Skip to content
New issue

Sprint 1

Past due by about 1 year 100% complete
  1. [email protected] - Friends fan - Prepare datasets for Fine Tuning and Evaluation
    a. Fetch data
    b. Preprocessing, etc
    c. Questions with answers + 3 options
  2. Advaith Shyamsunder Rao- Pediatrician - Prepare datasets for Fine Tuning and Evaluation
  3. [email protected] Host base model on Ilabs. (GGML/GGUF for CPU+GPU inference)
  4. Advaith Sh…
  1. [email protected] - Friends fan - Prepare datasets for Fine Tuning and Evaluation
    a. Fetch data
    b. Preprocessing, etc
    c. Questions with answers + 3 options
  2. Advaith Shyamsunder Rao- Pediatrician - Prepare datasets for Fine Tuning and Evaluation
  3. [email protected] Host base model on Ilabs. (GGML/GGUF for CPU+GPU inference)
  4. Advaith Shyamsunder Rao Create API for inference of base model + Port forwarding to colab notebook
  5. [email protected] Evaluation code - This should cover a function that does the following,
    a. Study prompt engg, langchain
    b. Prompt engineering -> Few shot for answer formatting
    c. Measure evaluation 2: Prompt engg vs non-prompt (both w/ fine-tuned model) -> 4.c) Advaith Shyamsunder Rao
    d. Get prediction for the LLM on the dataset provided
    e. Return accuracy
    [email protected] put the base LLM against the dataset and see how well it does.
    a. If accuracy < 40%, then we can go ahead with Finetuning the model for our persona.
    b. Elif try different model [email protected]; task 3 repeats
    c. Else: drop persona: task 1 repeats

There are no open issues in this milestone.

Add issues to milestones to help organize your work for a particular release or project.

Create new issue

Or find and add issues with no milestone in this repo.