Sensitive content

This media may contain sensitive content.

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。

Sandrone [full with sound] VA: Marzicc 💫COMMS OPEN 🔞 NYASA Space Captain ✨ model by: Hairy Harzoo #GenshinImpact #Sandrone #nodkrai

OrangeDonut

92,669 subscribers

325,398 次观看 • 11 个月前 •via X (Twitter)

NSFW #GenshinImpact #Sandrone #nodkrai

Anya Rossi• Live Now

Private livecam show

0 条评论

暂无评论

原始帖子的评论将显示在这里

相关视频

sesbian lex (4K ON YT) Sandrone: ally kirser ✨⚔️ VA | VTUBER Columbina: CeladonVA 💫🦴 #GenshinImpact #Columbina #Sandrone

sesbian lex (4K ON YT) Sandrone: ally kirser ✨⚔️ VA | VTUBER Columbina: CeladonVA 💫🦴 #GenshinImpact #Columbina #Sandrone

yespengoo

691,431 次观看 • 9 个月前

Sounded version of precedent animation for Shiretsuna, VA by @PhosporusVA Sounded by SockoDrawing🔞🐻 with sound effect from LeHornySFX3D🔞 (COMMS OPEN) Studio Stray SFX OpenNSFW Model by Valz

Sensitive content

Sounded version of precedent animation for Shiretsuna, VA by @PhosporusVA Sounded by SockoDrawing🔞🐻 with sound effect from LeHornySFX3D🔞 (COMMS OPEN) Studio Stray SFX OpenNSFW Model by Valz

Proto1340 | 🔞NSFW ANIMATOR

66,188 次观看 • 2 年前

THE WINNER TAKES IT ALL - MLP G5 Fan Animation #mlpg5 #MLP #mylittlepony Song Cover by November Ultra Voiced by IMShadow007 2D Artworks by 🌞💫Kris ✧ Comms Open!💫🌞 Twilight Sparkle Model by @Dreeze_GH Watch the full video on YT!

THE WINNER TAKES IT ALL - MLP G5 Fan Animation #mlpg5 #MLP #mylittlepony Song Cover by November Ultra Voiced by IMShadow007 2D Artworks by 🌞💫Kris ✧ Comms Open!💫🌞 Twilight Sparkle Model by @Dreeze_GH Watch the full video on YT!

Makaryo

87,522 次观看 • 2 年前

💗Gentle Dickolas💗 Art: 🔞💖Artdoer69💖🔞 Editing: SpicyRabbitVA - COMISSIONS OPEN (4/7) Blitz: 🔞✨SED21✨Artist/VA! Voice Comms OPEN Stolas: Athena - Stolas but unfortunately more Bri'ish Full Video Here: 💖 #NSFW #HelluvaBossFanart #helluvabossnsfw #Stolitz #HelluvabossStolas #HelluvaBossBlitz

Sensitive content

💗Gentle Dickolas💗 Art: 🔞💖Artdoer69💖🔞 Editing: SpicyRabbitVA - COMISSIONS OPEN (4/7) Blitz: 🔞✨SED21✨Artist/VA! Voice Comms OPEN Stolas: Athena - Stolas but unfortunately more Bri'ish Full Video Here: 💖 #NSFW #HelluvaBossFanart #helluvabossnsfw #Stolitz #HelluvabossStolas #HelluvaBossBlitz

Athena - Stolas but unfortunately more Bri'ish

57,788 次观看 • 3 年前

Exploring Deep Space with SweetNoteVA! [Tentacle Monster] [Tickling] [N/SFW] Tsu & Sweet board a derelict ship in search of valuables, but they encounter something much worse... a monster with a taste for cute space explorers! Special thank you to SFX Editor 🔞RayTracingVA🐕⚜️ | VA & EDITOR | COMMS OPEN!!!

Exploring Deep Space with SweetNoteVA! [Tentacle Monster] [Tickling] [N/SFW] Tsu & Sweet board a derelict ship in search of valuables, but they encounter something much worse... a monster with a taste for cute space explorers! Special thank you to SFX Editor 🔞RayTracingVA🐕⚜️ | VA & EDITOR | COMMS OPEN!!!

Tsuma-ASMR

22,564 次观看 • 1 年前

【Chibi Vtuber Showcase】 🪶Vtuber Birdie Boba🧋🐾 🎨Live2D art by amazing neko 🍎 comms open ⚙️Live2D rigging by Eli ELI • Live2D 💫 Congratulations on the chibi model debut!! What a cute little alien cat! 🐱✨ #Live2D #Vtuber #VGenComm #Chibi

【Chibi Vtuber Showcase】 🪶Vtuber Birdie Boba🧋🐾 🎨Live2D art by amazing neko 🍎 comms open ⚙️Live2D rigging by Eli ELI • Live2D 💫 Congratulations on the chibi model debut!! What a cute little alien cat! 🐱✨ #Live2D #Vtuber #VGenComm #Chibi

ELI • Live2D 💫

10,926 次观看 • 3 个月前

🔞 NSFW COMIC DUB 🔞 Episode 2 of 4 Artist: Stripen (commissions full!)🐧🔞 Main Male VA: Ambrose Kincaid / Comms OPEN Main Female VA: (Me!) Just Dex Sound design: (Me!) Just Dex ~~~ Next episode drops next Friday unless we can get this to 200 likes this time 👀 Thank you for the support so far!

Sensitive content

🔞 NSFW COMIC DUB 🔞 Episode 2 of 4 Artist: Stripen (commissions full!)🐧🔞 Main Male VA: Ambrose Kincaid / Comms OPEN Main Female VA: (Me!) Just Dex Sound design: (Me!) Just Dex ~~~ Next episode drops next Friday unless we can get this to 200 likes this time 👀 Thank you for the support so far!

Just Dex

105,044 次观看 • 2 年前

Not to mention the top-notch VAs, like Cinder - Adult VA 💋✨, Endy 🔞 Lewd Sheep, ✦𝙿𝚒𝚡𝚒𝚎𝚆𝚒𝚕𝚕𝚘𝚠✦ | Voice Actress & Sound, Missus🎙️🔞💋 | COMMS OPEN, Wildflower Phoenix | The Velvet Siren and our Scarlet, who make such a work that worth way more than their regular rates.

Sensitive content

Not to mention the top-notch VAs, like Cinder - Adult VA 💋✨, Endy 🔞 Lewd Sheep, ✦𝙿𝚒𝚡𝚒𝚎𝚆𝚒𝚕𝚕𝚘𝚠✦ | Voice Actress & Sound, Missus🎙️🔞💋 | COMMS OPEN, Wildflower Phoenix | The Velvet Siren and our Scarlet, who make such a work that worth way more than their regular rates.

Sweet Pea Art Studio

19,051 次观看 • 1 年前

Madame Rockwell has a guest to entertain for the day, with her using her silver tongue and other skills at her disposal to make them bend to her will Art drawn by Nachmochok (COMMS OPEN)⭐ Video edited by Tsuma-ASMR Madame Rockwell voiced by 🫦Alexia Renaldis🫦 - Adult VA🔞

Madame Rockwell has a guest to entertain for the day, with her using her silver tongue and other skills at her disposal to make them bend to her will Art drawn by Nachmochok (COMMS OPEN)⭐ Video edited by Tsuma-ASMR Madame Rockwell voiced by 🫦Alexia Renaldis🫦 - Adult VA🔞

Vinny-De-Loud

15,672 次观看 • 3 个月前

What if Atlantis didn't fall by accident... but by breaking universal laws? In her sessions, Dolores Cannon described a civilization that had mastered incredible technologies. Some Atlanteans began experimenting with forces powerful enough to bend space, open portals, and even attempt the creation of new universes. Many believe humanity today stands at another turning point. The difference is that our awareness is growing faster than ever before. Do you feel humanity is waking up in time to choose a different path than Atlantis? ✨🙌🏾💫

What if Atlantis didn't fall by accident... but by breaking universal laws? In her sessions, Dolores Cannon described a civilization that had mastered incredible technologies. Some Atlanteans began experimenting with forces powerful enough to bend space, open portals, and even attempt the creation of new universes. Many believe humanity today stands at another turning point. The difference is that our awareness is growing faster than ever before. Do you feel humanity is waking up in time to choose a different path than Atlantis? ✨🙌🏾💫

🧬Maxpein🧬

32,669 次观看 • 2 个月前

🌼VBridger Mouth Rig Raffle🌼 (🔊Preview has sound | 🇹🇭 ภาษาไทยด้านล่าง) ✨1 Winner will get VBridger Mouth Rigging by Azato Stacia 💤 Outer God VTuber ! (For model with existing base rig only) 🌼Rules🌼 ✨Retweet this post ✨Follow both エーデルワイス & Azato Stacia 💤 Outer God VTuber ❣️You must have your own Live2D Model already rigged and must be able to get permission from your rigger for the base rigging file and provide it to us. ❣️This raffle only includes VBridger Mouth Rigging, Not Full Model Rig! ❣️Tongue Out toggle is NOT included! _________________________ 🌼สุ่มแจก VBridger Mouth Rig🌼 (🔊สามารถเปิดเสียงวิดีโอเพื่อดูตัวอย่างได้) ✨ผู้โชคดี 1 ท่านจะได้รับ VBridger Mouth Rigging โดย Azato Stacia 💤 Outer God VTuber มูลค่า 12,000 บาท! (สำหรับโมเดล Live2D ที่มีริกเต็มตัว / ครึ่งตัวเป็นของตนเองแล้วเท่านั้น) 💡VBridger คือโปรแกรมเสริมสำหรับ VTube Studio ที่จะทำให้โมเดลสามารถจับรูปปากได้ตรงกับการขยับปากของเรามากขึ้น เหมาะกับผู้ที่ต้องการทำคอนเทนต์ร้องเพลง ลิปซิงก์ และอื่น ๆ 🌼กฎในการเข้าร่วม🌼 ✨Retweet โพสต์นี้ ✨กด Follow エーデルワイス และ Azato Stacia 💤 Outer God VTuber ❣️ผู้โชคดีต้องมีโมเดล Live2D ที่พร้อมใช้งานเป็นของตนเองอยู่แล้ว และสามารถขอรับไฟล์ต้นฉบับจากริกเกอร์ของท่าน เพื่อส่งให้เราริก VBridger เสริมให้ได้ ❣️รางวัลจากการสุ่มแจกมีเพียงริกปาก VBridger เท่านั้น ไม่ใช่ริกโมเดลเต็มตัว ❣️ริกปากที่ได้รับจากการสุ่มแจกจะไม่สามารถแลบลิ้นได้ #EdelweissMonthlyRaffle

🌼VBridger Mouth Rig Raffle🌼 (🔊Preview has sound | 🇹🇭 ภาษาไทยด้านล่าง) ✨1 Winner will get VBridger Mouth Rigging by Azato Stacia 💤 Outer God VTuber ! (For model with existing base rig only) 🌼Rules🌼 ✨Retweet this post ✨Follow both エーデルワイス & Azato Stacia 💤 Outer God VTuber ❣️You must have your own Live2D Model already rigged and must be able to get permission from your rigger for the base rigging file and provide it to us. ❣️This raffle only includes VBridger Mouth Rigging, Not Full Model Rig! ❣️Tongue Out toggle is NOT included! _________________________ 🌼สุ่มแจก VBridger Mouth Rig🌼 (🔊สามารถเปิดเสียงวิดีโอเพื่อดูตัวอย่างได้) ✨ผู้โชคดี 1 ท่านจะได้รับ VBridger Mouth Rigging โดย Azato Stacia 💤 Outer God VTuber มูลค่า 12,000 บาท! (สำหรับโมเดล Live2D ที่มีริกเต็มตัว / ครึ่งตัวเป็นของตนเองแล้วเท่านั้น) 💡VBridger คือโปรแกรมเสริมสำหรับ VTube Studio ที่จะทำให้โมเดลสามารถจับรูปปากได้ตรงกับการขยับปากของเรามากขึ้น เหมาะกับผู้ที่ต้องการทำคอนเทนต์ร้องเพลง ลิปซิงก์ และอื่น ๆ 🌼กฎในการเข้าร่วม🌼 ✨Retweet โพสต์นี้ ✨กด Follow エーデルワイス และ Azato Stacia 💤 Outer God VTuber ❣️ผู้โชคดีต้องมีโมเดล Live2D ที่พร้อมใช้งานเป็นของตนเองอยู่แล้ว และสามารถขอรับไฟล์ต้นฉบับจากริกเกอร์ของท่าน เพื่อส่งให้เราริก VBridger เสริมให้ได้ ❣️รางวัลจากการสุ่มแจกมีเพียงริกปาก VBridger เท่านั้น ไม่ใช่ริกโมเดลเต็มตัว ❣️ริกปากที่ได้รับจากการสุ่มแจกจะไม่สามารถแลบลิ้นได้ #EdelweissMonthlyRaffle

Edelweiss Studio

39,792 次观看 • 2 年前

Today we’re celebrating 10 years of the Meta FAIR lab in Paris by sharing a collection of new models, datasets and some exciting milestones in the impacts of open source — all laddering up to our ongoing work to achieve Advanced Machine Intelligence (AMI). 1️⃣ Meta PARTNR is a framework for human-robot collaboration that builds on our existing work in this space with a new dataset and a large planning model enabling robots to accomplish complex tasks alongside humans. 2️⃣ Meta Audiobox Aesthetics enables the automatic evaluation of audio aesthetics, providing a comprehensive assessment of audio quality across speech, music and sound. 3️⃣ Open Source Machine Translation Benchmark is a carefully crafted collection with the aim of building an unprecedented multilingual machine translation benchmark for the community. 4️⃣ Two new breakthrough studies using AI to further our understanding of language in the brain.

Today we’re celebrating 10 years of the Meta FAIR lab in Paris by sharing a collection of new models, datasets and some exciting milestones in the impacts of open source — all laddering up to our ongoing work to achieve Advanced Machine Intelligence (AMI). 1️⃣ Meta PARTNR is a framework for human-robot collaboration that builds on our existing work in this space with a new dataset and a large planning model enabling robots to accomplish complex tasks alongside humans. 2️⃣ Meta Audiobox Aesthetics enables the automatic evaluation of audio aesthetics, providing a comprehensive assessment of audio quality across speech, music and sound. 3️⃣ Open Source Machine Translation Benchmark is a carefully crafted collection with the aim of building an unprecedented multilingual machine translation benchmark for the community. 4️⃣ Two new breakthrough studies using AI to further our understanding of language in the brain.

AI at Meta

85,975 次观看 • 1 年前

Revisiting the golden era of the 90s with a soulful rendition of Meri Sanson Mein Basa Hai—a 90s classic, originally sung by Udit Narayan. Now streaming on the Rahul Deshpande Collective for you. Let nostalgia take the lead. 💫🎶 full video : ✨ Special Thanks: Vishal, Shweta, Vishwa & Shiv Chordia 🎵 Original Credits: Singer - @uditnarayanmusic Composers - Nusrat Fateh Ali Khan Lyricist - @javedjaduofficial Movie - Aur Pyaar Ho Gaya (1997) 🎵 Video Credits: Acoustic Guitar - @sanjoygtr Tabla - @adiathalye Keyboard - Anay Gadgil Octapad - @rohanvanage Bass Guitar - @fenderlove93 Harmonium - Nusrat Apoorv Mix and Mastered By - Ishaan Devasthali Sound - Mahesh Gaikwad & Lakir Mehta Video By - Aditya Ganpule, penny georgiou.garge & team Managed By - @komalkhode27 Don’t forget to Like, Share, and Comment! #RahulDeshpande #RahulDeshpandeCollective #MeriSansonMeinBasaHai #UditNarayan #Covers #IndianMusic #IndianSinger

Revisiting the golden era of the 90s with a soulful rendition of Meri Sanson Mein Basa Hai—a 90s classic, originally sung by Udit Narayan. Now streaming on the Rahul Deshpande Collective for you. Let nostalgia take the lead. 💫🎶 full video : ✨ Special Thanks: Vishal, Shweta, Vishwa & Shiv Chordia 🎵 Original Credits: Singer - @uditnarayanmusic Composers - Nusrat Fateh Ali Khan Lyricist - @javedjaduofficial Movie - Aur Pyaar Ho Gaya (1997) 🎵 Video Credits: Acoustic Guitar - @sanjoygtr Tabla - @adiathalye Keyboard - Anay Gadgil Octapad - @rohanvanage Bass Guitar - @fenderlove93 Harmonium - Nusrat Apoorv Mix and Mastered By - Ishaan Devasthali Sound - Mahesh Gaikwad & Lakir Mehta Video By - Aditya Ganpule, penny georgiou.garge & team Managed By - @komalkhode27 Don’t forget to Like, Share, and Comment! #RahulDeshpande #RahulDeshpandeCollective #MeriSansonMeinBasaHai #UditNarayan #Covers #IndianMusic #IndianSinger

Rahul Deshpande

13,114 次观看 • 1 年前

YOUR DREAMS ARE NOT RANDOM.. (They’re portals) - Every night you’re either dreaming… or being dreamed through. - If your plasma shields aren’t sealed, your consciousness can drift into false light matrices where energy is harvested, implanted, or looped while you sleep. - This is why so many people wake up anxious, fragmented, or haunted by dreams they can’t shake. - You didn’t just have a nightmare—you left the door open. - But here’s the truth: you are the gatekeeper. - Sleep isn’t just rest, it’s a portal. - And you can seal your dream space every night by commanding it. - Repeat this before bed: “I now seal my body, field, and dream space in eternal source light. - Only energies aligned with truth, love, and eternal law may enter. All others are dissolved and returned to source. I sleep sovereign, protected, and whole.” - Your dreams can be messages, not manipulations—when you choose sovereignty. ✨🙌🏾💫 © Samuel Lee

YOUR DREAMS ARE NOT RANDOM.. (They’re portals) - Every night you’re either dreaming… or being dreamed through. - If your plasma shields aren’t sealed, your consciousness can drift into false light matrices where energy is harvested, implanted, or looped while you sleep. - This is why so many people wake up anxious, fragmented, or haunted by dreams they can’t shake. - You didn’t just have a nightmare—you left the door open. - But here’s the truth: you are the gatekeeper. - Sleep isn’t just rest, it’s a portal. - And you can seal your dream space every night by commanding it. - Repeat this before bed: “I now seal my body, field, and dream space in eternal source light. - Only energies aligned with truth, love, and eternal law may enter. All others are dissolved and returned to source. I sleep sovereign, protected, and whole.” - Your dreams can be messages, not manipulations—when you choose sovereignty. ✨🙌🏾💫 © Samuel Lee

🧬Maxpein🧬

25,886 次观看 • 7 个月前

YOUR DREAMS ARE NOT RANDOM.. (They’re portals) - Every night you’re either dreaming… or being dreamed through. - If your plasma shields aren’t sealed, your consciousness can drift into false light matrices where energy is harvested, implanted, or looped while you sleep. - This is why so many people wake up anxious, fragmented, or haunted by dreams they can’t shake. - You didn’t just have a nightmare—you left the door open. - But here’s the truth: you are the gatekeeper. - Sleep isn’t just rest, it’s a portal. - And you can seal your dream space every night by commanding it. - Repeat this before bed: “I now seal my body, field, and dream space in eternal source light. - Only energies aligned with truth, love, and eternal law may enter. All others are dissolved and returned to source. I sleep sovereign, protected, and whole.” - Your dreams can be messages, not manipulations—when you choose sovereignty. ✨🙌🏾💫 © Samuel Lee

YOUR DREAMS ARE NOT RANDOM.. (They’re portals) - Every night you’re either dreaming… or being dreamed through. - If your plasma shields aren’t sealed, your consciousness can drift into false light matrices where energy is harvested, implanted, or looped while you sleep. - This is why so many people wake up anxious, fragmented, or haunted by dreams they can’t shake. - You didn’t just have a nightmare—you left the door open. - But here’s the truth: you are the gatekeeper. - Sleep isn’t just rest, it’s a portal. - And you can seal your dream space every night by commanding it. - Repeat this before bed: “I now seal my body, field, and dream space in eternal source light. - Only energies aligned with truth, love, and eternal law may enter. All others are dissolved and returned to source. I sleep sovereign, protected, and whole.” - Your dreams can be messages, not manipulations—when you choose sovereignty. ✨🙌🏾💫 © Samuel Lee

🧬Maxpein🧬

43,393 次观看 • 10 个月前

Google just proved that bigger isn't always better. Their 308M parameter model is outperforming models 2x its size. Google just released 𝗘𝗺𝗯𝗲𝗱𝗱𝗶𝗻𝗴𝗚𝗲𝗺𝗺𝗮, and it's proving that lightweight embedding models can punch way above their weight class. At just 308M parameters (578MB), it's the new state-of-the-art for models under 500M parameters across MTEB multilingual, English, and code benchmarks. But the really impressive part is that it ranks 8th overall on MTEB(Multilingual, v2) - that's 𝟭𝟳 𝗽𝗹𝗮𝗰𝗲𝘀 above the second-best sub-500M model, and it's delivering performance 𝗰𝗼𝗺𝗽𝗮𝗿𝗮𝗯𝗹𝗲 𝘁𝗼 𝗺𝗼𝗱𝗲𝗹𝘀 𝗻𝗲𝗮𝗿𝗹𝘆 𝗱𝗼𝘂𝗯𝗹𝗲 𝗶𝘁𝘀 𝘀𝗶𝘇𝗲. There are three key parts of their training recipe that sets it apart: 𝟭. 𝗘𝗻𝗰𝗼𝗱𝗲𝗿-𝗗𝗲𝗰𝗼𝗱𝗲𝗿 𝗜𝗻𝗶𝘁𝗶𝗮𝗹𝗶𝘇𝗮𝘁𝗶𝗼𝗻 Instead of starting from a decoder-only Gemma 3 model, they first adapted it to encoder-decoder, then used just the encoder. By basing EmbeddingGemma off an LLM that already has world and language understanding, it gives it a stronger starting point. 𝟮. 𝗧𝗵𝗿𝗲𝗲-𝗟𝗼𝘀𝘀 𝗧𝗿𝗮𝗶𝗻𝗶𝗻𝗴 They combine three different loss functions, instead of just having one: • Contrastive loss (NCE) with in-batch negatives and hardness weighting • Spread-out regularization to ensure embeddings utilize the full space (for quantization and ANN retrieval) • Embedding matching distillation from Gemini Embedding - not just learning from relevance scores, but directly aligning the embedding space with the teacher model 𝟯. 𝗠𝗼𝗱𝗲𝗹 𝗦𝗼𝘂𝗽𝗶𝗻𝗴 Rather than just averaging checkpoints from the same training run, they use optimization techniques to find multiple specialized training mixtures. Each mixture creates an "expert" model in different domains, and averaging all their parameters creates a final model that's actually better than individual models. Extras: • Matryoshka embeddings supporting 768, 512, 256, and 128 dimensions • Quantization-aware training - maintains quality even at int4 precision • 100+ languages from Gemma 3 pretraining • Exceptional performance on low-resource languages (check their XTREME-UP results) Is it the absolute best embedding model? No - Gemini Embedding still leads overall. But that's not really the point. EmbeddingGemma proves you can achieve state-of-the-art performance in a small package that's actually deployable on-device, in low-latency applications, and in resource-constrained environments. This makes good embeddings accessible for use cases that I'm seeing more and more: offline applications, privacy-sensitive deployments, and high-throughput scenarios where inference cost actually matters. Full paper: Shoutout to the EmbeddingGemma team at Google DeepMind for this awesome open source work 💙 and to Daniel Williams for helping me with this video! 🫶

Google just proved that bigger isn't always better. Their 308M parameter model is outperforming models 2x its size. Google just released 𝗘𝗺𝗯𝗲𝗱𝗱𝗶𝗻𝗴𝗚𝗲𝗺𝗺𝗮, and it's proving that lightweight embedding models can punch way above their weight class. At just 308M parameters (578MB), it's the new state-of-the-art for models under 500M parameters across MTEB multilingual, English, and code benchmarks. But the really impressive part is that it ranks 8th overall on MTEB(Multilingual, v2) - that's 𝟭𝟳 𝗽𝗹𝗮𝗰𝗲𝘀 above the second-best sub-500M model, and it's delivering performance 𝗰𝗼𝗺𝗽𝗮𝗿𝗮𝗯𝗹𝗲 𝘁𝗼 𝗺𝗼𝗱𝗲𝗹𝘀 𝗻𝗲𝗮𝗿𝗹𝘆 𝗱𝗼𝘂𝗯𝗹𝗲 𝗶𝘁𝘀 𝘀𝗶𝘇𝗲. There are three key parts of their training recipe that sets it apart: 𝟭. 𝗘𝗻𝗰𝗼𝗱𝗲𝗿-𝗗𝗲𝗰𝗼𝗱𝗲𝗿 𝗜𝗻𝗶𝘁𝗶𝗮𝗹𝗶𝘇𝗮𝘁𝗶𝗼𝗻 Instead of starting from a decoder-only Gemma 3 model, they first adapted it to encoder-decoder, then used just the encoder. By basing EmbeddingGemma off an LLM that already has world and language understanding, it gives it a stronger starting point. 𝟮. 𝗧𝗵𝗿𝗲𝗲-𝗟𝗼𝘀𝘀 𝗧𝗿𝗮𝗶𝗻𝗶𝗻𝗴 They combine three different loss functions, instead of just having one: • Contrastive loss (NCE) with in-batch negatives and hardness weighting • Spread-out regularization to ensure embeddings utilize the full space (for quantization and ANN retrieval) • Embedding matching distillation from Gemini Embedding - not just learning from relevance scores, but directly aligning the embedding space with the teacher model 𝟯. 𝗠𝗼𝗱𝗲𝗹 𝗦𝗼𝘂𝗽𝗶𝗻𝗴 Rather than just averaging checkpoints from the same training run, they use optimization techniques to find multiple specialized training mixtures. Each mixture creates an "expert" model in different domains, and averaging all their parameters creates a final model that's actually better than individual models. Extras: • Matryoshka embeddings supporting 768, 512, 256, and 128 dimensions • Quantization-aware training - maintains quality even at int4 precision • 100+ languages from Gemma 3 pretraining • Exceptional performance on low-resource languages (check their XTREME-UP results) Is it the absolute best embedding model? No - Gemini Embedding still leads overall. But that's not really the point. EmbeddingGemma proves you can achieve state-of-the-art performance in a small package that's actually deployable on-device, in low-latency applications, and in resource-constrained environments. This makes good embeddings accessible for use cases that I'm seeing more and more: offline applications, privacy-sensitive deployments, and high-throughput scenarios where inference cost actually matters. Full paper: Shoutout to the EmbeddingGemma team at Google DeepMind for this awesome open source work 💙 and to Daniel Williams for helping me with this video! 🫶

Victoria Slocum

21,610 次观看 • 9 个月前

Open weight Kimi K3 performs at Claude Fable 5 level on 3D arcade games! We gave both models the same prompts: ✦ Pac-Man with four hunting ghosts ✦ Nokia Snake remade in 3D ✦ Retro space pinball table with real physics Outputs (thinking included): ✦ Kimi K3 (local, 8x B300): ~237K tokens, $0 ✦ Claude Fable 5: 92K tokens, $4.50 Every one of Kimi's tokens was free on our GPUs. Even via Moonshot's API the same Kimi run would be $3.60, still under Fable. Kimi held its quality against the strongest cloud model. Its snake is the liveliest scene of all six: soft shadows, floating dust, a flicking tongue, and a bot that checks it can still reach its own tail before every move. Fable's snake plays clean but looks flat next to it. Pac-Man is the closest call of the three. Kimi builds a fresh labyrinth on every restart and draws a live minimap in the corner while the red ghost tracks the player turn by turn through the corridors, and it even caught our bot on camera. Fable's maze answers with pearl pellets, +10 popups bursting over the floor and a wide-eyed cartoon ghost patrol. Fable won the pinball table: painted playfield art, a chrome ball that mirrors the neon lights, even little synth sound effects.

Open weight Kimi K3 performs at Claude Fable 5 level on 3D arcade games! We gave both models the same prompts: ✦ Pac-Man with four hunting ghosts ✦ Nokia Snake remade in 3D ✦ Retro space pinball table with real physics Outputs (thinking included): ✦ Kimi K3 (local, 8x B300): ~237K tokens, $0 ✦ Claude Fable 5: 92K tokens, $4.50 Every one of Kimi's tokens was free on our GPUs. Even via Moonshot's API the same Kimi run would be $3.60, still under Fable. Kimi held its quality against the strongest cloud model. Its snake is the liveliest scene of all six: soft shadows, floating dust, a flicking tongue, and a bot that checks it can still reach its own tail before every move. Fable's snake plays clean but looks flat next to it. Pac-Man is the closest call of the three. Kimi builds a fresh labyrinth on every restart and draws a live minimap in the corner while the red ghost tracks the player turn by turn through the corridors, and it even caught our bot on camera. Fable's maze answers with pearl pellets, +10 popups bursting over the floor and a wide-eyed cartoon ghost patrol. Fable won the pinball table: painted playfield art, a chrome ball that mirrors the neon lights, even little synth sound effects.

Atomic Agent

309,981 次观看 • 1 个月前

AI TENNIS ANALYSIS. A FULL COMPUTER VISION SYSTEM. BUILT ON YOLO, PYTORCH, AND KEYPOINT EXTRACTION. Take any tennis match broadcast, any camera angle, any resolution. Feed it into the pipeline. YOLO detects both players and the tennis ball frame by frame. No manual labeling, no pre-annotated dataset. A fine-tuned YOLOv5 model trained on a Roboflow tennis ball dataset handles the ball - the hardest object to track in any sport. Tiny, fast, constantly occluded. The model finds it anyway. Trackers maintain identity across frames so Player 1 stays Player 1 from the first serve to match point. But detection is just the start. A ResNet50 CNN trained in PyTorch predicts court keypoints from every frame - the corners, service lines, baselines, net posts. Fourteen points that define the entire playing surface geometry. From those keypoints the system builds a homography matrix and warps the broadcast perspective into a top-down mini court with real coordinates. Now every player has a position in real space, not pixel space. Every frame becomes a measurement. Every rally becomes a dataset. Player movement speed - calculated from position deltas between frames, converted to meters per second through the homography. Ball shot speed - measured from the ball trajectory across consecutive detections. Number of shots per rally - counted automatically through ball direction changes. All of this rendered live on the video as an overlay. A mini court in the corner showing both players as dots moving in real time. Stats updating after every point. OpenCV handles the rendering. Pandas handles the math. PyTorch handles the intelligence. YOLO handles the eyes. No Hawkeye subscription, no court-embedded sensors, no tracking chips in the ball. A Python script, a trained model, and a GPU. The full code is on GitHub. The tutorial walks through every module - from ball detector training to court keypoint extraction to the final statistical overlay. Professional teams used to need broadcast deals and proprietary hardware for this kind of analysis. Now you build it in an afternoon with open-source tools. Trading here: Computer vision didn't just enter tennis. It made the expensive stuff free.

AI TENNIS ANALYSIS. A FULL COMPUTER VISION SYSTEM. BUILT ON YOLO, PYTORCH, AND KEYPOINT EXTRACTION. Take any tennis match broadcast, any camera angle, any resolution. Feed it into the pipeline. YOLO detects both players and the tennis ball frame by frame. No manual labeling, no pre-annotated dataset. A fine-tuned YOLOv5 model trained on a Roboflow tennis ball dataset handles the ball - the hardest object to track in any sport. Tiny, fast, constantly occluded. The model finds it anyway. Trackers maintain identity across frames so Player 1 stays Player 1 from the first serve to match point. But detection is just the start. A ResNet50 CNN trained in PyTorch predicts court keypoints from every frame - the corners, service lines, baselines, net posts. Fourteen points that define the entire playing surface geometry. From those keypoints the system builds a homography matrix and warps the broadcast perspective into a top-down mini court with real coordinates. Now every player has a position in real space, not pixel space. Every frame becomes a measurement. Every rally becomes a dataset. Player movement speed - calculated from position deltas between frames, converted to meters per second through the homography. Ball shot speed - measured from the ball trajectory across consecutive detections. Number of shots per rally - counted automatically through ball direction changes. All of this rendered live on the video as an overlay. A mini court in the corner showing both players as dots moving in real time. Stats updating after every point. OpenCV handles the rendering. Pandas handles the math. PyTorch handles the intelligence. YOLO handles the eyes. No Hawkeye subscription, no court-embedded sensors, no tracking chips in the ball. A Python script, a trained model, and a GPU. The full code is on GitHub. The tutorial walks through every module - from ball detector training to court keypoint extraction to the final statistical overlay. Professional teams used to need broadcast deals and proprietary hardware for this kind of analysis. Now you build it in an afternoon with open-source tools. Trading here: Computer vision didn't just enter tennis. It made the expensive stuff free.

zostaff

120,370 次观看 • 4 个月前