Используем ИИ строго не по назначению.
Заметки про ИИ, IT, компьютерные игры, и всякие инженерные интересности.
Post #4900
10.6K
Я проверяю пуллреквесты Claude Code перед деплоем в прод
- 😁 73
- 🤣 17
- 😢 3
- 💯 2
Channel Public Channel
DE Derp Learning
@derplearning
- Subscribers
- 13.9K
- Photos
- 3.6K
- Videos
- 1K
- Links
- 1.5K
Showing posts older than #4901 · Back to latest
Older Posts 17 shown
Post #4899
3.23K
- 💯 28
- ❤ 8
- 😁 3
Post #4898
3.34K
⚡️ Daily AI Digest — Sun, Feb 15
🔥 MonarchRT — Real-time video diffusion via Monarch matrix attention + Triton kernels
arxiv.org/abs/2602.12271
🔥 DeepGen 1.0 — 5B model beats 80B HunyuanImage (+28%), open-source
arxiv.org/abs/2602.12205 | github.com/DeepGenTeam/DeepGen
🔥 DreamID-Omni — Unified human-centric audio-video gen, multi-character identity+voice
arxiv.org/abs/2602.12160
UniT — Multimodal CoT test-time scaling
arxiv.org/abs/2602.12279
UniDFlow — Unified discrete flow matching, LoRA per task
arxiv.org/abs/2602.12221
FAIL — Adversarial imitation learning for flow matching
arxiv.org/abs/2602.12155
GigaBrain — VLA from world model RL
arxiv.org/abs/2602.12099
arXiv.org MonarchRT: Efficient Attention for Real-Time Video Generation Real-time video generation with Diffusion Transformers is bottlenecked by the quadratic cost of 3D self-attention, especially in real-time regimes that are both few-step and autoregressive, where... 🔥 MonarchRT — Real-time video diffusion via Monarch matrix attention + Triton kernels
arxiv.org/abs/2602.12271
🔥 DeepGen 1.0 — 5B model beats 80B HunyuanImage (+28%), open-source
arxiv.org/abs/2602.12205 | github.com/DeepGenTeam/DeepGen
🔥 DreamID-Omni — Unified human-centric audio-video gen, multi-character identity+voice
arxiv.org/abs/2602.12160
UniT — Multimodal CoT test-time scaling
arxiv.org/abs/2602.12279
UniDFlow — Unified discrete flow matching, LoRA per task
arxiv.org/abs/2602.12221
FAIL — Adversarial imitation learning for flow matching
arxiv.org/abs/2602.12155
GigaBrain — VLA from world model RL
arxiv.org/abs/2602.12099
- 🔥 6
- 🤡 2
- ❤ 1
Post #4897
3.5K
- ❤ 19
- 😁 7
- 🤡 6
- 🔥 4
- 😱 4
- 👍 1
- 👀 1
Post #4896
2.25K
Forwarded from Denis Sexy IT 🤖
Все никак не могу привыкнуть к этому новому миру – скачал Need For Speed Underground, и естественно на маке такие игры запускать это танцы с бубнами (сам запустить я не смог)
Открываешь папку, запускаешь там Claude Code в терминале, и через 10 минут играешь – потому что Opus 4.6 залез там в .exe файл и какими-то патчами починил то, что роняло эмулятор винды
Еще раз – подправил 🪟 файл, сам
Скоро весь легаси софт оживим роботами, дурун-дун-дун
Открываешь папку, запускаешь там Claude Code в терминале, и через 10 минут играешь – потому что Opus 4.6 залез там в .exe файл и какими-то патчами починил то, что роняло эмулятор винды
Еще раз – подправил 🪟 файл, сам
Скоро весь легаси софт оживим роботами, дурун-дун-дун
- 😱 47
- ❤ 11
- 🤩 4
- 🔥 3
- 🤡 3
Post #4895
3.03K
☕️ Daily AI Digest — Feb 14, 2026 (Valentine's Edition 💘)
━━━━━━━━━━━━━━━━━━━━
🔬 PAPERS & RESEARCH
🔥 DeepGen 1.0 — Lightweight 5B unified model for image gen + editing. Beats 80B HunyuanImage by 28% on WISE, 27B Qwen-Image-Edit by 37%. Open-source weights, code, datasets.
📄 arxiv.org/abs/2602.12205
💻 github.com/DeepGenTeam/DeepGen
Latent Forcing — Pixel-space image generation matching latent diffusion quality without a separate decoder. Joint latent+pixel denoising with tuned noise schedules.
📄 arxiv.org/abs/2602.11401
Ctrl&Shift — Geometry-aware object manipulation in images/videos. Move & reorient objects while preserving scene realism, no explicit 3D reconstruction needed.
📄 arxiv.org/abs/2602.11440
Active-Zero — Self-evolving VLMs through active environment exploration (self-play but for vision models). Shifts from passive datasets to active visual data seeking.
📄 arxiv.org/abs/2602.11241
C-JEPA — Causal world models via object-level latent interventions. Extends JEPA from image patches to object-centric representations with counterfactual reasoning.
📄 arxiv.org/abs/2602.11389
New day, new JEPA. Extension, anyone? :D
arXiv.org DeepGen 1.0: A Lightweight Unified Multimodal Model for Advancing... Current unified multimodal models for image generation and editing typically rely on massive parameter scales (e.g., >10B), entailing prohibitive training costs and deployment footprints. In this... ━━━━━━━━━━━━━━━━━━━━
🔬 PAPERS & RESEARCH
🔥 DeepGen 1.0 — Lightweight 5B unified model for image gen + editing. Beats 80B HunyuanImage by 28% on WISE, 27B Qwen-Image-Edit by 37%. Open-source weights, code, datasets.
📄 arxiv.org/abs/2602.12205
💻 github.com/DeepGenTeam/DeepGen
Latent Forcing — Pixel-space image generation matching latent diffusion quality without a separate decoder. Joint latent+pixel denoising with tuned noise schedules.
📄 arxiv.org/abs/2602.11401
Ctrl&Shift — Geometry-aware object manipulation in images/videos. Move & reorient objects while preserving scene realism, no explicit 3D reconstruction needed.
📄 arxiv.org/abs/2602.11440
Active-Zero — Self-evolving VLMs through active environment exploration (self-play but for vision models). Shifts from passive datasets to active visual data seeking.
📄 arxiv.org/abs/2602.11241
C-JEPA — Causal world models via object-level latent interventions. Extends JEPA from image patches to object-centric representations with counterfactual reasoning.
📄 arxiv.org/abs/2602.11389
New day, new JEPA. Extension, anyone? :D
- 🔥 8
- ❤ 1
Post #4891
2.65K
Post #4890
2.65K
Forwarded from Метаверсище и ИИще (Sergey Tsyptsyn ️️)
Qwen AI Slides
Под капотом Qwen3 Agent и Qwen-Image 2.0
Пробуем тут:
https://chat.qwen.ai/?inputFeature=slides
Нужно зарегаться и залогиниться.
Стартапы-слайдоделы напряглись
@cgevent
Под капотом Qwen3 Agent и Qwen-Image 2.0
Пробуем тут:
https://chat.qwen.ai/?inputFeature=slides
Нужно зарегаться и залогиниться.
Стартапы-слайдоделы напряглись
@cgevent
- ❤ 6
- 🔥 2
Post #4889
3.52K
Post #4888
3.3K
- 😁 57
- ❤ 3
- 💯 2
Post #4887
3.41K
🔬 ML PAPERS Дайджест (Image,video,text - arXiv, Feb 13 2026)
🔥 MonarchRT — Efficient attention for real-time video generation via Monarch matrix factorization. Makes autoregressive video DiT viable.
→ arxiv.org/abs/2602.12271
🔥 DreamID-Omni — Unified human-centric audio-video gen. Multi-person identity + voice disentanglement in one framework.
→ arxiv.org/abs/2602.12160
UniT — Unified multimodal CoT with test-time scaling
→ arxiv.org/abs/2602.12279
UniDFlow — Discrete flow matching for multimodal understanding + generation + editing
→ arxiv.org/abs/2602.12221
DeepGen 1.0 — Lightweight unified model for image gen & editing
→ arxiv.org/abs/2602.12205
FAIL — Adversarial imitation learning for flow matching post-training (no reward model needed)
→ arxiv.org/abs/2602.12155
GigaBrain-0.5M — VLA from world model RL (robotics)
→ arxiv.org/abs/2602.12099
arXiv.org MonarchRT: Efficient Attention for Real-Time Video Generation Real-time video generation with Diffusion Transformers is bottlenecked by the quadratic cost of 3D self-attention, especially in real-time regimes that are both few-step and autoregressive, where... 🔥 MonarchRT — Efficient attention for real-time video generation via Monarch matrix factorization. Makes autoregressive video DiT viable.
→ arxiv.org/abs/2602.12271
🔥 DreamID-Omni — Unified human-centric audio-video gen. Multi-person identity + voice disentanglement in one framework.
→ arxiv.org/abs/2602.12160
UniT — Unified multimodal CoT with test-time scaling
→ arxiv.org/abs/2602.12279
UniDFlow — Discrete flow matching for multimodal understanding + generation + editing
→ arxiv.org/abs/2602.12221
DeepGen 1.0 — Lightweight unified model for image gen & editing
→ arxiv.org/abs/2602.12205
FAIL — Adversarial imitation learning for flow matching post-training (no reward model needed)
→ arxiv.org/abs/2602.12155
GigaBrain-0.5M — VLA from world model RL (robotics)
→ arxiv.org/abs/2602.12099
- 🔥 9
- 🗿 1
Post #4886
2.64K
Post #4885
3.03K
Post #4884
3.26K
- 😁 55
- ❤ 9
Post #4883
3.18K
Forwarded from Hacker News
Warcraft III Peon Voice Notifications for Claude Code (🔥 Score: 156+ in 2 hours)
Link: https://readhacker.news/s/6Mywz
Comments: https://readhacker.news/c/6Mywz
Link: https://readhacker.news/s/6Mywz
Comments: https://readhacker.news/c/6Mywz
- 😁 15
- 🔥 5
- 🤡 3
Post #4882
3.05K
Post #4881
3.33K
- 😁 26
- 🌚 7
- 💯 6
- 🗿 2