diff --git a/index.html b/index.html index 73d6357..596a465 100644 --- a/index.html +++ b/index.html @@ -299,7 +299,7 @@

Approach

We introduce SIMPLER, a suite of open-source simulated evaluation environments for common real robot manipulation setups. All environments expose a uniform Gym API. Additionally, we open-source all policy inference code (e.g., RT-1, RT-1-X, Octo) for real-to-sim evaluation, and we provide a detailed guide for evaluating new policies and creating new evaluation - environments. + environments. All environments can be imported with a single line of code and can be interacted with through a standard Gym interface.

Metrics for Real-to-Sim Evaluation

@@ -336,30 +336,36 @@

Visual Matching to mitigate Real-to-Sim Visual Gap

-

Experiment

+

Experiments

-

Long-Horizon Spatial Goal Navigation from Language

+

Evaluating Policies

- With open-vocabulary landmark indexing, VLMaps enables long-horizon spatial goal navigation with - natural language instructions + Our approach yields a strong correlation between real-world and simulated performance for various open-source robot policies, + across two commonly used robot embodiments (Google Robot and WidowX) and over ∼1500 evaluation episodes.

Sequence 1

VLMaps

CLIP on Wheels

+
+ +

Octo-Small stack green block on yellow block

+
@@ -367,16 +373,22 @@

Long-Horizon Spatial Goal Navigation from Language

Sequence 2

VLMaps

CLIP on Wheels

+
+ +

RT1 move redbull can near orange

+