Chinese AI labs DeepSeek and Moonshot AI reportedly trained on outputs from OpenAI's Astra model for their latest releases.
#AI #OpenAI #MoonshotAI #DeepSeep #GPT6Astra #ChinaAI
Chinese AI labs DeepSeek and Moonshot AI reportedly trained on outputs from OpenAI's Astra model for their latest releases.
#AI #OpenAI #MoonshotAI #DeepSeep #GPT6Astra #ChinaAI
Lots of community PRs have been merged, and some were ported from the 4x recipe!
Speed and responsiveness
- A sho…
— from @MiaAI_lab (https://x.com/MiaAI_lab/status/2115903199207215611)
Lots of community PRs have been merged, and some were ported from the 4x recipe!
Speed and responsiveness
- A sho…
— from @MiaAI_lab (https://x.com/MiaAI_lab/status/2115903199207215611)
DeepSeep v4.1 Flash for 4x DGX Sparks is getting... ridicules 🤯
Decode improvements
- Prose decode is now 87 tok/s single stream
- On 4 streams it's 163.6 tok/s
- Code decode is 124.8 tok/s…
— from @MiaAI_lab (https://x.com/MiaAI_lab/status/2103511749114687702)
DeepSeep v4.1 Flash for 4x DGX Sparks is getting... ridicules 🤯
Decode improvements
- Prose decode is now 87 tok/s single stream
- On 4 streams it's 163.6 tok/s
- Code decode is 124.8 tok/s…
— from @MiaAI_lab (https://x.com/MiaAI_lab/status/2103511749114687702)
We took MIR, '20-'24 to test open LLMs. Llama 3.1 based models, like Aloe, reach +80 in accuracy.
Deepseep R1 reaches +88. Boosted by a RAG system, 92.
buff.ly/4bbbXMw
buff.ly/4hLrhBV
We took MIR, '20-'24 to test open LLMs. Llama 3.1 based models, like Aloe, reach +80 in accuracy.
Deepseep R1 reaches +88. Boosted by a RAG system, 92.
buff.ly/4bbbXMw
buff.ly/4hLrhBV
Si j'ai bien compris le bidule DeepSeep pompe dans dautres Ai et adapte sa reponse
Si j'ai bien compris le bidule DeepSeep pompe dans dautres Ai et adapte sa reponse
It is possible... if you use the open source small models!
My AI content creation is at like 10 to 15W----
Like about an LED light if I use it!
It is possible... if you use the open source small models!
My AI content creation is at like 10 to 15W----
Like about an LED light if I use it!
Jan 24
翻了下推特, DeepSeek 让 Meta 和 ChatGPT 的员工陷入了恐慌,一方面高昂的训练成本没法解释了,一方面取消订阅的会员逐日增多,DeepSeep 的价格只是 ChatGPT 的 1/30,许多人认为这一低成本高效果的奇迹,是咱们国家层面在推动...
国家层面推动倒有可能,不过
DeepSeek 让 Meta 和 ChatGPT 的员工陷入了恐慌
你不拿我不拿史密斯专员怎么拿😂
Jan 24
翻了下推特, DeepSeek 让 Meta 和 ChatGPT 的员工陷入了恐慌,一方面高昂的训练成本没法解释了,一方面取消订阅的会员逐日增多,DeepSeep 的价格只是 ChatGPT 的 1/30,许多人认为这一低成本高效果的奇迹,是咱们国家层面在推动...
国家层面推动倒有可能,不过
DeepSeek 让 Meta 和 ChatGPT 的员工陷入了恐慌
你不拿我不拿史密斯专员怎么拿😂
Ma pensa te, pensa te che faccia tosta! Coloro che hanno successo in Silicon Valley sono certi doppiostandardisti (...ma forse in USA in generale) cape di cazzo che... mamma mia, non fatemi dire maleparole,…
Ma pensa te, pensa te che faccia tosta! Coloro che hanno successo in Silicon Valley sono certi doppiostandardisti (...ma forse in USA in generale) cape di cazzo che... mamma mia, non fatemi dire maleparole,…
James Grugett / @jahooma:
DeepSeep V4 Pro will remain 100% free in our coding agent 🎉 Try it now: npm i -g freebuff
James Grugett / @jahooma:
DeepSeep V4 Pro will remain 100% free in our coding agent 🎉 Try it now: npm i -g freebuff
"... DeepSeep..."
"Other companies did not inovate due lost time in woke meetings!"
"'woke' is that thing about a company creating a black doll called 'lawyer' to make more money and white people cries 'white persecussion!' or 'the great replacement is happening!'?"
"... DeepSeep..."
"Other companies did not inovate due lost time in woke meetings!"
"'woke' is that thing about a company creating a black doll called 'lawyer' to make more money and white people cries 'white persecussion!' or 'the great replacement is happening!'?"
DeepSeep R1 dispo chez OVHcloud:
endpoints.ai.cloud.ovh.net/models/a0115...
DeepSeep R1 dispo chez OVHcloud:
endpoints.ai.cloud.ovh.net/models/a0115...
🙃👍
🙃👍
Americká společnost OpenAI znovu obvinila čínský startup zabývající se umělou inteligencí (AI) DeepSeek z neoprávněného využívání svých modelů, a to prostřednictvím techniky zvané destilace. Podle agentury Reuters to vyplývá...
Americká společnost OpenAI znovu obvinila čínský startup zabývající se umělou inteligencí (AI) DeepSeek z neoprávněného využívání svých modelů, a to prostřednictvím techniky zvané destilace. Podle agentury Reuters to vyplývá...
The real price of developing the new deepneeeep models are unknown, however, since a figure cited in a single search card may not understand the full picture of their costs. "I don't believe in 6 months million,…
The real price of developing the new deepneeeep models are unknown, however, since a figure cited in a single search card may not understand the full picture of their costs. "I don't believe in 6 months million,…
Chinan Chatbot Chatbot has moved the Openai's chat of Openai as the most discharge app on Apple Apple and the market is panic. The stocks for the most companies aware as nvidia fell on Monday after the news. Deepseek has released his…
Chinan Chatbot Chatbot has moved the Openai's chat of Openai as the most discharge app on Apple Apple and the market is panic. The stocks for the most companies aware as nvidia fell on Monday after the news. Deepseek has released his…