issues
search
caio-freitas
/
GraphDiffusionImitate
Diffusion-based graph generative policies for imitation learning in robotics tasks 🧠🤖
MIT License
1
stars
1
forks
source link
Fix and Use EMA in Diffusion Policy
#54
Open
caio-freitas
opened
11 months ago
caio-freitas
commented
11 months ago
Find out why it only works with pred_horizon > obs_horizon + action_horizon
Use EMA (Exponential Moving Average) model in training