Got good results; from RSME 52 to 32 a goof 39 % improvement in failure prediction . but trying few more things to get it more in range . let me knnow if anything worked other folks workign on C-MAPSS data. #DataEngineering#IndustrialAI#MachineLearning
Been working on a small project .. plotted all graphs on the C-MAPSS jet engine dataset .. many graphs were straight lines . I was like how can jet engine data not change with time .. learn more on on my linkedin https://t.co/LadBG1H2DF
started going through Activation functions.. Reading on all basics and understanding what these functions are useful for.. Got a good understanding on them Happy learning #learn#ML
@santtiagom_ my approach so far has been to build what I really want to ...like a Twitter bot which will post on my behalf and use both local LLM model posts regularly .. The good part is you learn a lot doing this.. ππ
So I am working on using langchain and online embedding models to try and get better results .. if anyone has worked on such projects how have you approached this issue with Bad PDF data? πππ’π’
Working on a Different approch for my RAG project .. I guess trial and error is the way to go.. Simply reading pdf creating chunks and embeddings is not good .. and reading and trying to clean the pdf is also a headache specially when you cant properly read the PDf's
starting to re structure fre thing in the RAG project I am working on. Model is getting data from pdf but random jibberish, also for all the user input I don't want to get vector embeddings as a response looking for a better approach. #RAG#AI#LLMs
it's always good to share some new learnings, Completed a course on vector dbs. Learned a lot on fundamentals of vector dbs; did hands-on coding for same, understood langchain framework as well. Overall a really productive course
Been busy with some office work. I am quiet surprised why don't we we have sample templates people can modify to create different applications or this are out there and I am just oblivious ..!!
"AI agents are revolutionizing productivity, turning complex tasks into seamless workflows. ππ€ I am starting my learnings and projects towards this new trend. let me know if anyone has some good starting points #ArtificialIntelligence#AIAgents#FutureOfWork"
I am currently working on a RAG project. Having some difficulty with setting up the llm model fine tuning it.
The idea is creating a backend fast api to expose the llm model and using it and get response. #development#buildinginpublic
Finally back from a Break. Got married and now starting with a new RAG Project. Starting new year with a lot changes and new determination!!!πππ‘π‘
Just Finsished initial portion of data Engineering project I am working on. Fetching data from API and storing in correct format to postgrace table. Will be adding some changes to handle SCD and also try to add some Machine Learning to predict weather as well. πππ
Of course, initial tweet was not good I was working on adjustments. I have some obvious features I can add Developement has never been this much fun with AI tools to help. @cursor_ai@amanrsanger And team great product.
Added some code to create different tweets. Will be working on scheduling this bot and making that profile look better. Have more ideas on how to make this bot more engaging. I want to see this account get 1k followers someday. #Progress#LearnInPublic