Загрузка видео...

Не удалось загрузить видео

На главную

Introducing Moka: a Go-playing model small enough to fit in a website Loading an existing model (KataGo) is 17 MB over the network (!!) Moka is 135 KB (100x smaller) and still ranks ~2 kyu

aidenybai's profile picture

Aiden Bai

69,603 subscribers

63,140 просмотров • 2 месяцев назад •via X (Twitter)

Наука и технологии
LIVE

Anya Rossi• Live Now

Private livecam show

Комментарии: 0

Нет доступных комментариев

Здесь появятся комментарии из оригинального поста

Похожие видео

are you kidding me 😭 in their spotify ad, illit members sent messages to moka and celebrated her birthday. they even made a playlist for her.... the chat room is named "moka birthday room" and the car license plate is "04mo 1008" her birthdate 🥹
0:13

are you kidding me 😭 in their spotify ad, illit members sent messages to moka and celebrated her birthday. they even made a playlist for her.... the chat room is named "moka birthday room" and the car license plate is "04mo 1008" her birthdate 🥹

moka pics

41,815 просмотров • 12 дней назад

It's been over 10 years since the last redesign of the Jeep Cherokee, and while the new 2027 Hybrid model is an improvement it's still not good enough
0:18

It's been over 10 years since the last redesign of the Jeep Cherokee, and while the new 2027 Hybrid model is an improvement it's still not good enough

Edmunds

25,717 просмотров • 5 месяцев назад

Minji is beautiful Yunah is a man Minju reconstructed her face Wonhee is slow and failed a year Illit copied newjeans AND blackpink in the span of 2 days Iroha has horse teeth Moka can't sing
0:13

Minji is beautiful Yunah is a man Minju reconstructed her face Wonhee is slow and failed a year Illit copied newjeans AND blackpink in the span of 2 days Iroha has horse teeth Moka can't sing

zayn

19,623 просмотров • 6 месяцев назад

How does Edge0 run a 35B model with only 1–2.5 GB of peak memory? In simple terms: • The full model stays in storage. • Only the parameters needed for the current step are loaded into memory. • The next parameters are prepared while the current step is running. • Memory is reused instead of loading the entire model at once. Large model. Small working set. Fully on-device.
0:10

How does Edge0 run a 35B model with only 1–2.5 GB of peak memory? In simple terms: • The full model stays in storage. • Only the parameters needed for the current step are loaded into memory. • The next parameters are prepared while the current step is running. • Memory is reused instead of loading the entire model at once. Large model. Small working set. Fully on-device.

Samuel Zeng

34,026 просмотров • 23 дней назад

She is studying to be a doctor 🫣 If my doctor was an IG model I would just get in a Joby and fly to the next town over 😅
0:13

She is studying to be a doctor 🫣 If my doctor was an IG model I would just get in a Joby and fly to the next town over 😅

Matt Wallace

30,198 просмотров • 18 дней назад

I wanted to try Maple-Preview by DeepGrove on iPhone so I simply ported the model to MLX Swift using Bionic Maple runs at ~80tk/s on iPhone 17 Pro with my quick port, it’s a 20B-A1B model at 2-bit The model is close to Qwen3.5 35-A2B in many benchs and runs easily on iPhone
0:15

I wanted to try Maple-Preview by DeepGrove on iPhone so I simply ported the model to MLX Swift using Bionic Maple runs at ~80tk/s on iPhone 17 Pro with my quick port, it’s a 20B-A1B model at 2-bit The model is close to Qwen3.5 35-A2B in many benchs and runs easily on iPhone

Adrien Grondin

13,376 просмотров • 1 месяц назад

Introducing Jan v3, with updates to Jan Desktop v0.7.6 💛 Jan v3 is our first v3 model, a 4B base built for fine-tuning and fast local use, with stronger math and coding. This release also includes a small Jan Desktop update, starting with a UI refresh as we move toward a more unified Jan experience. Try it: - Jan v3 is available in Jan Desktop and at - Get the latest desktop app at Model: - Jan-v3-4B: - Jan-v3-4B-GGUF: Thanks Qwen for the base model and Georgi Gerganov for llama.cpp 💛
0:17

Introducing Jan v3, with updates to Jan Desktop v0.7.6 💛 Jan v3 is our first v3 model, a 4B base built for fine-tuning and fast local use, with stronger math and coding. This release also includes a small Jan Desktop update, starting with a UI refresh as we move toward a more unified Jan experience. Try it: - Jan v3 is available in Jan Desktop and at - Get the latest desktop app at Model: - Jan-v3-4B: - Jan-v3-4B-GGUF: Thanks Qwen for the base model and Georgi Gerganov for llama.cpp 💛

👋 Jan

53,900 просмотров • 8 месяцев назад

Okay this new model is insane. Generate an image, then give it to the model to turn it into a 3D model with matching texture in seconds. Creativity on Spielwerk will be endless with this
0:10

Okay this new model is insane. Generate an image, then give it to the model to turn it into a 3D model with matching texture in seconds. Creativity on Spielwerk will be endless with this

Eike Drescher

37,987 просмотров • 10 месяцев назад

Small detail, but I like how the Model Y’s brake pedal no longer moves while in FSD. The Model Y now uses a dual master cylinder brake system. One master cylinder is connected directly to the driver’s brake pedal, and the other is controlled electronically by FSD.
0:16

Small detail, but I like how the Model Y’s brake pedal no longer moves while in FSD. The Model Y now uses a dual master cylinder brake system. One master cylinder is connected directly to the driver’s brake pedal, and the other is controlled electronically by FSD.

Nic Cruz Patane

145,485 просмотров • 5 месяцев назад

Tesla’s Supercharger network had a ridiculous Q2. 2 terawatt-hours delivered in 3 months. That is roughly enough electricity to power 180,000+ U.S. homes for a year. 60 million charging sessions. 266 kWh sold per plug per day. Wait times below 0.5%. 5,000 site maps live. 17% YoY growth. The wild part is that the network is getting busier while the waiting gets smaller. A very satisfying graph. Tesla X Freeze / Writer: Annette, Designer: Janné
0:16

Tesla’s Supercharger network had a ridiculous Q2. 2 terawatt-hours delivered in 3 months. That is roughly enough electricity to power 180,000+ U.S. homes for a year. 60 million charging sessions. 266 kWh sold per plug per day. Wait times below 0.5%. 5,000 site maps live. 17% YoY growth. The wild part is that the network is getting busier while the waiting gets smaller. A very satisfying graph. Tesla X Freeze / Writer: Annette, Designer: Janné

Mario Nawfal

76,855 просмотров • 3 месяцев назад

After playing with the Jev hybrid model, I started looking into OpenJev, and started with a halite 1 implementation with: GLM 5.3 Flash + Jev [vs] GLM 5.3 Flash + OpenJev OpenJev is just a 4B model handily defeating Jev (in this tiny silly toy example ofc). Now I'm super curious to learn what Jev's actual model size is.
0:13

After playing with the Jev hybrid model, I started looking into OpenJev, and started with a halite 1 implementation with: GLM 5.3 Flash + Jev [vs] GLM 5.3 Flash + OpenJev OpenJev is just a 4B model handily defeating Jev (in this tiny silly toy example ofc). Now I'm super curious to learn what Jev's actual model size is.

Harrison Kinsley

24,958 просмотров • 11 дней назад

FLUX.2 is live on Runware! D0 drop! 🔥 Built by Black Forest Labs, FLUX.2 is the new SOTA model that brings better control over structure, text, and references in image generation. This is a huge step forward for image gen & image editing. And we’re here to offer you the best prices on D0. This model comes in three versions. Details below.
0:15

FLUX.2 is live on Runware! D0 drop! 🔥 Built by Black Forest Labs, FLUX.2 is the new SOTA model that brings better control over structure, text, and references in image generation. This is a huge step forward for image gen & image editing. And we’re here to offer you the best prices on D0. This model comes in three versions. Details below.

Runware

293,432 просмотров • 10 месяцев назад

1/ Gemini 2.5 is here, and it’s our most intelligent AI model ever. Our first 2.5 model, Gemini 2.5 Pro Experimental is a state-of-the-art thinking model, leading in a wide range of benchmarks – with impressive improvements in enhanced reasoning and coding and now #1 on Arena by a significant margin. With a model this intelligent, we wanted to get it to people as quickly as possible. Find it on Google AI Studio and in the Google Gemini for Gemini Advanced users now – and in Vertex in the coming weeks. This is the start of a new era of thinking models – and we can’t wait to see where things go from here.
0:14

1/ Gemini 2.5 is here, and it’s our most intelligent AI model ever. Our first 2.5 model, Gemini 2.5 Pro Experimental is a state-of-the-art thinking model, leading in a wide range of benchmarks – with impressive improvements in enhanced reasoning and coding and now #1 on Arena by a significant margin. With a model this intelligent, we wanted to get it to people as quickly as possible. Find it on Google AI Studio and in the Google Gemini for Gemini Advanced users now – and in Vertex in the coming weeks. This is the start of a new era of thinking models – and we can’t wait to see where things go from here.

Sundar Pichai

864,841 просмотров • 1 год назад

Haven't seen someone more beautiful than her in a while. Eva Delonne is an AI-generated virtual model and influencer, Half black, Half white. Ladies and gentlemen AI is taking over
0:15

Haven't seen someone more beautiful than her in a while. Eva Delonne is an AI-generated virtual model and influencer, Half black, Half white. Ladies and gentlemen AI is taking over

iamtakudzwa

55,009 просмотров • 7 месяцев назад

The new PrismML’s Bonsai 27B model running on iPhone 17 Pro Bonsai 27B is the first 27B-class model to run on a phone and it uses less than 5GB of memory Pushing the limits on what can run a phone
0:14

The new PrismML’s Bonsai 27B model running on iPhone 17 Pro Bonsai 27B is the first 27B-class model to run on a phone and it uses less than 5GB of memory Pushing the limits on what can run a phone

Adrien Grondin

67,095 просмотров • 2 месяцев назад