@willfrancis24: Run DeepSeek completely privately and offline using LM Studio, which is totally free and easy to use, even if you're not a tech geek. Get the deep reasoning, writing, maths and coding skills of the latest AI tool to hit the headlines, but on your own PC or Mac, thanks to the LLM being open source (and LM Studio, which is itself open source). #ai #techtok #artificialintelligence #chatgpt #learnontikok #llm #Tech #genai #aitools #aiapps #opensource #deepseek
Will Francis - AI + Marketing
Region: GB
Wednesday 29 January 2025 22:53:02 GMT
Music
Download
Comments
Thanos Sofroniou :
Am I the only one who doesn't care about this data thing?
2025-01-29 23:47:19
518
Kisshot :
Fun thing is that even openAI is logged and heavily censored
2025-01-30 13:16:24
107
M1KE2345 :
I don’t think running locally bypasses the censorship. I ran the local DeepSeek R1 7b model through ollama and it still wouldn’t tell me anything about what happened on that day in June 1989.
2025-01-29 23:11:07
12
Lumian :
you can even use it with ollama if you have the hardware for it
2025-01-30 12:17:11
2
Uros S :
competition worries about ds privacy 😂
2025-01-30 15:54:36
9
Gökbörü🐺 :
Then u need a super expensive PC to get an inferior performance, ain’t it right?
2025-01-30 08:30:06
5
HowToBJöRN ™ :
I use the distilled with Llama, 8B, it’s 6GB
2025-01-30 11:28:37
4
ᅟ :
i just asked this question via Qwen & surprisingly Deepseek turned out it got better score according to Qwen 2.5 response
2025-01-30 17:51:50
1
Michael Scott :
depends if the censorship is baked into the training... and you can hardly run the 700B param model locally
2025-01-30 00:18:58
7
Donniwaha :
"Will my data go to the USA" never bothered anyone, apparently.
2025-01-30 16:38:56
11
Klvn :
I’m using it now it’s so user friendly 😭 especially I have no idea about coding. And yeah offline performance is not that good…
2025-01-30 06:36:33
3
ThePuddleMaker™ :
it takes ages to get an answer offline... with a normal high-end GPU, the 20GB version is the most performant one you can run. r1 is 404gb 😐= 100GPUs
2025-01-30 06:14:08
2
WTF :
I can't set up an account..... the send code doesn't send a code to my email..... any ideas as to why please?
2025-01-29 23:46:12
2
Ellen :
Thank you so much.
2025-01-29 22:59:16
3
Beyond Da Meaning :
Microsoft just adds deepseek on azure
2025-01-30 05:08:44
7
chicken emas :
no one talks about who's hacking deepseek?
2025-01-30 15:35:57
7
Onlyユザー⭐🌟 :
ChatGPT can't draw copyright characters, If this can, I am switching.
2025-01-30 13:07:42
2
⠀ ⠀ ⠀ ⠀ ⠀ ⠀ ⠀ ⠀ ⠀ ⠀crox :
Can LM studio Access my Home wifi Internet?
2025-01-30 13:25:19
0
5T3W13_G :
Run it with webUI you can utilize file upload or even make it search (you just need an API from google to use google search
2025-01-30 14:27:13
2
nfbdhejehe :
I did this today but it was so slowww!! I don’t want to wait 5 minutes for every prompt. Have I done something wrong or is my computer slow
2025-01-30 06:22:27
1
Jakob :
The distil models actually don’t have the deepseek r1 architectures but use the llama or qwen architecture. So it’s not the same
2025-01-30 13:05:00
3
Steeve 🇪🇺 :
One thing that I like about chat gpt at the moment is the memory part. It remembers past conversations, context etc. Can I do this with deep seek on my machine? (Without sending my data to anyone? )
2025-01-30 13:26:08
2
babak :
not people on earth, only us ,and we dont care, the world isn't only us,us empire is over believe or not
2025-01-30 11:41:51
2
Trisha :
Aha! Now I know how to do this - thanks for your vid.
2025-01-30 13:06:56
1
LeatherBass 🏴🇮🇪 :
Same. Facebook, Amazon & eBay know much more about me than is healthy
2025-01-30 12:54:08
1
To see more videos from user @willfrancis24, please go to the Tikwm
homepage.