This is the model for our paper PIG-Nav: Key Insights for Pretrained Image-Goal Navigation Models.
Description of this model:
- This model (PIG-Nav-NoEarlyfuse) is the model pretrained without early fusing architecture in ViT, where other setups are kept the same as PIG-Nav. The model is trained by a total of 156K iterations with a batch size of 128.
- Please follow our github repo for detailed use.
Inference Providers
NEW
This model isn't deployed by any Inference Provider.
🙋
Ask for provider support