Загрузка видео...

Не удалось загрузить видео

На главную

this is basically as far as i've been able to push deepseek 0731 flash to "capture the essence" of rainworld. It's very much hampered by its lack of eyes. Switching to the new qwen to see how well it does

yacineMTB's profile picture

kache

395,629 subscribers

13,903 просмотров • 2 месяцев назад •via X (Twitter)

LIVE

Anya Rossi• Live Now

Private livecam show

Комментарии: 0

Нет доступных комментариев

Здесь появятся комментарии из оригинального поста

Похожие видео

Artemis II WDR 2: HOLD as planned at T-33 seconds. Next is to recycle to T-10 mins as part of the WDR. So far this has gone very well indeed!
0:15

Artemis II WDR 2: HOLD as planned at T-33 seconds. Next is to recycle to T-10 mins as part of the WDR. So far this has gone very well indeed!

NSF - NASASpaceflight.com

20,858 просмотров • 7 месяцев назад

DeepSeek V4 Flash 0731 first test. DeepSeek V4 Flash 0731 is ridiculously cheap compared to Kimi K3, GPT-5.6, and Opus 5 For its price, this is a surprisingly capable model. First challenge: Generate a railway station in Three.js. The result exceeded my expectations for such a low-cost model. It’s definitely one to watch if you’re building 3D projects on a tight budget.
0:15

DeepSeek V4 Flash 0731 first test. DeepSeek V4 Flash 0731 is ridiculously cheap compared to Kimi K3, GPT-5.6, and Opus 5 For its price, this is a surprisingly capable model. First challenge: Generate a railway station in Three.js. The result exceeded my expectations for such a low-cost model. It’s definitely one to watch if you’re building 3D projects on a tight budget.

FHILY👑

21,825 просмотров • 2 месяцев назад

DeepSeek R1 is *the* best model available right now. It's at the level of o1, but you can use it for free, and it's much faster. A huge leap forward that nobody saw coming. No wonder so many people are throwing tantrums online trying to discredit the Chinese students who built this. You can use DeepSeek in Visual Studio Code right now: 1. Install the Qodo Gen AI extension 2. Select DeepSeek R1 from their list of models The Qodo team is hosting DeepSeek on their servers, so none of your data will go to China. I've been building a Tetris game using DeepSeek, and this is the most impressive model I've seen so far.
0:17

DeepSeek R1 is *the* best model available right now. It's at the level of o1, but you can use it for free, and it's much faster. A huge leap forward that nobody saw coming. No wonder so many people are throwing tantrums online trying to discredit the Chinese students who built this. You can use DeepSeek in Visual Studio Code right now: 1. Install the Qodo Gen AI extension 2. Select DeepSeek R1 from their list of models The Qodo team is hosting DeepSeek on their servers, so none of your data will go to China. I've been building a Tetris game using DeepSeek, and this is the most impressive model I've seen so far.

Santiago

1,224,473 просмотров • 1 год назад

Last one for week, as I’m out points or whatever. Really working to capture the essence of villains winning. What other content would you all like to see?
0:15

Last one for week, as I’m out points or whatever. Really working to capture the essence of villains winning. What other content would you all like to see?

hypno fantasy

13,639 просмотров • 2 месяцев назад

Trump on the war: Well, hopefully we're toward the end of the war. They want to make a deal, and we'll see how that works out. They're very much wanting to make—Iran is very much wanting to make a deal. We'll see how that works.
0:10

Trump on the war: Well, hopefully we're toward the end of the war. They want to make a deal, and we'll see how that works out. They're very much wanting to make—Iran is very much wanting to make a deal. We'll see how that works.

MeidasTouch

52,642 просмотров • 25 дней назад

I gotta be honest. As much as I love the Post, and appreciate the new weapons, this Wemby GoTo has to be tuned in some kind of way. It just covers way too much space in such little time, and I've yet to see footage of a contest other than "Open".
0:10

I gotta be honest. As much as I love the Post, and appreciate the new weapons, this Wemby GoTo has to be tuned in some kind of way. It just covers way too much space in such little time, and I've yet to see footage of a contest other than "Open".

BlaqKobra 🐍

267,438 просмотров • 1 год назад

Flow is as much about the cat's fear of water as it's fear of others. Water is used as a visual metaphor to convey how the cat feels towards the rest of the characters.
0:10

Flow is as much about the cat's fear of water as it's fear of others. Water is used as a visual metaphor to convey how the cat feels towards the rest of the characters.

Gints Zilbalodis

170,413 просмотров • 1 год назад

Ver. 11.0.0 also adds a new visual indicator of how much health a damaged Inkling or Octoling has left. Use this info to help see which of your allies is in trouble or decide if it's time to push the advantage.
0:10

Ver. 11.0.0 also adds a new visual indicator of how much health a damaged Inkling or Octoling has left. Use this info to help see which of your allies is in trouble or decide if it's time to push the advantage.

Splatoon North America

208,927 просмотров • 8 месяцев назад

✨You know the clam is irresistible, so you need to crack it open as soon as possible. 🦪The all-new red loot Tears of Ocean is basically HoA underwater, except it's iridescent.
0:18

✨You know the clam is irresistible, so you need to crack it open as soon as possible. 🦪The all-new red loot Tears of Ocean is basically HoA underwater, except it's iridescent.

Delta Force Game

32,193 просмотров • 1 год назад

No need to go as far as Zaragoza in Spain to witness the climate disaster, this is Sheffield yesterday after flash flooding wreaked havoc. Why aren’t we talking about this? We need to treat climate change as the emergency it is.
0:19

No need to go as far as Zaragoza in Spain to witness the climate disaster, this is Sheffield yesterday after flash flooding wreaked havoc. Why aren’t we talking about this? We need to treat climate change as the emergency it is.

Claudia Webbe

1,065,153 просмотров • 3 лет назад

We see Dort attempt to basically pull the chair. I disagree that there's a "ton" of contact and refs have allowed this amount all series long. Could it have been called? Sure. But I can see how they let that one go as well
0:15

We see Dort attempt to basically pull the chair. I disagree that there's a "ton" of contact and refs have allowed this amount all series long. Could it have been called? Sure. But I can see how they let that one go as well

BBALLBREAKDOWN

248,594 просмотров • 1 год назад

This is what Palestine has been subjected to by Israel since the very first day of its existence.
0:12

This is what Palestine has been subjected to by Israel since the very first day of its existence.

Haitham in Gaza 🇵🇸🍉

70,991 просмотров • 4 месяцев назад

Okay, I am fairly confident in my hypothesis now. The secret sauce behind Qwen 3.8 27B becomes almost immediately evident during testing. It is not the training data. In fact, I doubt any SFT was involved at all. The model was simply allowed GRPO with a more liberal reasoning context length. As I have explained to you before, distillation through GRPO is powerful, and in certain contexts vastly superior to K/L divergence distillation. Qwen used the outputs of Qwen 3.8 Max as the reinforcement learning objective for Qwen 3.8 27B and told the model reason as much as it has to in order to match the output. Remember, Intelligence is solution, not compression. This is like nesting the output of a function as an additional input to itself, and it is absolutely brilliant.
0:17

Okay, I am fairly confident in my hypothesis now. The secret sauce behind Qwen 3.8 27B becomes almost immediately evident during testing. It is not the training data. In fact, I doubt any SFT was involved at all. The model was simply allowed GRPO with a more liberal reasoning context length. As I have explained to you before, distillation through GRPO is powerful, and in certain contexts vastly superior to K/L divergence distillation. Qwen used the outputs of Qwen 3.8 Max as the reinforcement learning objective for Qwen 3.8 27B and told the model reason as much as it has to in order to match the output. Remember, Intelligence is solution, not compression. This is like nesting the output of a function as an additional input to itself, and it is absolutely brilliant.

Aesther

123,124 просмотров • 1 месяц назад

An interesting characteristic of the Russian “Yolka” interceptor drone is the lack of a warhead, reportedly due to regulatory barriers during its development. It relies solely on kinetic energy to knock down a UAV, which, as seen here, does not always suffice.
0:15

An interesting characteristic of the Russian “Yolka” interceptor drone is the lack of a warhead, reportedly due to regulatory barriers during its development. It relies solely on kinetic energy to knock down a UAV, which, as seen here, does not always suffice.

Roy🇨🇦

234,406 просмотров • 4 месяцев назад

It is very, very clear to see how much the Chelsea Squad respects and listens to Enzo Fernández.
0:13

It is very, very clear to see how much the Chelsea Squad respects and listens to Enzo Fernández.

~

113,660 просмотров • 2 лет назад