The “Nvidia & Langchain AI Agents contest” concluded on Tuesday at 2 in the morning, and I’m proud that I crossed the finish line 3 hours early with a little something: Introducing 𝑨𝒊𝒍𝒆𝒆𝒏 𝟐: an AI office agent and my entry into the ‘NVIDIA and LangChain Generative AI Agents Developer Contest’ in the Small Language Model category. AI agent... Read more 14 May 2024 - 1 minute read
Ethan Mollick remarks that “people realize how capable mutlimodal AI is, right now, out of the box”. I agree: For those who want to try and compare different models, there is https://huggingface.co/spaces/WildVision/vision-arena. Besides the usual suspects, I recommend trying Reka AI Flash. Read more 07 May 2024 - less than 1 minute read
A lot of background on the practicalities from one of the Sora private beta-testing teams: https://www.fxguide.com/fxfeatured/actually-using-sora/ „Air Head still needed a large amount of editorial and human direction to produce this engaging and funny story film.“ This included using Adobe After Effects. N.B.: The article states that Sora is ... Read more 25 Apr 2024 - less than 1 minute read
Urchade Zaratiana has released GLiNER, an NER model created “that can identify any type of entity using a bidirectional transformer encoder”. One of the ToDos listed is „Allow longer context“, and a closer look reveals that: The Colab caps out at 384 tokens, and confirms DebertaV2 architecture. Interesting, but too restricted as of now. Read more 25 Apr 2024 - less than 1 minute read
How does one record a live stream in absence, perhaps using AWS EC2? A discussion on Reddit lead me to base everything on a g5.xlarge instance. Problems: ⚒️ On the Windows Server, there initially was no soundcard. Fixed by simply installing Virtual Audio Cable (VAC)? Fixed by running Microsoft Teams in an OBS Source of type “Browser”: this will... Read more 25 Sep 2024 (Updated) - 2 minute read