Navila Satnav Continue 1ep 8f Sample Hk7 Fs7 Stopx4

--- license: llama3 language: - en library_name: transformers pipeline_tag: video-text-to-text base_model: a8cheng/navila-llama3-8b-8f base_model_relation: finetune tags: - navila - satnav - vision-language-navigation - embodied-ai - robotics --- # NaVILA for SatNav — Navigation-pretrained 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-llama3-8b-8f` - 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 - `continue`: navigation-pretrained initialization from the NaVILA navigation SFT checkpoint - `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 (navigation-pretrained)**. 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`) | 25.0 | 24.9 | 35.0 | 93.88 | 51.47 | | Test Unseen (`val_unseen`) | 18.6 | 18.4 | 31.6 | 123.49 | 59.57 | ## Usage Use this checkpoint with the [SatNav NaVILA integration](https://github.com/Eku127/SatNav/blob/master/docs/en-US/training/vlm/NAVILA.md). It is also listed in the [SatNav Baseline Model Zoo](https://huggingface.co/collections/Eku127/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.

--- license: llama3 language: - en library_name: transformers pipeline_tag: video-text-to-text base_model: a8cheng/navila-llama3-8b-8f base_model_relation: finetune tags: - navila - satnav - vision-language-navigation - embodied-ai - robotics --- # NaVILA SatNav Continue 1ep 8f sample-hk7-fs7-stopx4 This repository contains a Hugging Face upload-ready copy of the NaVILA SatNav continue baseline `navila-satnav-continue-1ep-8f-sample-hk7-fs7-stopx4`. ## Model Details - Model type: NaVILA-style vision-language-navigation policy. - Training mode: SatNav fine-tuning from `a8cheng/navila-llama3-8b-8f`. - Language backbone family: Llama 3. - Visual backbone family: SigLIP, packaged under the NaVILA model layout. - Local archive source: `output/model_zoo/baseline/navila-continue0404-r2-20260427-141143-sample-hk7-fs7-stopx4`. - Sampling setting: `sample-hk7-fs7-stopx4`. ## Repository Contents This release directory keeps only NaVILA inference/evaluation artifacts: - `config.json` - `llm/` - `mm_projector/` - `vision_tower/` - model card, license, and notice files Training logs, trainer state, DeepSpeed state, RNG state, and duplicated `checkpoint-*` directories are intentionally omitted. ## Training Data and Procedure The model was fine-tuned for 1 epoch on SatNav trajectory data using the NaVILA SatNav training pipeline. The model card does not redistribute SatNav episodes, images, simulator scenes, or training logs. ## License This release is marked with the Hugging Face `llama3` license tag because the released weights contain a Llama 3 language backbone. NaVILA code and other third-party components remain under their respective upstream licenses and terms.

Source context: 34 downloads · 0 likes · Pipeline video-text-to-text · Library transformers · Repo Eku127/navila-satnav-continue-1ep-8f-sample-hk7-fs7-stopx4