Downloads · 30 days
0
Legend2727/evidence_grounded_taxonomy_eval_v3
evidence_grounded_taxonomy_eval_v3 is a machine learning model from Legend2727. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as mit.
This directory is the local release bundle for the final model artifact in this repository.
Downloads · 30 days
0
Access
Public
Updated Apr 6, 2026
Repo size
1.5 GB
Likes
0
Public
Click a slice to open those files.
.pt1.5 GB · 100%
From the Hugging Face model README
This directory is the local release bundle for the final model artifact in this repository.
Release status:
https://huggingface.co/Legend2727/evidence_grounded_taxonomy_eval_v3Tracked metadata files in this directory:
README.mdartifact_manifest.jsonmetrics.jsonlabel_map.jsonbest_thresholds.jsonbest_thresholds_by_label.jsonLocal-only large file:
best_model.ptbest_model.pt is intentionally ignored by git. During cleanup, the local checkpoint was copied here from the older root import folder so this directory can serve as the upload-ready bundle.
Example live inference command:
python scripts/infer_evidence_grounded_taxonomy.py \
--image-root /path/to/image_root \
--checkpoint models/released/evidence_grounded_taxonomy_eval_v3/best_model.pt \
--thresholds models/released/evidence_grounded_taxonomy_eval_v3/best_thresholds.json \
--source-path images/source/shard_00/example.png \
--target-path images/target/shard_00/example.png \
--instruction "Remove the object from the left side of the image." \
--lang en
This bundle has been published to:
https://huggingface.co/Legend2727/evidence_grounded_taxonomy_eval_v3Refresh workflow:
huggingface-cli login
huggingface-cli upload-large-folder Legend2727/evidence_grounded_taxonomy_eval_v3 models/released/evidence_grounded_taxonomy_eval_v3 --repo-type model
Tracked evaluation/report files also live under:
reports/evidence_grounded/v3/That report bundle contains:
artifact_manifest.jsonmetrics.jsonlabel_map.jsonbest_thresholds.jsonbest_thresholds_by_label.jsonper_class_test_benchmark.csvtrain_history_v3.jsonIf the original local import folder is still available, it may also contain additional raw exports such as prediction JSONL files that were intentionally not migrated into normal git tracking.
The repo now includes the recovered lineage notebook and wrapper scripts for the evidence-grounded path. What is still not claimed is a fresh end-to-end retraining verification inside this cleaned repo.
A one-click gallery for the branch-coverage examples is available at: