Datasets:
trying to install with oobabooga on vicuna13b.
NamaskaraH and DanyavadaH for this great dataset!
Hi, I'm trying to install this dataset into oobabooga software for training of a vicuna13b model. I want to teach my bot sanskrit so I can practice with it and gain experience training the bot. If I were able to download the dataset and it is formatted correctly, oobabooga should be able to use it for training, but when I try to download from oobabooga it says I don't have permission. I am logged into huggingface while running oobabooga, but I assume oobabooga is not sending log in credentials to huggingface.
Other things I tried:
I downloaded the csv files via the browser, but oobabooga will not accept them.
I tried running the python script(itihasa.csv) and it ran, but didn't do anything that I noticed.
I even asked chatgpt3 for advice, and it wrote a script for me to try, but that didn't work either.
Any advice would be greatly appreciated.
I don't know what a oobabooga software is. Can you share a link or something?
Here is the link to oobabooga. It's a gui for making and running chat robots and has one click installers for us less technically inclined folks.
So it turns out I ran out of memory. I was able to install on vicuna7b and it worked! The thing speaks sanskrit! At least it tries to speak sanskrit. So thank you sir, for your work on sanskrit and ai. Now I will try to find a dictionary in a reasonable format as well as a grammar book that is well formatted(might be hard). It did read the whittney grammar, but it didn't do as well as reading your itihasa. I'm hoping by getting it grammar and a large corpus of words it can start utilizing conversational sanskrit. Pannini grammar might work for it since it's already organized into small succinct bits.
So much fun, Rahular!