Eku127/swiftvln-satnav-3b-1ep-f32s4-overlap0-pf-h8-pool-s2-initial-noembed

SwiftVLN for SatNav — Initial Observation

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 augments the reference visual memory with the episode's initial observation.

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, the initial observation, and up to 8 history frames
• —Prediction horizon: 4 actions
• —Memory: per-frame average pooling with stride 2

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
• —pf-h8-pool-s2: uses the reference per-frame memory configuration
• —initial: inserts the episode's initial observation into the system prompt
• —noembed: uses no embedding enhancement

Reference Results

Results reported in the SatNav paper, memory-design ablation, row Initial observation.

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 an input-augmentation ablation in the SwiftVLN SatNav Ablation Model Zoo.