Instructions to use NoeFlandre/osm-polygon-sentence-classifier with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use NoeFlandre/osm-polygon-sentence-classifier with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-classification", model="NoeFlandre/osm-polygon-sentence-classifier")# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("NoeFlandre/osm-polygon-sentence-classifier", device_map="auto") - Notebooks
- Google Colab
- Kaggle
OSM Polygon Sentence Classifier
This public repository contains organized, immutable outputs for the OSM polygon sentence-classification studies.
Repository layout
- Final model:
studies/place-relevance-v2-ablations/a04-head-lr1e-3/run-a4d7a6500ac2c19fb18d/final/(load with the Transformerssubfolderargument). - Complete checkpoints:
studies/place-relevance-v2-ablations/a04-head-lr1e-3/run-a4d7a6500ac2c19fb18d/checkpoints/step-N/. - Each run has its own experiment and run directory; no model files are stored at the repository root.
Study registry
Completed landuse ablations:
studies/landuse-v1/README.md.Protocol:
studies/landuse-v1/study.json.Results:
studies/landuse-v1/results.json.Worldwide V2 baseline:
studies/place-relevance-v2/README.md.Protocol:
studies/place-relevance-v2/study.json.Results:
studies/place-relevance-v2/results.json.Worldwide V2 ablations:
studies/place-relevance-v2-ablations/README.md.Ablation protocol:
studies/place-relevance-v2-ablations/study.json.Ablation results:
studies/place-relevance-v2-ablations/results.json.
Experiment tracking
- V1 landuse: NoeFlandre/osm-polygon-sentence-classifier-trackio.
Static data bucket:
NoeFlandre/osm-polygon-sentence-classifier-trackio-data. - V2 place relevance: NoeFlandre/osm-polygon-sentence-classifier-v2-trackio.
Static data bucket:
NoeFlandre/osm-polygon-sentence-classifier-v2-trackio-data.
Training identity
- Task:
place-relevance-v2 - Task description: place relevance
- Base model:
jhu-clsp/mmBERT-small - Base-model revision:
abc32620dd4f6ab06f5fbe905dc25f310618e09f - Run directory:
studies/place-relevance-v2-ablations/a04-head-lr1e-3/run-a4d7a6500ac2c19fb18d - Current run Trackio dashboard: NoeFlandre/osm-polygon-sentence-classifier-v2-trackio
The generated model card inside the final directory contains the recorded configuration and evaluation metrics.