Sublime

QwQ-32b outperforms 20x bigger Model DeepSeek-r1! That's completely crazy! DeepSeek-r1 was released in January and shocked the world. For the first time there was an OpenSource version of a model at OpenAI's o1 level. And now, only two months later (!) we have a new OpenSource model with the... See more

Build a LLM app with RAG to chat with PDF files using GPT-4 o mini in just 4 lines of Python code:

Perplexity CEO tells the story of how OpenAi Cofounder, Ilya Sutskever planned on creating AGI: https://t.co/A3YKhEI7Uj

Build a LLM app with RAG to chat with PDF files using the latest open-source Llama 3.1 model in just 4 lines of Python code:

.@ilyasut full talk at neurips 2024 "pre-training as we know it will end" and what comes next is superintelligence: agentic, reasons, understands and is self aware https://t.co/CUyQN5zA47

Interview with Geoff Hinton https://t.co/hU6wKVndkj

@giffmana Great read! You may find this blog from @ppwwyyxx interesting. It goes into details of different resize/resample options in CV libraries and their consequences. This understanding helped us a lot back in the days: https://t.co/w8UVA7S6yz

My research career: 2014: coded sparse recursive GP 2015: coded SMC^2 for infinite HMM 2015/6: co-invented MuProp/Gumbel-Softmax 2016: deep Q net on real robot! ... (CoRL best paper 🔥) ... 2022: pretrained Wikipedia to solve halfcheetah 2022: added... See more