From 01beb915db9b322bdf110b360cd6638fe39b00c9 Mon Sep 17 00:00:00 2001 From: Karthik Dharmarajan Date: Mon, 29 Apr 2024 01:21:05 -0700 Subject: [PATCH] Added README --- README.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/README.md b/README.md index 2760463..23c26ec 100644 --- a/README.md +++ b/README.md @@ -69,7 +69,7 @@ source mirage/mirage/ros_ws/install/setup.bash cd mirage/mirage/benchmark/robosuite python3 run_robosuite_benchmark.py --config config/example_config.yaml ``` -Please take a look at the example_config and the different parameters that can be set to run different tasks, agents, and robots. For the above code to work, you must change the agents to the path for the model checkpoints in robosuite. +Please take a look at the example_config and the different parameters that can be set to run different tasks, agents, and robots. For the above code to work, you must change the agents to the path for the model checkpoints in robosuite. We have provided the sample models used for evaluation in `mirage/mirage/models/{task}/{state_input}/{model_name}.pt`. The tasks are can, lift, square, stack, and two piece. The state inputs could be image_no_proprio (RGB observation only), image_proprio (RGB + proprio state), and low_dim (proprio + manipulated object position state). ## Citation If you utilized the benchmark, please consider citing the paper: