mHC: Manifold-Constrained Hyper-Connections
Paper
•
2512.24880
•
Published
•
125
AI & ML interests
None defined yet.
Recent Activity
View all activity
AdinaY
submitted
a
paper
to Daily Papers
2 days ago
Post
658
Following up on LLaDA 2.0 , the paper is now out on Daily Papers🔥
It has sparked a lot of discussion in the community for showing how discrete diffusion LLMs can scale to 100B and run faster than traditional AR models.
LLaDA2.0: Scaling Up Diffusion Language Models to 100B (2512.15745)
It has sparked a lot of discussion in the community for showing how discrete diffusion LLMs can scale to 100B and run faster than traditional AR models.
LLaDA2.0: Scaling Up Diffusion Language Models to 100B (2512.15745)
Post
4556
Finch 💰 an enterprise-grade benchmark that measures whether AI agents can truly handle real world finance & accounting work.
FinWorkBench/Finch
✨ Built from real enterprise data (Enron + financial institutions), not synthetic tasks
✨ Tests end-to-end finance workflows
✨ Multimodal & cross-file reasoning
✨ Expert annotated (700+ hours) and genuinely challenging hard
FinWorkBench/Finch
✨ Built from real enterprise data (Enron + financial institutions), not synthetic tasks
✨ Tests end-to-end finance workflows
✨ Multimodal & cross-file reasoning
✨ Expert annotated (700+ hours) and genuinely challenging hard
AdinaY
authored
a
paper
18 days ago
Post
3367
Kimi K2 Thinking is now live on the hub 🔥
moonshotai/Kimi-K2-Thinking
✨ 1T MoE for deep reasoning & tool use
✨ Native INT4 quantization = 2× faster inference
✨ 256K context window
✨ Modified MIT license
moonshotai/Kimi-K2-Thinking
✨ 1T MoE for deep reasoning & tool use
✨ Native INT4 quantization = 2× faster inference
✨ 256K context window
✨ Modified MIT license
Post
722
Chinese open source AI in October wasn’t about bigger models, it was about real world impact 🔥
https://huggingface.co/collections/zh-ai-community/october-2025-china-open-source-highlights
✨ Vision-Language & OCR wave 🌊
- DeepSeek-OCR : 3B
- PaddleOCR-VL : 0.9B
- Qwen3-VL : 2B / 4B / 8B / 32B /30B-A3B
- Open-Bee: Bee-8B-RL
- http://Z.ai Glyph :10B
OCR is industrializing, the real game now is understanding the (long context) document, not just reading it.
✨ Text generation: scale or innovation?
- MiniMax-M2: 229B
- Antgroup Ling-1T & Ring-1T
- Moonshot Kimi-Linear : linear-attention challenger
- Kwaipilot KAT-Dev
Efficiency is the key.
✨ Any-to-Any & World-Model : one step forward to the real world
- BAAI Emu 3.5
- Antgroup Ming-flash-omni
- HunyuanWorld-Mirror: 3D
Aligning with the “world model” globally
✨ Audio & Speech + Video & Visual: released from entertainment labs to delivery platforms
- SoulX-Podcast TTS
- LongCat-Audio-Codec & LongCat-Video by Meituan delivery paltform
- xiabs DreamOmni 2
Looking forward to what's next 🚀
https://huggingface.co/collections/zh-ai-community/october-2025-china-open-source-highlights
✨ Vision-Language & OCR wave 🌊
- DeepSeek-OCR : 3B
- PaddleOCR-VL : 0.9B
- Qwen3-VL : 2B / 4B / 8B / 32B /30B-A3B
- Open-Bee: Bee-8B-RL
- http://Z.ai Glyph :10B
OCR is industrializing, the real game now is understanding the (long context) document, not just reading it.
✨ Text generation: scale or innovation?
- MiniMax-M2: 229B
- Antgroup Ling-1T & Ring-1T
- Moonshot Kimi-Linear : linear-attention challenger
- Kwaipilot KAT-Dev
Efficiency is the key.
✨ Any-to-Any & World-Model : one step forward to the real world
- BAAI Emu 3.5
- Antgroup Ming-flash-omni
- HunyuanWorld-Mirror: 3D
Aligning with the “world model” globally
✨ Audio & Speech + Video & Visual: released from entertainment labs to delivery platforms
- SoulX-Podcast TTS
- LongCat-Audio-Codec & LongCat-Video by Meituan delivery paltform
- xiabs DreamOmni 2
Looking forward to what's next 🚀
AdinaY
authored
a
paper
about 2 months ago
Post
568
Kimi Linear🚀 Hybrid linear attention model from Moonshot AI
https://huggingface.co/collections/moonshotai/kimi-linear-a3b
✨ 48B total/ 3B active - MIT license
✨ Up to 1M context
✨ 84.3 on RULER (128k) with 3.98× speedup
✨ Hybrid KDA + MLA architecture for peak throughput & quality
https://huggingface.co/collections/moonshotai/kimi-linear-a3b
✨ 48B total/ 3B active - MIT license
✨ Up to 1M context
✨ 84.3 on RULER (128k) with 3.98× speedup
✨ Hybrid KDA + MLA architecture for peak throughput & quality
Post
3861
🤖 Did you know your voice might be cloned without your consent from just *one sentence* of audio?
That's not great. So with @frimelle , we brainstormed a new idea for developers who want to curb malicious use: ✨The Voice Consent Gate.✨
Details, code, here: https://huggingface.co/blog/voice-consent-gate
That's not great. So with @frimelle , we brainstormed a new idea for developers who want to curb malicious use: ✨The Voice Consent Gate.✨
Details, code, here: https://huggingface.co/blog/voice-consent-gate
- 3 replies
Post
1782
Ming-flash-omni Preview 🚀 Multimodal foundation model from AntGroup
inclusionAI/Ming-flash-omni-Preview
✨ Built on Ling-Flash-2.0: 10B total/6B active
✨ Generative segmentation-as-editing
✨ SOTA contextual & dialect ASR
✨ High-fidelity image generation
inclusionAI/Ming-flash-omni-Preview
✨ Built on Ling-Flash-2.0: 10B total/6B active
✨ Generative segmentation-as-editing
✨ SOTA contextual & dialect ASR
✨ High-fidelity image generation
Post
1886
Glyph 🔥 a framework that scales context length by compressing text into images and processing them with vision–language models, released by Z.ai.
Paper:https://huggingface.co/papers/2510.17800
Model:https://huggingface.co/zai-org/Glyph
✨ Compresses long sequences visually to bypass token limits
✨ Reduces computational and memory costs
✨ Preserves meaning through multimodal encoding
✨ Built on GLM-4.1V-9B-Base
Post
2671
HunyuanWorld Mirror🔥a versatile feed forward model for universal 3D world reconstruction by Tencent
tencent/HunyuanWorld-Mirror
✨ Any prior in → 3D world out
✨ Mix camera, intrinsics, depth as priors
✨ Predict point clouds, normals, Gaussians & more in one pass
✨ Unified architecture for all 3D task
tencent/HunyuanWorld-Mirror
✨ Any prior in → 3D world out
✨ Mix camera, intrinsics, depth as priors
✨ Predict point clouds, normals, Gaussians & more in one pass
✨ Unified architecture for all 3D task
Post
698
PaddleOCR VL🔥 0.9B Multilingual VLM by Baidu
PaddlePaddle/PaddleOCR-VL
✨ Ultra-efficient NaViT + ERNIE-4.5 architecture
✨ Supports 109 languages 🤯
✨ Accurately recognizes text, tables, formulas & charts
✨ Fast inference and lightweight for deployment
PaddlePaddle/PaddleOCR-VL
✨ Ultra-efficient NaViT + ERNIE-4.5 architecture
✨ Supports 109 languages 🤯
✨ Accurately recognizes text, tables, formulas & charts
✨ Fast inference and lightweight for deployment
Post
1826
Bee-8B 🐝 open 8B Multimodal LLM built on high quality data, released by
TencentHunyuan
Paper: Bee: A High-Quality Corpus and Full-Stack Suite to Unlock Advanced Fully Open MLLMs (2510.13795)
Model: https://huggingface.co/collections/Open-Bee/bee-8b-68ecbf10417810d90fbd9995
✨ Trained on Honey-Data-15M, a 15M-sample SFT corpus with dual-level CoT reasoning
✨ Backed by HoneyPipe, a transparent & reproducible open data curation suite
TencentHunyuan
Paper: Bee: A High-Quality Corpus and Full-Stack Suite to Unlock Advanced Fully Open MLLMs (2510.13795)
Model: https://huggingface.co/collections/Open-Bee/bee-8b-68ecbf10417810d90fbd9995
✨ Trained on Honey-Data-15M, a 15M-sample SFT corpus with dual-level CoT reasoning
✨ Backed by HoneyPipe, a transparent & reproducible open data curation suite
Post
349
Reflection ≠ self-correction
Interesting paper on long-chain reasoning 📑 from Miromind AI.
First Try Matters: Revisiting the Role of Reflection in Reasoning Models (2510.08308)
It dives into how LLMs think 🧠 Most reflections confirm, not fix, and true improvement comes from stronger initial reasoning.
Interesting paper on long-chain reasoning 📑 from Miromind AI.
First Try Matters: Revisiting the Role of Reflection in Reasoning Models (2510.08308)
It dives into how LLMs think 🧠 Most reflections confirm, not fix, and true improvement comes from stronger initial reasoning.
Post
511
Ring-1T🔥 the trillion-parameter thinking model released by Ant group, the company behind Alipay
inclusionAI/Ring-1T
✨ 1T params (50B active)- MIT license
✨ 128K context (YaRN)
✨ RLVR, Icepop, and ASystem make trillion-scale RL stable
inclusionAI/Ring-1T
✨ 1T params (50B active)- MIT license
✨ 128K context (YaRN)
✨ RLVR, Icepop, and ASystem make trillion-scale RL stable
Post
532
KAT-Dev-72B-Exp🔥 Kuaishou's ( the company behind Kring AI ) new open model for software engineering
Kwaipilot/KAT-Dev-72B-Exp
✨ 72B - Apache2.0
✨ Redesigned attention kernel & training engine for efficient context-aware RL
✨ 74.6% accuracy on SWE-Bench Verified
Kwaipilot/KAT-Dev-72B-Exp
✨ 72B - Apache2.0
✨ Redesigned attention kernel & training engine for efficient context-aware RL
✨ 74.6% accuracy on SWE-Bench Verified
Post
4518
🌎 AI ethics and sustainability are two sides of the same coin.
In our new blog post with Dr. Sasha Luccioni, we argue that separating them (as is too often the case) means missing the bigger picture of how AI systems impact both people and the planet.
Ethical and sustainable AI development can’t be pursued in isolation. The same choices that affect who benefits or is harmed by AI systems also determine how much energy and resources they consume.
We explore how two key concepts, evaluation and transparency, can serve as bridges between these domains:
📊 Evaluation, by moving beyond accuracy or performance metrics to include environmental and social costs, as we’ve done with tools like the AI Energy Score.
🔍 Transparency, by enabling reproducibility, accountability, and environmental reporting through open tools like the Environmental Transparency Space.
AI systems mirror our priorities. If we separate ethics from sustainability, we risk building technologies that are efficient but unjust, or fair but unsustainable.
Read our blog post here: https://huggingface.co/blog/sasha/ethics-sustainability
AIEnergyScore/Leaderboard
sasha/environmental-transparency
In our new blog post with Dr. Sasha Luccioni, we argue that separating them (as is too often the case) means missing the bigger picture of how AI systems impact both people and the planet.
Ethical and sustainable AI development can’t be pursued in isolation. The same choices that affect who benefits or is harmed by AI systems also determine how much energy and resources they consume.
We explore how two key concepts, evaluation and transparency, can serve as bridges between these domains:
📊 Evaluation, by moving beyond accuracy or performance metrics to include environmental and social costs, as we’ve done with tools like the AI Energy Score.
🔍 Transparency, by enabling reproducibility, accountability, and environmental reporting through open tools like the Environmental Transparency Space.
AI systems mirror our priorities. If we separate ethics from sustainability, we risk building technologies that are efficient but unjust, or fair but unsustainable.
Read our blog post here: https://huggingface.co/blog/sasha/ethics-sustainability
AIEnergyScore/Leaderboard
sasha/environmental-transparency
- 1 reply
Post
4431
At the close of the National Holiday🇨🇳, Antgroup drops a new SoTA model.
Ling-1T 🔥 the trillion-parameter flagship of the Ling 2.0 series.
inclusionAI/Ling-1T
✨1T total / 50B active params per token
✨20T+ reasoning-dense tokens (Evo-CoT)
✨128K context via YaRN
✨FP8 training: 15%+ faster, same precision as BF16
✨Hybrid Syntax-Function-Aesthetics reward for front-end & visual generation
Ling-1T 🔥 the trillion-parameter flagship of the Ling 2.0 series.
inclusionAI/Ling-1T
✨1T total / 50B active params per token
✨20T+ reasoning-dense tokens (Evo-CoT)
✨128K context via YaRN
✨FP8 training: 15%+ faster, same precision as BF16
✨Hybrid Syntax-Function-Aesthetics reward for front-end & visual generation
- 1 reply
sasha
authored
a
paper
3 months ago