From 54d8b56939460d5ad05450cc5ab54711e783e62d Mon Sep 17 00:00:00 2001
From: Joshua David <jshuadvd@icloud.com>
Date: Fri, 19 Jul 2024 21:44:13 -0700
Subject: [PATCH] Update README.md to be more detailed

---
 README.md | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/README.md b/README.md
index 4390c4f..ff584b4 100644
--- a/README.md
+++ b/README.md
@@ -204,6 +204,10 @@ To train on a custom dataset:
 Hyperparameter Tuning
 LongRoPE's performance can be sensitive to hyperparameters. Key parameters to tune include:
 
+`population_size`, `num_mutations`, and `num_crossovers` in the lambda factor search
+Learning rate and scheduler parameters for fine-tuning
+gradient_accumulation_steps for training stability
+
 - **Base Length**: The initial context length used for fine-tuning.
 - **Target Length**: The final context length to extend to.
 - **Population Size**: The number of candidate solutions to search for during the search algorithm.