Telegram Super

41 posts

Telegram Super

Telegram Super

@TelegramSuper

انضم Aralık 2022
22 يتبع1 المتابعون
Telegram Super
Telegram Super@TelegramSuper·
@JNYBGR You copied me, I made this a while ago. I posted it on the showcase. 😉
English
0
0
0
32
moondream
moondream@moondreamai·
SOTA segmenting through prompts. Player blocking the shot, a jersey, armband, Steph's face. No sweat.
English
9
15
303
13.7K
Telegram Super
Telegram Super@TelegramSuper·
@skalskip92 Vision models, such as sam 3 don't get the love that LLM get. Most LLM are basically the same, so much so they had to create weird benchmarks to justify the hype
English
0
0
2
1.1K
SkalskiP
SkalskiP@skalskip92·
SAM3 video tracking is so good yesterday: collect data, train custom object detector, use tracker to estimate object motion - days today: track anything with text prompt - seconds
English
45
243
2.5K
706.5K
Telegram Super
Telegram Super@TelegramSuper·
Built an agent to explain the concepts in research papers. It generates a podcast and it can highlight quotes in the paper
English
1
0
5
76
Telegram Super
Telegram Super@TelegramSuper·
@KevinQHLin At the moment it does to pick the quote from a page. But that can be automated using rag and an llm
English
0
0
0
23
Kevin Lin
Kevin Lin@KevinQHLin·
@TelegramSuper Look pretty interesting! so does it requires human interact?
English
1
0
0
34
Kevin Lin
Kevin Lin@KevinQHLin·
😫Struggle with preparing presentation video before Deadline (such as NeurIPS)? 🔥🔥Thrilled to share our latest work — Paper2Video — automatically generates presentation videos from papers!! 🚀🚀Just provide your paper➕a portrait photo➕a short audio sample — our Paper2Video will create a full presentation video for you. Try Paper2Video and let us know your thoughts!! 💻 GitHub: github.com/showlab/Paper2… 🌐 Website: showlab.github.io/Paper2Video/ 📜 arXiv: arxiv.org/abs/2510.05096 🤗 HF Dataset: huggingface.co/datasets/ZaynZ… 🤗Daily Paper: huggingface.co/papers/2510.05… 🎉Paper2Video is accpeted by SEA Workshop @ NeurIPS 2025 sea-workshop.github.io and will be presented this year Decemebr! 🙏Kudos to the amazing @zayn42682 and @MikeShou1 Our work is built on top of the multi-agent framework @CamelAIOrg by @guohao_li. Huge thanks @_akhaliq for sharing our work!
AK@_akhaliq

Paper2Video Automatic Video Generation from Scientific Papers

English
7
23
104
27.1K
Telegram Super
Telegram Super@TelegramSuper·
@KevinQHLin Yeah I'm building it component by component (like a react app) there will be others
English
2
0
1
29
Telegram Super
Telegram Super@TelegramSuper·
@KevinQHLin Here is what you could possibly do with the technologies mentioned and a little elbow grease.
English
3
0
0
46
Kevin Lin
Kevin Lin@KevinQHLin·
@TelegramSuper It was great to chat with you! Are you interested in becoming a contributor to this open source project? :D
English
1
0
0
27
Kevin Lin
Kevin Lin@KevinQHLin·
Thanks! I will check out the react you mentioned. For latex, I want to clarify that we focus on academic style slide rather than fancy appearance, there is why we choose latex / beamer (both input and output are by latex) for accurately control. For talking head, we try several candidate methods and found Hallo2 is the best, to ensure the face quality.
English
1
0
2
44
Kevin Lin
Kevin Lin@KevinQHLin·
@TelegramSuper I think the codebase is highly modular. you can first identify the speed bottleneck.
English
3
0
1
27
Telegram Super
Telegram Super@TelegramSuper·
@KevinQHLin With react you can build the templates on the fly. And with js you can do more complex animations even 3d animations
English
0
0
0
15
Telegram Super
Telegram Super@TelegramSuper·
@stevenbjohnson @NotebookLM @EricTopol AI automator📷VOXT — 10:38 Hey I am working on a custom create video feature for NBLM, Inspired by the current notebooklm video feature...what features would you like to have?
English
0
0
0
2