正在加载视频...

视频加载失败

加载此视频时出现问题。这可能是由于临时网络问题，或视频可能不可用。

💥 Introducing MiniCPM-o 2.6: An 8B size, GPT-4o level Omni Model runs on device ✨ Highlights: ~Match GPT-4o-202405 in vision, audio and multimodal live streaming ~End-to-end real-time bilingual audio conversation ~Voice cloning & emotion control ~Advanced OCR & video understanding ~Offline iPad-compatible multimodal live streaming 🔗 Try it out:... show more

OpenBMB

8,365 subscribers

97,663 次观看 • 1 年前 •via X (Twitter)

教育科学技术

Anya Rossi• Live Now

Private livecam show

11 条评论

merve 的头像

merve1 年前

very hyped about this 🥹🤩 congratulations! if you feel like hosting your demo at HF Spaces we can provide GPU support ☺️

AssemblyAI 的头像

AssemblyAI1 年前

Announcing: Our most advanced speech-to-text model goes beyond accuracy to capture the real-world complexity of human conversation and deliver reliable, source-of-truth audio data. Explore Universal-2 updates 👇

Zhiyuan Liu 的头像

Zhiyuan Liu1 年前

gpt-4o level on-device model, a new step of densing law. 💥💥💥

Al BigBang 的头像

Al BigBang1 年前

This is truly Open AI,not closed AI. @elonmusk @sama @OpenAI @xai

PrimerYang 的头像

PrimerYang1 年前

Can the delay be reduced to below 200ms? If possible, it will be a very noteworthy model!

Lynncc 的头像

Lynncc1 年前

They're extremely cracked，what to expect next?

GDP (not GPT) 的头像

GDP (not GPT)1 年前

Why does it say I was developed by OpenAI???😀

Benjamin Shehu 的头像

Benjamin Shehu1 年前

@teortaxesTex bmb giving out strong anime protagonist energy

AK 的头像

AK1 年前

Her, on device👁️👂🏻👄

Nicolas Maton 的头像

Nicolas Maton1 年前

Is the ipad code available somewhere? Love to tinker with this.

AJ - e/Acc 🚀🇨🇴🇵🇷 的头像

AJ - e/Acc 🚀🇨🇴🇵🇷1 年前

Looks pretty cool! Got a 2nd brain in my iPad now? 😏 Cant wait to see how it handles real-time audio and vision. Thanks for sharing the links! Gonna check it out and see what I can mess around with. 👀

相关视频

🚀Introducing MiniCPM-V 2.6! 🔥 1、Surpassing GPT-4V in single image, multi-image and video understanding 📸🎥 2、Outperforms GPT-4o mini and Gemini 1.5 on OpenCompass 🏆 3、Real-time video analysis on iPad 📱💨 Try out the best on-device multimodal LLM here！ 👑 GitHub： Huggingface： #MLLM #MiniCPM

🚀Introducing MiniCPM-V 2.6! 🔥 1、Surpassing GPT-4V in single image, multi-image and video understanding 📸🎥 2、Outperforms GPT-4o mini and Gemini 1.5 on OpenCompass 🏆 3、Real-time video analysis on iPad 📱💨 Try out the best on-device multimodal LLM here！ 👑 GitHub： Huggingface： #MLLM #MiniCPM

OpenBMB

196,343 次观看 • 1 年前

🚀 Introducing MiniCPM-V 4.5 8B: pushing the boundary of multimodal AI! ～ SOTA VL Capability: Surpasses GPT-4o, Gemini 2.0 Pro, Qwen2.5-VL 72B on OpenCompass! ～ "Eagle Eye" Video: 96x visual token compression for high refresh rate and long video understanding ～ Controllable Hybrid Fast/Deep Thinking ～ Strong OCR & Doc Parsing: Surpasses GPT-4o & Gemini 2.5 on OmniDocBench Get ready for the future of multimodal AI 👉 Huggingface｜ Github｜ Gradio｜ #AI #MiniCM #GPT #Gemini #OpenBMB #ArtificialIntelligence #MachineLearning

🚀 Introducing MiniCPM-V 4.5 8B: pushing the boundary of multimodal AI! ～ SOTA VL Capability: Surpasses GPT-4o, Gemini 2.0 Pro, Qwen2.5-VL 72B on OpenCompass! ～ "Eagle Eye" Video: 96x visual token compression for high refresh rate and long video understanding ～ Controllable Hybrid Fast/Deep Thinking ～ Strong OCR & Doc Parsing: Surpasses GPT-4o & Gemini 2.5 on OmniDocBench Get ready for the future of multimodal AI 👉 Huggingface｜ Github｜ Gradio｜ #AI #MiniCM #GPT #Gemini #OpenBMB #ArtificialIntelligence #MachineLearning

OpenBMB

25,136 次观看 • 10 个月前

🥳 Introducing MiniCPM-o 4.5 The first full-duplex omni-modal LLM in open-source community 🎬🎙️ 🔥 Key Highlights: • Full-duplex Omni-modal Live Streaming: The model can see, listen, and speak simultaneously in a real-time conversation without mutual blocking • Proactive Interaction: Moving beyond reactive QA to performing proactive interaction, such as initiating reminders • Leading Performance: Scoring 77.6 on OpenCompass, it outperforms GPT-4o & Gemini 2.0 Pro in vision-language tasks with 9B params The best part? You can experience all above on your PC! #MiniCPM #OpenSource #MultimodalAI #LLM

🥳 Introducing MiniCPM-o 4.5 The first full-duplex omni-modal LLM in open-source community 🎬🎙️ 🔥 Key Highlights: • Full-duplex Omni-modal Live Streaming: The model can see, listen, and speak simultaneously in a real-time conversation without mutual blocking • Proactive Interaction: Moving beyond reactive QA to performing proactive interaction, such as initiating reminders • Leading Performance: Scoring 77.6 on OpenCompass, it outperforms GPT-4o & Gemini 2.0 Pro in vision-language tasks with 9B params The best part? You can experience all above on your PC! #MiniCPM #OpenSource #MultimodalAI #LLM

OpenBMB

397,772 次观看 • 5 个月前

New open Omni model released! 👀OpenBMB MiniCPM-o 2.6 is a new 8B parameters, any-to-any multimodal model that can understand vision, speech, and language and runs on edge devices like phones and tablets. TL;DR: 🧠 8B total parameters (SigLip-400M + Whisper-300M + ChatTTS-200M + Qwen2.5-7B) 🔥 Outperforms GPT-4V on visual tasks with 70.2 average score on OpenCompass 🎙️ Best-in-class bilingual speech capabilities with real-time conversation and voice cloning 🎬 Supports multimodal streaming with support for continuous video/audio processing 📱 Runs on iPads and phones and supports 30+ languages 🖼️ Processes images up to 1.8M pixels (1344x1344) with OCR capabilities 🛠️ Easy integration with popular frameworks (llama.cpp, vLLM, Gradio) 🤗 commercial friendly (< 1 mio DAU) and available on Hugging Face

New open Omni model released! 👀OpenBMB MiniCPM-o 2.6 is a new 8B parameters, any-to-any multimodal model that can understand vision, speech, and language and runs on edge devices like phones and tablets. TL;DR: 🧠 8B total parameters (SigLip-400M + Whisper-300M + ChatTTS-200M + Qwen2.5-7B) 🔥 Outperforms GPT-4V on visual tasks with 70.2 average score on OpenCompass 🎙️ Best-in-class bilingual speech capabilities with real-time conversation and voice cloning 🎬 Supports multimodal streaming with support for continuous video/audio processing 📱 Runs on iPads and phones and supports 30+ languages 🖼️ Processes images up to 1.8M pixels (1344x1344) with OCR capabilities 🛠️ Easy integration with popular frameworks (llama.cpp, vLLM, Gradio) 🤗 commercial friendly (< 1 mio DAU) and available on Hugging Face

Philipp Schmid

19,409 次观看 • 1 年前

MiniCPM-o 4.5: Seeing, Listening, and Speaking — All at Once. 👁️👂🗣️ ✨Beyond traditional turn-taking, we’ve built a Native Full-Duplex engine that allows a 9B model to see, listen, and speak in one concurrent, non-blocking stream. Watch how it masters real-world complexity in real-time: 🔔 Proactive Auditory Interaction: Interrupts itself to alert you when it hears a "Ding!" while reading cards. 🎨 Temporal Flow Tracking: Follows your pen in real-time, narrating and "mind-reading" your drawing as you sketch. 🍎 Omni-Perception: Scans groceries & identifies prices on the fly. ✨Why it’s a category-leader: 📌Performance: Surpasses GPT-4o and Gemini 2.0 Pro on OpenCompass (Avg. 77.6). 📌Architecture: End-to-end fusion of SigLip2, Whisper, and CosyVoice2 on a Qwen3-8B base. 📌Efficiency: Full-duplex live streaming now runs locally on PCs via llama.cpp-omni. The era of "Wait-and-Response" AI is over. Proactive, real-time intelligence is now open-source. 🚀Experience it on Hugging Face: 🔗 #MiniCPM #Omnimodal #FullDuplex #EdgeAI #OpenSource #ComputerVision

MiniCPM-o 4.5: Seeing, Listening, and Speaking — All at Once. 👁️👂🗣️ ✨Beyond traditional turn-taking, we’ve built a Native Full-Duplex engine that allows a 9B model to see, listen, and speak in one concurrent, non-blocking stream. Watch how it masters real-world complexity in real-time: 🔔 Proactive Auditory Interaction: Interrupts itself to alert you when it hears a "Ding!" while reading cards. 🎨 Temporal Flow Tracking: Follows your pen in real-time, narrating and "mind-reading" your drawing as you sketch. 🍎 Omni-Perception: Scans groceries & identifies prices on the fly. ✨Why it’s a category-leader: 📌Performance: Surpasses GPT-4o and Gemini 2.0 Pro on OpenCompass (Avg. 77.6). 📌Architecture: End-to-end fusion of SigLip2, Whisper, and CosyVoice2 on a Qwen3-8B base. 📌Efficiency: Full-duplex live streaming now runs locally on PCs via llama.cpp-omni. The era of "Wait-and-Response" AI is over. Proactive, real-time intelligence is now open-source. 🚀Experience it on Hugging Face: 🔗 #MiniCPM #Omnimodal #FullDuplex #EdgeAI #OpenSource #ComputerVision

OpenBMB

115,435 次观看 • 5 个月前

VITA Towards Open-Source Interactive Omni Multimodal LLM discuss: The remarkable multimodal capabilities and interactive experience of GPT-4o underscore their necessity in practical applications, yet open-source models rarely excel in both areas. In this paper, we introduce VITA, the first-ever open-source Multimodal Large Language Model (MLLM) adept at simultaneous processing and analysis of Video, Image, Text, and Audio modalities, and meanwhile has an advanced multimodal interactive experience. Starting from Mixtral 8x7B as a language foundation, we expand its Chinese vocabulary followed by bilingual instruction tuning. We further endow the language model with visual and audio capabilities through two-stage multi-task learning of multimodal alignment and instruction tuning. VITA demonstrates robust foundational capabilities of multilingual, vision, and audio understanding, as evidenced by its strong performance across a range of both unimodal and multimodal benchmarks. Beyond foundational capabilities, we have made considerable progress in enhancing the natural multimodal human-computer interaction experience. To the best of our knowledge, we are the first to exploit non-awakening interaction and audio interrupt in MLLM. VITA is the first step for the open-source community to explore the seamless integration of multimodal understanding and interaction. While there is still lots of work to be done on VITA to get close to close-source counterparts, we hope that its role as a pioneer can serve as a cornerstone for subsequent research.

VITA Towards Open-Source Interactive Omni Multimodal LLM discuss: The remarkable multimodal capabilities and interactive experience of GPT-4o underscore their necessity in practical applications, yet open-source models rarely excel in both areas. In this paper, we introduce VITA, the first-ever open-source Multimodal Large Language Model (MLLM) adept at simultaneous processing and analysis of Video, Image, Text, and Audio modalities, and meanwhile has an advanced multimodal interactive experience. Starting from Mixtral 8x7B as a language foundation, we expand its Chinese vocabulary followed by bilingual instruction tuning. We further endow the language model with visual and audio capabilities through two-stage multi-task learning of multimodal alignment and instruction tuning. VITA demonstrates robust foundational capabilities of multilingual, vision, and audio understanding, as evidenced by its strong performance across a range of both unimodal and multimodal benchmarks. Beyond foundational capabilities, we have made considerable progress in enhancing the natural multimodal human-computer interaction experience. To the best of our knowledge, we are the first to exploit non-awakening interaction and audio interrupt in MLLM. VITA is the first step for the open-source community to explore the seamless integration of multimodal understanding and interaction. While there is still lots of work to be done on VITA to get close to close-source counterparts, we hope that its role as a pioneer can serve as a cornerstone for subsequent research.

AK

23,958 次观看 • 1 年前

🚀 🚀Excited to announce the technical report of MiniCPM-o 4.5! MiniCPM-o 4.5 transitions #AI interaction from traditional turn-based processing to a real-time, native full-duplex stream-based paradigm. 🌊 The Omni-Flow Framework Instead of traditional VAD-based workarounds, we introduce the #Omni-#Flow framework. This unified stream paradigm aligns video, audio, and text on a synchronized millisecond timeline. • Native Full-Duplex: Simultaneous perception and response. • Proactive Interaction: Natively manages turn-taking without external VAD, supports proactive reminding. 📉 9B Scale, SOTA Performance MiniCPM-o 4.5 demonstrates SOTA multimodal intelligence at its scale: • Multimodal Benchmarks: Comparable to #Gemini 2.5 Flash on MMBench EN (87.6) and MathVista (80.1). • Streaming Evaluation: 54.4% win rate on LiveSports-3K-CC, surpassing specialized models. 💻 The Ultimate Edge AI — Fully Functional without Network Connection We are providing one-click installers for Windows (12G VRAM,RTX 5070) and macOS (M1-M5 Max/ M5 Pro). • Local API Support: Deploy your own inference server to integrate native full-duplex into custom apps. • Free Access: We are offering free community API services for exploration. • 100% Private: Your data never leaves your machine. Deploy in under 10 minutes. 🛠️👇 👐 Join the Open Future The weights are open. The protocol is public. 📄 Technical Report: 💻 GitHub: 🤗 HuggingFace: 🌐 Web Demo: #MiniCPMo #OpenSourceAI #EdgeAI #MachineLearning #ComputerVision #LLM

🚀 🚀Excited to announce the technical report of MiniCPM-o 4.5! MiniCPM-o 4.5 transitions #AI interaction from traditional turn-based processing to a real-time, native full-duplex stream-based paradigm. 🌊 The Omni-Flow Framework Instead of traditional VAD-based workarounds, we introduce the #Omni-#Flow framework. This unified stream paradigm aligns video, audio, and text on a synchronized millisecond timeline. • Native Full-Duplex: Simultaneous perception and response. • Proactive Interaction: Natively manages turn-taking without external VAD, supports proactive reminding. 📉 9B Scale, SOTA Performance MiniCPM-o 4.5 demonstrates SOTA multimodal intelligence at its scale: • Multimodal Benchmarks: Comparable to #Gemini 2.5 Flash on MMBench EN (87.6) and MathVista (80.1). • Streaming Evaluation: 54.4% win rate on LiveSports-3K-CC, surpassing specialized models. 💻 The Ultimate Edge AI — Fully Functional without Network Connection We are providing one-click installers for Windows (12G VRAM,RTX 5070) and macOS (M1-M5 Max/ M5 Pro). • Local API Support: Deploy your own inference server to integrate native full-duplex into custom apps. • Free Access: We are offering free community API services for exploration. • 100% Private: Your data never leaves your machine. Deploy in under 10 minutes. 🛠️👇 👐 Join the Open Future The weights are open. The protocol is public. 📄 Technical Report: 💻 GitHub: 🤗 HuggingFace: 🌐 Web Demo: #MiniCPMo #OpenSourceAI #EdgeAI #MachineLearning #ComputerVision #LLM

OpenBMB

147,824 次观看 • 2 个月前

1/5 MiniCPM-V 4.6 (1.3B) is now live 🚀🚀 High-res visual processing, optimized for consumer-grade and mobile hardware. We’ve leveraged the latest LLaVA-UHD v4 technique to cut vision encoding costs by 55%, enabling native edge deployment with extreme efficiency. 🔥 Beats Gemma4-E2B-it and Qwen3.5-0.8B across key multimodal and Artificial Analysis benchmarks — scoring higher than Qwen3.5-0.8B using just 2.5% of its token budget. ⚡ TTFT (75.7ms) 2.2x Faster than Qwen3.5-0.8B even with 3136² high-res images. 🏗️ ~1.5x Token Throughput compared with Qwen3.5-0.8B on a single RTX 4090. Try the model here: 🤗 Hugging Face: 💻 GitHub: 🔭 Modelscope: 🌐 Web Demo: 📱 App Demo:

1/5 MiniCPM-V 4.6 (1.3B) is now live 🚀🚀 High-res visual processing, optimized for consumer-grade and mobile hardware. We’ve leveraged the latest LLaVA-UHD v4 technique to cut vision encoding costs by 55%, enabling native edge deployment with extreme efficiency. 🔥 Beats Gemma4-E2B-it and Qwen3.5-0.8B across key multimodal and Artificial Analysis benchmarks — scoring higher than Qwen3.5-0.8B using just 2.5% of its token budget. ⚡ TTFT (75.7ms) 2.2x Faster than Qwen3.5-0.8B even with 3136² high-res images. 🏗️ ~1.5x Token Throughput compared with Qwen3.5-0.8B on a single RTX 4090. Try the model here: 🤗 Hugging Face: 💻 GitHub: 🔭 Modelscope: 🌐 Web Demo: 📱 App Demo:

OpenBMB

352,293 次观看 • 2 个月前

🚀 VoxCPM 2 is live! 🎉 Another open-source AI #TTS model from China — and one that stands shoulder to shoulder with Qwen3-TTS, while bringing everything into a single unified model. After rapid iterations from V1 (zero-shot cloning) to V1.5 (long-form + fine-tuning), #VoxCPM has consistently pushed quality and usability forward. Now, VoxCPM 2 takes it further: 🔹30+ languages — truly global, truly local. 🔹Infinite voice design — type it, hear it, control it. From a whisper to a booming cinematic voice. 🔹Studio-grade audio — 48kHz ultra-high fidelity with emotional depth 🔹Diffusion-Autoregressive cloning — preserves more acoustic and emotional detail than token-based models like Qwen3-TTS 💡 Big shoutout to Grok — used your multi-image video magic for our launch demo. It’s scarily good at keeping visuals consistent across shots. Elon Elon Musk, this one’s for you. 😉 Check the demo & start cloning your dream voice: 🌐 Hugging Face Space: 🤗 Hugging Face Model: 🤖 ModelScope Model: 💻 GitHub： #TTS #AI #VoiceCloning #GrokImagine #ElonMusk #OpenBMB #VoxCPM

🚀 VoxCPM 2 is live! 🎉 Another open-source AI #TTS model from China — and one that stands shoulder to shoulder with Qwen3-TTS, while bringing everything into a single unified model. After rapid iterations from V1 (zero-shot cloning) to V1.5 (long-form + fine-tuning), #VoxCPM has consistently pushed quality and usability forward. Now, VoxCPM 2 takes it further: 🔹30+ languages — truly global, truly local. 🔹Infinite voice design — type it, hear it, control it. From a whisper to a booming cinematic voice. 🔹Studio-grade audio — 48kHz ultra-high fidelity with emotional depth 🔹Diffusion-Autoregressive cloning — preserves more acoustic and emotional detail than token-based models like Qwen3-TTS 💡 Big shoutout to Grok — used your multi-image video magic for our launch demo. It’s scarily good at keeping visuals consistent across shots. Elon Elon Musk, this one’s for you. 😉 Check the demo & start cloning your dream voice: 🌐 Hugging Face Space: 🤗 Hugging Face Model: 🤖 ModelScope Model: 💻 GitHub： #TTS #AI #VoiceCloning #GrokImagine #ElonMusk #OpenBMB #VoxCPM

OpenBMB

557,835 次观看 • 3 个月前

Introducing Lightning V3 - it beats every model we tested against. ElevenLabs, Cartesia, OpenAI. Lightning sets a new SOTA with V3 in conversational text-to-speech. → Highest MOS score for conversational TTS at 3.9 → ~76% win rate vs gpt-4o-mini-tts on naturalness → 15 languages with mid-sentence code-switching → Built from scratch for voice agents, not read-aloud Every TTS model sounds clean in a demo. You type a sentence and you get beautiful audio. Voice agents don't work that way. They stream. They're generating audio in real-time chunks with half the context missing. That's where everything breaks. A great reading voice and a great conversational voice are fundamentally different things. A conversational voice has to sound like it's thinking - with the pauses, the rhythm shifts, the reactions. It has to handle the way real people actually talk, including switching languages mid-sentence. That's what V3 does. V3.1 also ships voice cloning. 5 to 15 seconds of audio, no fine-tuning, production-grade clone across 15 languages. Blog link in the comments.

Introducing Lightning V3 - it beats every model we tested against. ElevenLabs, Cartesia, OpenAI. Lightning sets a new SOTA with V3 in conversational text-to-speech. → Highest MOS score for conversational TTS at 3.9 → ~76% win rate vs gpt-4o-mini-tts on naturalness → 15 languages with mid-sentence code-switching → Built from scratch for voice agents, not read-aloud Every TTS model sounds clean in a demo. You type a sentence and you get beautiful audio. Voice agents don't work that way. They stream. They're generating audio in real-time chunks with half the context missing. That's where everything breaks. A great reading voice and a great conversational voice are fundamentally different things. A conversational voice has to sound like it's thinking - with the pauses, the rhythm shifts, the reactions. It has to handle the way real people actually talk, including switching languages mid-sentence. That's what V3 does. V3.1 also ships voice cloning. 5 to 15 seconds of audio, no fine-tuning, production-grade clone across 15 languages. Blog link in the comments.

Sudarshan Kamath

71,298 次观看 • 3 个月前

🤔The world’s best small models? We immediately compared Mistral-3-8B with our previous-gen model, MiniCPM-4.1 (Both in thinking) 😂The findings are compelling: ✅MiniCPM is still ～2x faster, maintaining a massive speed lead ✅It remains a full generation ahead in capabilities (excluding math/code) For developers prioritizing efficiency and speed, MiniCPM is undeniably the world's best small model.

🤔The world’s best small models? We immediately compared Mistral-3-8B with our previous-gen model, MiniCPM-4.1 (Both in thinking) 😂The findings are compelling: ✅MiniCPM is still ～2x faster, maintaining a massive speed lead ✅It remains a full generation ahead in capabilities (excluding math/code) For developers prioritizing efficiency and speed, MiniCPM is undeniably the world's best small model.

OpenBMB

260,271 次观看 • 7 个月前

Introducing MiniCPM 4.1-8B: First Open-Source Reasoning LLM with Trainable Sparse Attention ✅ Strong Reasoning Capability: Surpasses similar-sized models on 15 tasks! ✅ Fast Generation: 3x decoding speedup for reasoning ✅ Efficient Architecture: Trainable sparse attention, frequency-ranked speculative decoding Download Models: Huggingface: Github: Technical Report: #AI #MiniCPM #LLM #OpenBMB #ArtificialIntelligence #MachineLearning

Introducing MiniCPM 4.1-8B: First Open-Source Reasoning LLM with Trainable Sparse Attention ✅ Strong Reasoning Capability: Surpasses similar-sized models on 15 tasks! ✅ Fast Generation: 3x decoding speedup for reasoning ✅ Efficient Architecture: Trainable sparse attention, frequency-ranked speculative decoding Download Models: Huggingface: Github: Technical Report: #AI #MiniCPM #LLM #OpenBMB #ArtificialIntelligence #MachineLearning

OpenBMB

19,236 次观看 • 10 个月前

Why does a realistic voice matter? 🤔🤔 The same robot that feels creepy can transform into a trusted companion just by having a human-like voice. Think about the movie, Her. 🔥VoxCPM is a new paradigm: continuous, context-aware, and incredibly lifelike. ✅ Small size: 0.5B only ✅ Zero-Shot Voice Cloning ✅ Context-Aware, Expressive Speech Generation Try it 👉 🔗Huggingface | 🔗Github |

Why does a realistic voice matter? 🤔🤔 The same robot that feels creepy can transform into a trusted companion just by having a human-like voice. Think about the movie, Her. 🔥VoxCPM is a new paradigm: continuous, context-aware, and incredibly lifelike. ✅ Small size: 0.5B only ✅ Zero-Shot Voice Cloning ✅ Context-Aware, Expressive Speech Generation Try it 👉 🔗Huggingface | 🔗Github |

OpenBMB

126,770 次观看 • 10 个月前

Google I/O. Some thoughts: the model seems to be multimodal in, but not multimodal out. Imagen-3 and music gen models are still detached from Gemini as standalone components. Merging all modality I/O natively is the inevitable future: - enables tasks like "use a more robotic voice", "speak 2x faster", "edit this image iteratively", and "generate consistent comic strips". - does not lose information across modal boundaries, e.g. emotion and background sound. - opens up new in-context capabilities. You can teach the model to combine different senses in novel ways with few-shot examples. GPT-4o doesn't do it perfectly, but it gets the form factor correct. In words of Andrej's LLM-as-OS analogy: we need the model to natively support as many file extensions as possible. One thing Google is doing right: they are finally making serious efforts to integrate AI into the search box. I sense the agent flow: planning, real-time browsing, and multimodal input, all from the landing page. Google's strongest moat is distribution. Gemini doesn't have to be the best model to be the most used one in the world.

Google I/O. Some thoughts: the model seems to be multimodal in, but not multimodal out. Imagen-3 and music gen models are still detached from Gemini as standalone components. Merging all modality I/O natively is the inevitable future: - enables tasks like "use a more robotic voice", "speak 2x faster", "edit this image iteratively", and "generate consistent comic strips". - does not lose information across modal boundaries, e.g. emotion and background sound. - opens up new in-context capabilities. You can teach the model to combine different senses in novel ways with few-shot examples. GPT-4o doesn't do it perfectly, but it gets the form factor correct. In words of Andrej's LLM-as-OS analogy: we need the model to natively support as many file extensions as possible. One thing Google is doing right: they are finally making serious efforts to integrate AI into the search box. I sense the agent flow: planning, real-time browsing, and multimodal input, all from the landing page. Google's strongest moat is distribution. Gemini doesn't have to be the best model to be the most used one in the world.

Jim Fan

209,643 次观看 • 2 年前

🚀 Introducing AgentCPM-Explore: The First Open-Source 4B-Agent Model to Conquer GAIA & Complex Real-World Tasks! 🤗 Hugging Face: 🔗 GitHub: ✨ Key Highlights: ✅ SOTA Agentic Performance: Sets a new benchmark for 4B-scale agent models—outperforming all peers, surpassing 8B models, and rivaling select 30B+ and closed-source LLMs. 🧠 Deep Research Capability: Excels at long-horizon reasoning, supports 100+ turns of autonomous interaction with multi-source cross-validation, human-like self-correction, and dynamic tool use + strategy adaptation—just like a real researcher! 🔓 Full-Stack Open Source: We’re open-sourcing the entire end-to-end agent stack—not just the model! Empower your own innovations with - AgentRL: Asynchronous reinforcement learning framework - AgentDock: Secure, extensible tool sandbox - AgentToLeaP: An one-click evaluation platform for agent tool-learning capabilitie - Full training data pipeline & reproducible workflows #AgentCPM #OpenSourceAI #AgenticAI #AI #GAIA #LLM #OpenBMB #AIAgents #HuggingFace

🚀 Introducing AgentCPM-Explore: The First Open-Source 4B-Agent Model to Conquer GAIA & Complex Real-World Tasks! 🤗 Hugging Face: 🔗 GitHub: ✨ Key Highlights: ✅ SOTA Agentic Performance: Sets a new benchmark for 4B-scale agent models—outperforming all peers, surpassing 8B models, and rivaling select 30B+ and closed-source LLMs. 🧠 Deep Research Capability: Excels at long-horizon reasoning, supports 100+ turns of autonomous interaction with multi-source cross-validation, human-like self-correction, and dynamic tool use + strategy adaptation—just like a real researcher! 🔓 Full-Stack Open Source: We’re open-sourcing the entire end-to-end agent stack—not just the model! Empower your own innovations with - AgentRL: Asynchronous reinforcement learning framework - AgentDock: Secure, extensible tool sandbox - AgentToLeaP: An one-click evaluation platform for agent tool-learning capabilitie - Full training data pipeline & reproducible workflows #AgentCPM #OpenSourceAI #AgenticAI #AI #GAIA #LLM #OpenBMB #AIAgents #HuggingFace

OpenBMB

13,996 次观看 • 6 个月前

I know your timeline is flooded now with word salads of "insane, HER, 10 features you missed, we're so back". Sit down. Chill. Take a deep breath like Mark does in the demo . Let's think step by step: - Technique-wise, OpenAI has figured out a way to map audio to audio directly as first-class modality, and stream videos to a transformer in real-time. These require some new research on tokenization and architecture, but overall it's a data and system optimization problem (as most things are). High-quality data can come from at least 2 sources: 1) Naturally occurring dialogues on YouTube, podcasts, TV series, movies, etc. Whisper can be trained to identify speaker turns in a dialogue or separate overlapping speeches for automated annotation. 2) Synthetic data. Run the slow 3-stage pipeline using the most powerful models: speech1->text1 (ASR), text1->text2 (LLM), text2->speech2 (TTS). The middle LLM can decide when to stop and also simulate how to resume from interruption. It could output additional "thought traces" that are not verbalized to help generate better reply. Then GPT-4o distills directly from speech1->speech2, with optional auxiliary loss functions based on the 3-stage data. After distillation, these behaviors are now baked into the model without emitting intermediate texts. On the system side: the latency would not meet real-time threshold if every video frame is decompressed into an RGB image. OpenAI has likely developed their own neural-first, streaming video codec to transmit the motion deltas as tokens. The communication protocol and NN inference must be co-optimized. For example, there could be a small and energy-efficient NN running on the edge device that decides to transmit more tokens if the video is interesting, and fewer otherwise. - I didn't expect GPT-4o to be closer to GPT-5, the rumored "Arrakis" model that takes multimodal in and out. In fact, it's likely an early checkpoint of GPT-5 that hasn't finished training yet. The branding betrays a certain insecurity. Ahead of Google I/O, OpenAI would rather beat our mental projection of GPT-4.5 than disappoint by missing the sky-high expectation for GPT-5. A smart move to buy more time. - Notably, the assistant is much more lively and even a bit flirty. GPT-4o is trying (perhaps a bit too hard) to sound like HER. OpenAI is eating Character AI's lunch, with almost 100% overlap in form factor and huge distribution channels. It's a pivot towards more emotional AI with strong personality, which OpenAI seemed to actively suppress in the past. - Whoever wins Apple first wins big time. I see 3 levels of integration with iOS: 1) Ditch Siri. OpenAI distills a smaller-tier, purely on-device GPT-4o for iOS, with optional paid upgrade to use the cloud. 2) Native features to stream the camera or screen into the model. Chip-level support for neural audio/video codec. 3) Integrate with iOS system-level action API and smart home APIs. No one uses Siri Shortcuts, but it's time to resurrect. This could become the AI agent product with a billion users from the get-go. The FSD for smartphones with a Tesla-scale data flywheel.

I know your timeline is flooded now with word salads of "insane, HER, 10 features you missed, we're so back". Sit down. Chill. Take a deep breath like Mark does in the demo . Let's think step by step: - Technique-wise, OpenAI has figured out a way to map audio to audio directly as first-class modality, and stream videos to a transformer in real-time. These require some new research on tokenization and architecture, but overall it's a data and system optimization problem (as most things are). High-quality data can come from at least 2 sources: 1) Naturally occurring dialogues on YouTube, podcasts, TV series, movies, etc. Whisper can be trained to identify speaker turns in a dialogue or separate overlapping speeches for automated annotation. 2) Synthetic data. Run the slow 3-stage pipeline using the most powerful models: speech1->text1 (ASR), text1->text2 (LLM), text2->speech2 (TTS). The middle LLM can decide when to stop and also simulate how to resume from interruption. It could output additional "thought traces" that are not verbalized to help generate better reply. Then GPT-4o distills directly from speech1->speech2, with optional auxiliary loss functions based on the 3-stage data. After distillation, these behaviors are now baked into the model without emitting intermediate texts. On the system side: the latency would not meet real-time threshold if every video frame is decompressed into an RGB image. OpenAI has likely developed their own neural-first, streaming video codec to transmit the motion deltas as tokens. The communication protocol and NN inference must be co-optimized. For example, there could be a small and energy-efficient NN running on the edge device that decides to transmit more tokens if the video is interesting, and fewer otherwise. - I didn't expect GPT-4o to be closer to GPT-5, the rumored "Arrakis" model that takes multimodal in and out. In fact, it's likely an early checkpoint of GPT-5 that hasn't finished training yet. The branding betrays a certain insecurity. Ahead of Google I/O, OpenAI would rather beat our mental projection of GPT-4.5 than disappoint by missing the sky-high expectation for GPT-5. A smart move to buy more time. - Notably, the assistant is much more lively and even a bit flirty. GPT-4o is trying (perhaps a bit too hard) to sound like HER. OpenAI is eating Character AI's lunch, with almost 100% overlap in form factor and huge distribution channels. It's a pivot towards more emotional AI with strong personality, which OpenAI seemed to actively suppress in the past. - Whoever wins Apple first wins big time. I see 3 levels of integration with iOS: 1) Ditch Siri. OpenAI distills a smaller-tier, purely on-device GPT-4o for iOS, with optional paid upgrade to use the cloud. 2) Native features to stream the camera or screen into the model. Chip-level support for neural audio/video codec. 3) Integrate with iOS system-level action API and smart home APIs. No one uses Siri Shortcuts, but it's time to resurrect. This could become the AI agent product with a billion users from the get-go. The FSD for smartphones with a Tesla-scale data flywheel.

Jim Fan

991,628 次观看 • 2 年前

The teams shipping AI agents right now are bleeding money on the dumbest possible expense: teaching a 400B-parameter model to read a file name. Every time an AI agent needs to "see" something today, it routes an image through a frontier model. OCR, object detection, checking if a button exists on screen. You're paying GPT-4o or Claude pricing for tasks that require perception, not reasoning. One agent workflow processing a few thousand screenshots per day can burn through more on vision calls than on the actual thinking. Perceptron's Isaac is 2B parameters. Built by the team that created Meta's Chameleon multimodal models. On perceptive benchmarks, it matches or beats models 50x its size. The VQA, OCR, and object detection scores are competitive with models running on infrastructure that costs orders of magnitude more. The MCP wrapper is the distribution play. One install command and every Claude Code agent can offload vision tasks to a model that runs on a single consumer GPU. The agent keeps its reasoning in the frontier model and routes perception to a specialist. That split is how you get vision-heavy agent workflows from "technically possible but expensive" to "cheap enough to run on everything." This is the same pattern that won in every other compute-intensive stack. General-purpose handles orchestration. Specialists handle the heavy lifting. Graphics went through it. Audio went through it. Video encoding went through it. Vision in AI agents is next. The teams building agents that see 10,000 images a day will care about this before anyone else does.

The teams shipping AI agents right now are bleeding money on the dumbest possible expense: teaching a 400B-parameter model to read a file name. Every time an AI agent needs to "see" something today, it routes an image through a frontier model. OCR, object detection, checking if a button exists on screen. You're paying GPT-4o or Claude pricing for tasks that require perception, not reasoning. One agent workflow processing a few thousand screenshots per day can burn through more on vision calls than on the actual thinking. Perceptron's Isaac is 2B parameters. Built by the team that created Meta's Chameleon multimodal models. On perceptive benchmarks, it matches or beats models 50x its size. The VQA, OCR, and object detection scores are competitive with models running on infrastructure that costs orders of magnitude more. The MCP wrapper is the distribution play. One install command and every Claude Code agent can offload vision tasks to a model that runs on a single consumer GPU. The agent keeps its reasoning in the frontier model and routes perception to a specialist. That split is how you get vision-heavy agent workflows from "technically possible but expensive" to "cheap enough to run on everything." This is the same pattern that won in every other compute-intensive stack. General-purpose handles orchestration. Specialists handle the heavy lifting. Graphics went through it. Audio went through it. Video encoding went through it. Vision in AI agents is next. The teams building agents that see 10,000 images a day will care about this before anyone else does.

Aakash Gupta

55,978 次观看 • 3 个月前

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

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

73,074 次观看 • 1 年前

Text-to-speech is moving way too fast. Just a few days ago, I tweeted about PersonaPlex-7B, NVIDIA's new open source TTS ( And today, Qwen just open-sourced Qwen3-TTS 🤯 It’s a revolutionary text-to-speech model built for control. Not just about generating speech, but about shaping how it sounds directly from language. You can guide the pace, the tone, and the expressiveness straight from text, without touching audio graphs or hand-tuning parameters. That’s the real shift! What makes Qwen3-TTS stand out is how practical it already is: → voice cloning from just a few seconds of audio → voice creation without any reference sample → support for 10 languages out of the box → end-to-end latency down to ~97ms → works in both streaming and non-streaming setups The models come in two sizes (0.6B and 1.7B), so you can trade off quality and hardware cost depending on your setup. You can work with curated voices, designed voices, or cloned ones, and it integrates cleanly with vLLM for production use. It also ships as a simple Python package you can pip install. If you’re building real-time voice systems, this removes a lot of friction! 100% free and open source. I put the repo in the 🧵↓

Text-to-speech is moving way too fast. Just a few days ago, I tweeted about PersonaPlex-7B, NVIDIA's new open source TTS ( And today, Qwen just open-sourced Qwen3-TTS 🤯 It’s a revolutionary text-to-speech model built for control. Not just about generating speech, but about shaping how it sounds directly from language. You can guide the pace, the tone, and the expressiveness straight from text, without touching audio graphs or hand-tuning parameters. That’s the real shift! What makes Qwen3-TTS stand out is how practical it already is: → voice cloning from just a few seconds of audio → voice creation without any reference sample → support for 10 languages out of the box → end-to-end latency down to ~97ms → works in both streaming and non-streaming setups The models come in two sizes (0.6B and 1.7B), so you can trade off quality and hardware cost depending on your setup. You can work with curated voices, designed voices, or cloned ones, and it integrates cleanly with vLLM for production use. It also ships as a simple Python package you can pip install. If you’re building real-time voice systems, this removes a lot of friction! 100% free and open source. I put the repo in the 🧵↓

Charly Wargnier

59,144 次观看 • 5 个月前

QVAC SDK 0.14.0 is live. This release makes the on-device stack faster on mobile, ships the developer-agent path, and takes local text-to-speech to 31 languages. Main highlights: - OpenCode and OpenClaw. The first official OpenCode plugin, plus a maintained OpenClaw compatibility path, both built on managed mode and qvac serve. Point a coding agent at a local model with far less setup and far fewer surprises. - Brain-computer interface transcription, on the SDK. Take recorded neural signal data and decode it into text, fully on-device, no cloud. Stream it in chunks through a simple API. In 0.14 it runs GPU-accelerated on iOS. - Text to Speech in 31 languages with our Supertonic3 upgrade. VOICE AND SPEECH - Supertonic3 multilingual TTS, 5 languages to 31. - Chatterbox and Supertonic now run on the Android GPU, with lower memory use (especially on iOS), quantized s3gen Chatterbox support, and a fix for Chatterbox occasionally emitting random speech. - Whisper transcription now runs on the iOS GPU. Parakeet runs on the Android GPU, with steadier real-time streaming. VISION AND OCR - VLM multi-tile batching: high-resolution Pan and Scan images are encoded in one pass instead of tile by tile, for faster vision throughput. - OCR on ggml (EasyOCR and DocTR) reaches full speed parity with the onnx path, across Metal, OpenCL, and Vulkan. PLATFORM AND RELIABILITY - Dynamic compute backends on Linux: one build picks the right backend at runtime, and opens the door to ROCm and CUDA support without per-backend builds. - Thinking tokens are kept out of the model context, so reasoning no longer fills the KV cache. SDK 0.14.0 is now leaner and faster to start. Let’s build.

QVAC SDK 0.14.0 is live. This release makes the on-device stack faster on mobile, ships the developer-agent path, and takes local text-to-speech to 31 languages. Main highlights: - OpenCode and OpenClaw. The first official OpenCode plugin, plus a maintained OpenClaw compatibility path, both built on managed mode and qvac serve. Point a coding agent at a local model with far less setup and far fewer surprises. - Brain-computer interface transcription, on the SDK. Take recorded neural signal data and decode it into text, fully on-device, no cloud. Stream it in chunks through a simple API. In 0.14 it runs GPU-accelerated on iOS. - Text to Speech in 31 languages with our Supertonic3 upgrade. VOICE AND SPEECH - Supertonic3 multilingual TTS, 5 languages to 31. - Chatterbox and Supertonic now run on the Android GPU, with lower memory use (especially on iOS), quantized s3gen Chatterbox support, and a fix for Chatterbox occasionally emitting random speech. - Whisper transcription now runs on the iOS GPU. Parakeet runs on the Android GPU, with steadier real-time streaming. VISION AND OCR - VLM multi-tile batching: high-resolution Pan and Scan images are encoded in one pass instead of tile by tile, for faster vision throughput. - OCR on ggml (EasyOCR and DocTR) reaches full speed parity with the onnx path, across Metal, OpenCL, and Vulkan. PLATFORM AND RELIABILITY - Dynamic compute backends on Linux: one build picks the right backend at runtime, and opens the door to ROCm and CUDA support without per-backend builds. - Thinking tokens are kept out of the model context, so reasoning no longer fills the KV cache. SDK 0.14.0 is now leaner and faster to start. Let’s build.

QVAC

23,973,950 次观看 • 18 天前