Do you have experience fine tuning GPS OSS models?
Hi I found this space called Affine. It is a daily reinforcement learning competition and I'm participating in it. One thing that I am looking for collaboration on is with fine tuning GPT OSS models to score well on the evaluations. I am wondering if anyone here is interested in mining? I feel that people here would have some good reinforcement learning tricks. These models are evaluated on a set of RL-environments with validators looking for the model which dominates the Pareto frontier. I'm specifically looking to see any improvements in the coding deduction environment and the new ELR environment they made. I would like to use a GPT OSS model here but its hard to fine-tune these models in GRPO. Here is the information I found on Affine: https://www.reddit.com/r/reinforcementlearning/comments/1mnq6i0/comment/n86sjrk/?utm_source=share&utm_medium=web3x&utm_name=web3xcss&utm_term=1&utm_content=share_button
Using Neural Network to Learn Profitable Trading in the FOREX Markets
I am using Neural Networks (NN) to teach them how to recognize profitable trading opportunities in the Foreign Exchange (FOREX) markets, using 10 currencies simultaneously. I am using 3rd-order Cubic Splines as input to give the NNs a sense of how the critical variables change over time. I am using free FOREX historical trading data to train the NNs how to trade profitably in the future. I don't just feed the trading levels of the FOREX currency pairs as input to the NNs. Instead, I use a variation of the computed DXY Index for all 10 currencies in order to isolate the value change of each of the individual currencies, using Cubic Splines to detail how those values change over various time periods. The end result is Neural Networks that recognize which currencies to Buy and which ones to Sell at the most profitable times. If anyone is interested in the details, please reach out and I will provide more details.
Personal AI
How could Microsoft get it so wrong. I tried to use Azure and AI and its just matrix calculations. Open-AI is like a brute force training method, but at least its visualized. I just want a dog, a pet cat. An annoying robot that helps me with math and does my homework and that i can play games with. That will learn and grow with me and all you have is just a bunch of worthless databases like your CD collection, or vintage games. Bunch of comic book nerds with their libraries of books. Its like walking into a library and you don't know how to read. Here i was thinking that paperclip from word would have progressed into an annoying helper, but at least i have dancing girls hows **bleep** i can vibrate. Seriously can you not just make something simple, someone somewhere in that huge office must be someone who knows where the AI program is and can make a simple program and is like a "ferbe". Thats as helpful as the paperclip with a comprehensive help toolkit. Something that works like an operating system of something robotic i want to build. Could someone please help me make something simple, that can do basic reading comprehension so i can at least talk to it, and get it to read a book and answer questions. Maybe one that can learn maths and can 3d model things so i can visualize it. Or maybe something that can just learn by watching me and can recall something i need for reference. "i saw something on facebook" (pulls out phone to show person) "asks ai something vague" instantly pulls it up.Tech Minutes Video - Project Trove
This post is Authored by Trinh Duong, Christian Liensberger and Giampaolo Battaglia Office of the CTO Team & AI/Innovation at Microsoft We recently launched the Innovation https://innovation.microsoft.com/en-us/developer, which are short, snackable informative tidbits from Microsoft researchers, developers and engineers all around the world on some of the latest and future technologies. https://innovation.microsoft.com/en-us/tech-minutes-project-trove, Christian Liensberger, Principal Program Manager and Advisor to Microsoft’s CTO shares new insights into Project Trove - a crowdsourcing marketplace where you can gather high-quality images for your AI models. Images are responsibly sourced from regular individuals and adhere to a rigid licensing and privacy framework, resulting in a more responsible data collection platform. In this Tech Minutes video, Christian shares the advantages of Trove, and also provides a walkthrough of Trove Web App from an AI Developer standpoint (selecting the right images for your model training), as well as showing how photo takers can upload their images through the Trove App on Android. https://innovation.microsoft.com/en-us/tech-minutes-project-trove Happy viewing & happy end of year! Trinh, Christian and Giampaolo
Jumping from Google's Teachable Machine to Azure. Help
I've been using Google's Teachable Machine for experiments for months, using two classes of images to train for recognition. I now need to switch the data source to tabular data (TM doesn't support this), and feel as though I've walked into Costco, Home Depot and Walgreens combined, with Azure. I've reviewed libraries of demos at studio.azureml.net and signed up for something else related to Azure, but, beyond uploading data, I have yet to find a way to replicate the workflow and simplicity of setup Teachable Machine offered. Any guidance is appreciated as (now knowing 9 computer programming languages) I'm not keen on learning yet another "ecosystem" over the course of X months.
common pitfall of using data bricks with pandas and not spark
Hi Team, I just want to understand what could be the common pitfall of using Pandas on Databricks instead of Spark. There are certain factors on which we have decided to go with Databricks instead of Azure AI platform (jupyter notebook). Experiment tracking using ML-ops Hyperparameter tuning with Spark trails which helps with parallelization I just wanted to understand that what could possibly go wrong if we train model on Databricks by just using pandas & sklearn . Deployment: we will deploy final model offline, will different env will cause an issue ? Cost: is AI platform supports points mentioned above, experiment tracking & parallel hyperparameter tuning Ease of use other advantage offered by AI platform (ex: automatic hyperparameter tuning) I am new to Azure service, It will be really helpful if you can share detail answer of above points with your preference. (what you would have chosen and why?) Thanks in Advance.
