Eku127/navila-satnav-scratch-1ep-8f-sample-hk7-fs7-stopx4

NaVILA for SatNav — Base initialization

This checkpoint is designed for use with SatNav, the continuous-state vision-and-language navigation platform for satellite maps. The SatNav NaVILA integration uses it to predict compact actions from a language instruction and eight RGB observations.

Model

• Starting checkpoint: a8cheng/navila-siglip-llama3-8b-v1.5-pretrain
• Training data: SatNav-v0.1 offline trajectories
• Training: 1 epoch, full-parameter fine-tuning
• Action format: stop, forward, left, and right

Repository name

• navila-satnav: NaVILA adapted to SatNav
• scratch: base initialization from the NaVILA pretraining checkpoint, before navigation SFT
• 1ep: trained for one epoch
• 8f: uses eight visual frames per sample
• sample: the following fields describe trajectory-step sampling
• hk7: keeps the first seven action steps
• fs7: keeps every seventh step in a consecutive forward run after the head; turning steps are kept
• stopx4: repeats each terminal stop sample four times

Reference Results

Results reported in the SatNav paper, baseline comparison, row NaVILA.

SR, SPL, and OS are percentages. NE is measured in meters. Steps is the average number of executed environment actions, including turns and stop.

Split	SR (%)	SPL (%)	OS (%)	NE (m)	Steps
Test Seen (val_seen)	18.1	18.0	27.6	93.05	48.08
Test Unseen (val_unseen)	13.0	12.7	23.7	128.88	53.71

Usage

Use this checkpoint with the SatNav NaVILA integration. It is also listed in the SatNav Baseline Model Zoo.

License

The checkpoint contains a Llama 3 backbone and is released under the applicable Llama 3 terms. NaVILA, SatNav, and other components retain their own licenses.