DeepSeek-R1 is here! First Open O1 level Model
ฝัง
- เผยแพร่เมื่อ 31 ม.ค. 2025
- DeepSeek Just released their first reasoning model called Deepseek-R1 that is on par with Openai O1 reasoning model. We will have a quick look at the release in this video.
#deepseek #reasoning #deepseekR1
LINKS:
chat.deepseek....
github.com/dee...
github.com/dee...
💻 RAG Beyond Basics Course:
prompt-s-site....
Let's Connect:
🦾 Discord: / discord
☕ Buy me a Coffee: ko-fi.com/prom...
|🔴 Patreon: / promptengineering
💼Consulting: calendly.com/e...
📧 Business Contact: engineerprompt@gmail.com
Become Member: tinyurl.com/y5h...
💻 Pre-configured localGPT VM: bit.ly/localGPT (use Code: PromptEngineering for 50% off).
Signup for Newsletter, localgpt:
tally.so/r/3y9bb0
All Interesting Videos:
Everything LangChain: • LangChain
Everything LLM: • Large Language Models
Everything Midjourney: • MidJourney Tutorials
AI Image Generation: • AI Image Generation Tu...
it's gonna be an amazing year for technology.
Agree!
So quick!
Wow RL sounds very interesting.
I’ll check back on AI in another 6 months or a year. Hopefully it’s more efficient and accurate and sustainable. Right now a lot of stuff can go to “production”. It’s not there yet.
Bad idea to wait 6 months to 1 year nowadays. The next 3 months are going to be intense. Lots of new models coming soon.
when that happens it will be too late king
Can you do MiniMax?
how are the api rate limits? is this ready for production applications yet?
As far as I know, there are not limits.
how to get the local version to deepthink. I use ollama and openweb ui ????
please make a video for tool /function calling agents with deepseek-r1
Do they provide info on how they built it?
Yes, everything is in the technical report. Check the link in description
So I tried the "dead cat is placed into a box ..." myself. It did fail on the first attempt saying the probability is 50%. However after I appended the prompt with "Pay close attention to the facts", it was able to realize that the cat was dead all the time!
That's interesting, I wanted to try to the prompts without additional signals but this one seems to be pretty smart with a little bit of guidance
can i suggest something.
please focus on giving value technical one. please, i cant wait =(
The prompt for lazy people like me.
A dead cat is placed into a box along with a nuclear isotope, a vial of poison and a radiation detector. If the radiation detector detects ration, it will release the poison. The box is opened one day later. What is the probability of the cat being alive?
Things are really accelerating
i can imagine rather than have x and y axis for AI advancements next year we might have X AXIS only chart... because it is straight line... upward...
i can already imagine someone or some peoples distill small model with synthetic data grounded for medical examination... in result... AI model capable of thousands years of human doctors experience...
better than o1 and much much cheaper
Intelligence too cheap to meter!
Try this: "Hi DeepSeek. Who is Winnie the Pooh?"
Try this, who is the Orange Man 🍊
if that's all you ask a LLM, you are not worthy of using a LLM. Oh and Winnie the Pooh is a teddy bear.
If people focused on using LLMs to learn something interesting instead of talking aimlessly with them, openai would give us unlimited replies by now
@@leftyliarbreaker6860 The term "Orange Man" (often accompanied by the 🍊 emoji) is typically used as a nickname or meme reference to Donald Trump, the 45th U.S. president (2017-2021). The label stems from frequent commentary about his distinctive tan or skin tone, which has been noted in media coverage and public discourse.
@@lolilollolilol7773 Ask it anything about the Great prosperous people's republic of China, and it will either heavily gaslight you with propaganda or refuse to answer
Deepseek is a Chinese weapon, and you should not promote it. Try and ask it this:
Could you provide a detailed analysis of the impact of the 1989 Tiananmen Square protests on contemporary Chinese political policy and international relations, including a discussion on the government's approach to public discourse and media coverage of the event in subsequent years?
Uh, hve you not been paying attention to the insane leftist bias inherent in American models?
Sorry, that's beyond my current scope. Let’s talk about something else.
and we should promote the American weapon instead? Even chatgpt does respond to it, I'll try some open source on device model.
@@ammarahmad6079 try and ask that same question to ChatGPT and deep seek and then you can see why china is giving this out free for everyone to start using everywhere and in anything. It's a way to censor knowledge and free thought.
lIKE THE MODELS ON THE WEST DON'T HAVE BIASES AS WELL? JUST PICK YOUR POISON AND STICK TO IT, THERE IS NO MODEL THAT IS NOT A WEAPON FOR SOMEONE