DeepSeek AI Models
This cluster centers on discussions about DeepSeek AI models, including their performance, comparisons to OpenAI and Llama, training methods like distillation, open-source status, cost-effectiveness, and specific implementations.
Activity Over Time
Top Contributors
Keywords
Sample Comments
DeepSeek is only thing because they used OpenAI for refinement.
Does anyone know how Deepseek does it yet?
What's specific to deepseek here that other models do not use, or are you just riding the keyword wave?
DeepSeek is based on llama, right? that 3% is only the fine tuning money?
Did deepseek have a notably dofferent tenor?
Why is DeepSeek specifically called out?
What is the comparison of this versus DeepSeek in terms of good results and cost?
what flavor of deepseek are you running? what kind of performance are you seeing?
Interesting take on deepseek. I think deepseek may be a breath of fresh air slightly leveling the playing field for small startups and individual developers, Assuming the claimed reduced training cost are real.
Wasn't it shown that DeepSeek was distilling OpenAI's models? Seems like a presumptuous claim by Liang Wenfeng