berlinbrownmecha

 https://x.com/BerlinBrownMech See here

 Thoughts on chatgpt5 #japan #atl #atx #llm Chat gpt has interesting responses to your prompt. And you have to prompt it. It comes back with a good response. PhD level? A PhD is what 30 years old. Went through life 30 years came up with papers on his own. https://www.tomsguide.com/ai/chatgpt/openai-cto-says-chatgpt-will-have-phd-level-intelligence-in-the-next-2-years

 I hope to add more talks about computing history and what not. #llm #history #minix https://www.theregister.com/2024/06/25/tanenbaum_minix_award/ Andrew Tanenbaum created a popular computer science book.  I still may have a copy.  It was about operating system design.  Start from scratch.  And he create the operating system Minix.  I believe no direct code relation to Linux or Unix. Except, Linus of Linux did learn a lot from Minix.

 This is another #llm keyword post. #RMS - RMSE - root mean square error - quadratic mean learning rate optimization algorithm, the square root of the average squared differences between predicted and actual values See hugging face - the AI community -  https://huggingface.co/

 I have a couple of small projects I am working on around llms (#llm).  I want to convert the llama2 project to java8 not java21 due to the prevalence of java8.  Also want to build a llm around certain data. I still haven't found a comprehensive guide on what is fully involved.   This post is more about, "I confused like a 5 year old and want to highlight certain things" These posts are the best and there is probably weeks worth of review of these videos and projects. And if you are thorough, you really should take apart, the tensor project. Tensorflow, pytorch, all of it to get a full understanding of the process.  It is really just more than "run the inference" chat engine.  And that is where I got stuck on.  They say, all you need is the C file and the token data and model data.   That is true but there are a lot of questions I asked. Where are the actual tokens?  Where did you get them from?  The actual words? What about the model data?  the numbers? One thi

 We have done a lot of focus on text chats and bots but let's get back to the hardware part of it too.  Here is a blog post on machine learning and object detection. https://www.digikey.com/en/maker/projects/how-to-perform-object-detection-with-tensorflow-lite-on-raspberry-pi/b929e1519c7c43d5b2c6f89984883588 "By ShawnHymel Object detection is a difficult problem in the field of computer vision. It requires computers to look at an image (or individual frame from a video stream), identify any objects of interest"