From 78ca6fcbc8e207a191ec8d7cf08b6b72f84e4c0b Mon Sep 17 00:00:00 2001 From: Benyahia Mohammed Oussama <mohammed.benyahia@etu.ec-lyon.fr> Date: Sat, 22 Feb 2025 13:30:24 +0000 Subject: [PATCH] Edit README.md --- README.md | 19 ++++++++++++++----- 1 file changed, 14 insertions(+), 5 deletions(-) diff --git a/README.md b/README.md index a2aed87..997609e 100644 --- a/README.md +++ b/README.md @@ -107,11 +107,20 @@ This repository contains my individual work for the **Hands-On Reinforcement Lea [A2C Panda-Reach Model](https://huggingface.co/oussamab2n/a2c-panda-reach) ### Evaluation -- **Evaluation Process:** - - The model was evaluated over **100 episodes**. - - An episode is considered **successful** if it reaches a total reward of **-0.25**. -- **Evaluation Results:** - - **97%** of episodes successfully met the reward threshold, indicating strong performance. + + +- **Evaluation Process:** + +- The model was evaluated over 100 episodes. +- An episode is considered successful if it reaches a total reward of -0.25. + +- **Updated Evaluation Results:** + +- **Total episodes with truncation:** 99/100 +- **Average reward at truncation:** -7.68 +- **Percentage of episodes meeting the reward threshold:** 97%, indicating strong performance. + + - **Evaluation Plot:**  *(Figure: The robotic arm’s performance on the PandaReachJointsDense-v3 environment.)* -- GitLab