn't just to bypass regulations to save time; it’s about using these synergies to develop better, more effective drugs. We are no longer just looking at linear regression or simple text; we are looking at the future of how humanity fights disease. #LiquidAI #InsilicoMedicine #GenerativeAI #DrugDiscovery #DeepTech #BiotechInnovation
1:53

For the first time in human history, we are teaching a Foundation Model to master the diverse tasks of medicinal chemists, biologists, and computational scientists all in one place. In our latest collaboration with Liquid AI, we are moving away from fragmented, specialized tools toward a single, super-intelligent model. What surprised me most? This model isn't just performing at reasonable levels—it has started outperforming specialist models across physics-based tasks, imaging, and longitudinal data. Why this changes everything: -Synergy over Specialization: Fine-tuning on specific tasks has unlocked unexpected capabilities in synergetic areas, opening a new frontier in multimodal AI research. -Zero-Shot Potential: We are building a model that can perform out-of-scope tasks, moving us closer to an "AI deity" for drug discovery. -Quality First: The goal isn't just to bypass regulations to save time; it’s about using these synergies to develop better, more effective drugs. We are no longer just looking at linear regression or simple text; we are looking at the future of how humanity fights disease. #LiquidAI #InsilicoMedicine #GenerativeAI #DrugDiscovery #DeepTech #BiotechInnovation

Alex Zhavoronkov, PhD (aka Aleksandrs Zavoronkovs)

10,544 görüntüleme • 6 ay önce

I’ve been testing Hy4 preview in WorkBuddy, and the most interesting part is not simply the model size, it’s how much practical work it can handle with a relatively focused active parameter count. Hy4 preview brings together stronger code understanding, generation, and editing; improved document and information processing; workflow automation; web and game development; cross-tool collaboration; and more reliable completion of complex, multi-step tasks. In other words, it is designed for work that requires planning, tool use, iteration, and follow-through, not just a quick answer in a chat window. Compared with its initial release, the current Hy4 preview is noticeably faster and better-performing in practical workflows. Following an upgrade released yesterday, it can complete tasks with fewer conversation rounds and lower token usage, while reasoning more quickly and making the overall user experience feel smoother from the first instruction to the final result. For my test, I gave it a demanding Three.js game-prototyping task with a 770B-parameter model and 49B active parameters. The result was more revealing than a simple first-look demo: Hy4 preview handled the core logic, edge cases, and follow-up changes while maintaining the broader context of the project. That combination of capability, speed, context, and active compute is what makes its cost-effectiveness worth examining. A fair evaluation should use the same prompt and environment configuration across models, changing only the model itself. That makes it easier to assess task completion, planning quality, tool-calling stability, reasoning speed, token efficiency, and performance over longer workflows without confusing the result with different settings. If you want to test the model yourself, access Hy4 preview through WorkBuddy and see how it performs on a real coding, document, automation, or creative task: Tencent Hy Tencent AI WorkBuddy
1:02

I’ve been testing Hy4 preview in WorkBuddy, and the most interesting part is not simply the model size, it’s how much practical work it can handle with a relatively focused active parameter count. Hy4 preview brings together stronger code understanding, generation, and editing; improved document and information processing; workflow automation; web and game development; cross-tool collaboration; and more reliable completion of complex, multi-step tasks. In other words, it is designed for work that requires planning, tool use, iteration, and follow-through, not just a quick answer in a chat window. Compared with its initial release, the current Hy4 preview is noticeably faster and better-performing in practical workflows. Following an upgrade released yesterday, it can complete tasks with fewer conversation rounds and lower token usage, while reasoning more quickly and making the overall user experience feel smoother from the first instruction to the final result. For my test, I gave it a demanding Three.js game-prototyping task with a 770B-parameter model and 49B active parameters. The result was more revealing than a simple first-look demo: Hy4 preview handled the core logic, edge cases, and follow-up changes while maintaining the broader context of the project. That combination of capability, speed, context, and active compute is what makes its cost-effectiveness worth examining. A fair evaluation should use the same prompt and environment configuration across models, changing only the model itself. That makes it easier to assess task completion, planning quality, tool-calling stability, reasoning speed, token efficiency, and performance over longer workflows without confusing the result with different settings. If you want to test the model yourself, access Hy4 preview through WorkBuddy and see how it performs on a real coding, document, automation, or creative task: Tencent Hy Tencent AI WorkBuddy

Tyler Wayne

56,152 görüntüleme • 1 ay önce

Reinforcement Learning from Human Feedback (RLHF) is gaining traction. This field aims to make AI more responsible by including human values and preferences. In this video, Nathan Lambert, a research scientist and RLHF team lead at Hugging Face explores its inner workings, applications and industry impact. RLHF has gained the spotlight in recent years. The growth of language models like Anthropic’s Claude and OpenAI's ChatGPT have increased interest in human-feedback integration. "There are some rumors that Open AI had two teams; one was doing RLHF and the other instruction fine-tuning. And the RLHF team kept getting more and more performance." Understanding RLHF The RLHF process has three main steps: Pre-training: Much like with GPT models, the journey starts with pre-training on a large corpus of data. This can range from text data, web scrapes, to specialized datasets. Reward Modeling: This is the RLHF counterpart of supervised fine-tuning in large language models. This stage involves creating a reward model that resonates with human values and preferences. RL Optimization: This stage parallels reward modeling and reinforcement learning in traditional AI models. The AI system fine-tunes itself based on the reward model, employing reinforcement learning algorithms for that extra layer of optimization. The Data Challenge Data collection and curation in RLHF closely resemble the challenges you'd encounter in large language model training. Datasets from organizations like OpenAI can serve as a useful foundation. However, the need for high-quality, task-specific data cannot be overstated. Implementing RLHF: A Practical Guide If you’re someone who loves getting hands-on with AI libraries like Hugging Face, implementing RLHF is right way to do. It’s essential to understand its limitations. Think about model stability, over-optimization, and exploration strategies, much like you would when prompt engineering. Ongoing Research and Next Steps While he suggests that some basics figured out, there are layers of complexity that still need to be unraveled: 1. New Benchmarks: How do we measure the effectiveness of RLHF? 2. Preference Modeling: How can the model be made to understand human preferences better? 3. Interpreting RLHF: Much like explainability in traditional models, how do we make RLHF more interpretable? 4. System-Wide Evaluation: Going beyond individual performance, how does RLHF affect an entire system? The Transformative Power of RLHF Whether you're an AI developer, a business analyst, or a marketer, RLHF promises to revolutionize your domain. Imagine customer service chatbots that understand human emotions better, or content generators that align more closely with human values. RLHF is an emerging field that focuses on enhancing machine learning models through human feedback. While it tackles important issues like bias and ethics, its broader goal is to improve system performance across various applications. Whether you're deeply invested in the ethics of AI or simply curious about advancements in machine learning, RLHF offers valuable insights. If you're interested in the next wave of AI development, this area is definitely one to watch.
2:00

Reinforcement Learning from Human Feedback (RLHF) is gaining traction. This field aims to make AI more responsible by including human values and preferences. In this video, Nathan Lambert, a research scientist and RLHF team lead at Hugging Face explores its inner workings, applications and industry impact. RLHF has gained the spotlight in recent years. The growth of language models like Anthropic’s Claude and OpenAI's ChatGPT have increased interest in human-feedback integration. "There are some rumors that Open AI had two teams; one was doing RLHF and the other instruction fine-tuning. And the RLHF team kept getting more and more performance." Understanding RLHF The RLHF process has three main steps: Pre-training: Much like with GPT models, the journey starts with pre-training on a large corpus of data. This can range from text data, web scrapes, to specialized datasets. Reward Modeling: This is the RLHF counterpart of supervised fine-tuning in large language models. This stage involves creating a reward model that resonates with human values and preferences. RL Optimization: This stage parallels reward modeling and reinforcement learning in traditional AI models. The AI system fine-tunes itself based on the reward model, employing reinforcement learning algorithms for that extra layer of optimization. The Data Challenge Data collection and curation in RLHF closely resemble the challenges you'd encounter in large language model training. Datasets from organizations like OpenAI can serve as a useful foundation. However, the need for high-quality, task-specific data cannot be overstated. Implementing RLHF: A Practical Guide If you’re someone who loves getting hands-on with AI libraries like Hugging Face, implementing RLHF is right way to do. It’s essential to understand its limitations. Think about model stability, over-optimization, and exploration strategies, much like you would when prompt engineering. Ongoing Research and Next Steps While he suggests that some basics figured out, there are layers of complexity that still need to be unraveled: 1. New Benchmarks: How do we measure the effectiveness of RLHF? 2. Preference Modeling: How can the model be made to understand human preferences better? 3. Interpreting RLHF: Much like explainability in traditional models, how do we make RLHF more interpretable? 4. System-Wide Evaluation: Going beyond individual performance, how does RLHF affect an entire system? The Transformative Power of RLHF Whether you're an AI developer, a business analyst, or a marketer, RLHF promises to revolutionize your domain. Imagine customer service chatbots that understand human emotions better, or content generators that align more closely with human values. RLHF is an emerging field that focuses on enhancing machine learning models through human feedback. While it tackles important issues like bias and ethics, its broader goal is to improve system performance across various applications. Whether you're deeply invested in the ethics of AI or simply curious about advancements in machine learning, RLHF offers valuable insights. If you're interested in the next wave of AI development, this area is definitely one to watch.

Muratcan Koylan

27,379 görüntüleme • 3 yıl önce

Cerebras inference is very fast. So fast that it changes how we think about configuring our LLMs for voice agent use cases. Kimi K2.6 is a 1T parameter reasoning model that Cerebras serves at 650 - 1,000 tokens per second (end-to-end throughput), with time to first token metrics as low as 150ms (latency). These numbers are two to three times faster than other similarly capable models. The biggest lever we get from this kind of speed is that we can use the model in reasoning mode, and still have excellent "time to first non-thinking token." This solves a big pain point we have in 2026 for voice agent use cases. Almost all recent innovation in post-training has focused on making models good at reasoning ("test time compute"). This is great, but it makes the user-facing model latency much, much slower. Which is a problem for conversational voice agents. We can run Kimi K2.6 with reasoning turned on, and get responses faster than other models produce with reasoning disabled. On my 30-turn voice agent benchmark, Kimi K2.6 with reasoning enabled ties GPT 5.1 and Haiku 4.5 with reasoning disabled, and is still about 200ms seconds faster! On my primary task agent benchmark, Kimi K2.6 is now the #2 model. It ranks just behind Gemini 3.5 Flash in "high" reasoning mode, and tied with GLM 5, Sonnet 4.6, and GPT 5.4 with reasoning set to "low." But Kimi K2.6 completes each turn in the agent loop in under 500ms. The other four models are all at least 3x slower. (Models only qualify for this benchmark if they can complete task turns at a P50 <4s.) A couple of other things that this speed buys us, for production voice agents: - Tool calls happen fast enough that we don't have to work around tool call latency in our pipeline design. - We can prompt the model to output structured data at the beginning of a response, followed by plain text for voice generation. This opens up possibilities like asking the model to do complex classification/generation tasks that influence the rest of the pipeline. For example, the model could create a detailed style prompt for a steerable TTS model, for each individual conversation turn. And, of course, you can use Kimi K2.6 with reasoning turned off. Cerebras calls this "instant" mode. Here's a video of a Cerebras Kimi K2.6 voice agent with voice-to-voice response time, measured at the client, under 500ms. This is the true response latency as perceived by the user, including all network and audio codec overhead, transcription and turn detection, Kimi K2.6 token generation, and voice generation. 500ms is, effectively, instant. So the Cerebras naming for this mode is a propos. :-)
0:40

Cerebras inference is very fast. So fast that it changes how we think about configuring our LLMs for voice agent use cases. Kimi K2.6 is a 1T parameter reasoning model that Cerebras serves at 650 - 1,000 tokens per second (end-to-end throughput), with time to first token metrics as low as 150ms (latency). These numbers are two to three times faster than other similarly capable models. The biggest lever we get from this kind of speed is that we can use the model in reasoning mode, and still have excellent "time to first non-thinking token." This solves a big pain point we have in 2026 for voice agent use cases. Almost all recent innovation in post-training has focused on making models good at reasoning ("test time compute"). This is great, but it makes the user-facing model latency much, much slower. Which is a problem for conversational voice agents. We can run Kimi K2.6 with reasoning turned on, and get responses faster than other models produce with reasoning disabled. On my 30-turn voice agent benchmark, Kimi K2.6 with reasoning enabled ties GPT 5.1 and Haiku 4.5 with reasoning disabled, and is still about 200ms seconds faster! On my primary task agent benchmark, Kimi K2.6 is now the #2 model. It ranks just behind Gemini 3.5 Flash in "high" reasoning mode, and tied with GLM 5, Sonnet 4.6, and GPT 5.4 with reasoning set to "low." But Kimi K2.6 completes each turn in the agent loop in under 500ms. The other four models are all at least 3x slower. (Models only qualify for this benchmark if they can complete task turns at a P50 <4s.) A couple of other things that this speed buys us, for production voice agents: - Tool calls happen fast enough that we don't have to work around tool call latency in our pipeline design. - We can prompt the model to output structured data at the beginning of a response, followed by plain text for voice generation. This opens up possibilities like asking the model to do complex classification/generation tasks that influence the rest of the pipeline. For example, the model could create a detailed style prompt for a steerable TTS model, for each individual conversation turn. And, of course, you can use Kimi K2.6 with reasoning turned off. Cerebras calls this "instant" mode. Here's a video of a Cerebras Kimi K2.6 voice agent with voice-to-voice response time, measured at the client, under 500ms. This is the true response latency as perceived by the user, including all network and audio codec overhead, transcription and turn detection, Kimi K2.6 token generation, and voice generation. 500ms is, effectively, instant. So the Cerebras naming for this mode is a propos. :-)

kwindla

40,593 görüntüleme • 4 ay önce

A team tested Pi0, Pi0 Fast, Gr00t, and ACT on real robot arms in manufacturing tasks. (🔖 Bookmark this for later!) The task was precise: place thin rectangular frames from a messy stack into a holder. The team fine-tuned each model on 100 real trajectories and compared training time, inference speed, motion quality, and success rates. ⬇️ Here’s a breakdown of what they found Pi0 (Original) ✅ Strongest overall performance in precise pick-and-place ✅ High success rate even in edge cases ✅ Longest training time (~11 hours, ~$30 per run) ✅ Inference time of 80 ms causes short pauses between actions Despite delays, it handles complex scenarios well… solid for high-precision tasks, but slow to train. Gr00t ✅ Trains fast (~2 hours, ~$5 per run) ✅ Performs almost as well as Pi0 on large-object tasks ✅ Struggles with fine precision; random movement in some trials ✅ More training didn’t fix jitter or random offsets Best suited for tasks where exact precision isn’t critical. Not ready for manufacturing-grade accuracy without more tuning. Pi0 Fast ✅ Promised faster training, but results were underwhelming ✅ Training at 6 hours still showed low success rates ✅ Inference was slower than expected ✅ Not reliable for generalizing even slightly new tasks Currently too unstable for real-world deployment. Doesn’t live up to the “Fast” name yet. ACT (Baseline) ✅ 200MB model—lightweight, but limited ✅ Struggles with stacked objects or ambiguous scenes ✅ Success rates around 70% in best-case setups ✅ Can’t match newer models on precision or generalization Still a solid baseline, but clearly a generation behind in robustness. 🚨 Extra Notes All newer models share a common issue: •Inference takes longer than a frame (80 ms vs 33 ms), so robots “pause” between chunks. •This results in jittery movements, but not a dealbreaker unless tasks are time-sensitive. Language-conditioned tasks also fell short: after training on two labeled tasks, the model couldn’t generalize to a third unseen combination using only text prompts. ✅ The good news? These models adapt well to new robot arms with quick fine-tuning. ❌ The bad news? There’s still no plug-and-play solution for improving performance after deployment. Reinforcement learning or DAgger-style data collection during real-world operation may be the next big step, something many teams in robotics are actively working on.
0:51

A team tested Pi0, Pi0 Fast, Gr00t, and ACT on real robot arms in manufacturing tasks. (🔖 Bookmark this for later!) The task was precise: place thin rectangular frames from a messy stack into a holder. The team fine-tuned each model on 100 real trajectories and compared training time, inference speed, motion quality, and success rates. ⬇️ Here’s a breakdown of what they found Pi0 (Original) ✅ Strongest overall performance in precise pick-and-place ✅ High success rate even in edge cases ✅ Longest training time (~11 hours, ~$30 per run) ✅ Inference time of 80 ms causes short pauses between actions Despite delays, it handles complex scenarios well… solid for high-precision tasks, but slow to train. Gr00t ✅ Trains fast (~2 hours, ~$5 per run) ✅ Performs almost as well as Pi0 on large-object tasks ✅ Struggles with fine precision; random movement in some trials ✅ More training didn’t fix jitter or random offsets Best suited for tasks where exact precision isn’t critical. Not ready for manufacturing-grade accuracy without more tuning. Pi0 Fast ✅ Promised faster training, but results were underwhelming ✅ Training at 6 hours still showed low success rates ✅ Inference was slower than expected ✅ Not reliable for generalizing even slightly new tasks Currently too unstable for real-world deployment. Doesn’t live up to the “Fast” name yet. ACT (Baseline) ✅ 200MB model—lightweight, but limited ✅ Struggles with stacked objects or ambiguous scenes ✅ Success rates around 70% in best-case setups ✅ Can’t match newer models on precision or generalization Still a solid baseline, but clearly a generation behind in robustness. 🚨 Extra Notes All newer models share a common issue: •Inference takes longer than a frame (80 ms vs 33 ms), so robots “pause” between chunks. •This results in jittery movements, but not a dealbreaker unless tasks are time-sensitive. Language-conditioned tasks also fell short: after training on two labeled tasks, the model couldn’t generalize to a third unseen combination using only text prompts. ✅ The good news? These models adapt well to new robot arms with quick fine-tuning. ❌ The bad news? There’s still no plug-and-play solution for improving performance after deployment. Reinforcement learning or DAgger-style data collection during real-world operation may be the next big step, something many teams in robotics are actively working on.

Ilir Aliu

21,844 görüntüleme • 1 yıl önce