SwiftVLN for SatNav — Segment Token Clustering

Related Repositories

• SwiftVLN: training and evaluation code for these checkpoints.
• SatNav: satellite-image navigation environments, datasets, and evaluation tools.

This checkpoint is designed for SwiftVLN on SatNav, the continuous-state vision-and-language navigation benchmark over satellite imagery. It divides visual history into temporal segments and clusters tokens within each segment.

Model

• Starting checkpoint: Qwen/Qwen2.5-VL-3B-Instruct
• Training data: SatNav-v0.1 offline expert trajectories
• Training: 1 epoch, full-parameter fine-tuning, learning rate 2e-5
• Context: 32 RGB frames plus the available visual history
• Prediction horizon: 4 actions
• Memory: eight temporal segments compressed to 512 total output tokens

Repository name

• swiftvln-satnav: SwiftVLN trained and evaluated on SatNav
• 3b: Qwen2.5-VL 3B backbone
• 1ep: trained for one epoch
• f32s4: uses a 32-frame window and predicts four actions
• overlap0: uses non-overlapping training windows
• sgtc-k512: applies segment token clustering with 512 total output tokens
• noembed: uses no embedding enhancement

Reference Results

Results reported in the SatNav paper, memory-design ablation, row Segment token clustering (STC).

SR, SPL, and OS are percentages. Changes in SR and OS are percentage points relative to the SwiftVLN reference model.

Usage

Use this checkpoint with the SwiftVLN evaluation guide. Keep the full repository name unchanged because SwiftVLN derives the evaluation configuration from it.

This checkpoint is a long-term-memory ablation in the SwiftVLN SatNav Ablation Model Zoo.