Text Classification
PEFT
Safetensors
English
entity-matching
person-name-matching
record-linkage
deduplication
lora
deberta-v3
Instructions to use LessLM/person-name-match-likelihood-v6 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use LessLM/person-name-match-likelihood-v6 with PEFT:
from peft import PeftModel from transformers import AutoModelForSequenceClassification base_model = AutoModelForSequenceClassification.from_pretrained("MoritzLaurer/DeBERTa-v3-base-mnli-fever-anli") model = PeftModel.from_pretrained(base_model, "LessLM/person-name-match-likelihood-v6") - Notebooks
- Google Colab
- Kaggle
| { | |
| "f1": 0.9682, | |
| "precision": 0.9568, | |
| "recall": 0.9798, | |
| "accuracy": 0.9733, | |
| "confusion_matrix": [ | |
| [ | |
| 1424, | |
| 46 | |
| ], | |
| [ | |
| 21, | |
| 1019 | |
| ] | |
| ], | |
| "total_pairs": 2510, | |
| "latency_mean_ms": 0.4, | |
| "latency_p50_ms": 0.39, | |
| "latency_p95_ms": 0.42, | |
| "latency_p99_ms": 0.42, | |
| "ece": 0.0162, | |
| "per_edge_case": { | |
| "unknown": { | |
| "correct": 665, | |
| "total": 682, | |
| "accuracy": 0.9751 | |
| }, | |
| "partial_overlap": { | |
| "correct": 345, | |
| "total": 353, | |
| "accuracy": 0.9773 | |
| }, | |
| "nickname": { | |
| "correct": 121, | |
| "total": 121, | |
| "accuracy": 1.0 | |
| }, | |
| "name_order": { | |
| "correct": 112, | |
| "total": 112, | |
| "accuracy": 1.0 | |
| }, | |
| "combined": { | |
| "correct": 32, | |
| "total": 32, | |
| "accuracy": 1.0 | |
| }, | |
| "typo": { | |
| "correct": 99, | |
| "total": 117, | |
| "accuracy": 0.8462 | |
| }, | |
| "similar_name": { | |
| "correct": 324, | |
| "total": 341, | |
| "accuracy": 0.9501 | |
| }, | |
| "transliteration": { | |
| "correct": 93, | |
| "total": 93, | |
| "accuracy": 1.0 | |
| }, | |
| "initial": { | |
| "correct": 112, | |
| "total": 112, | |
| "accuracy": 1.0 | |
| }, | |
| "tricky_non_match": { | |
| "correct": 324, | |
| "total": 331, | |
| "accuracy": 0.9789 | |
| }, | |
| "middle_name": { | |
| "correct": 50, | |
| "total": 50, | |
| "accuracy": 1.0 | |
| }, | |
| "title_suffix": { | |
| "correct": 112, | |
| "total": 112, | |
| "accuracy": 1.0 | |
| }, | |
| "hyphenation": { | |
| "correct": 15, | |
| "total": 15, | |
| "accuracy": 1.0 | |
| }, | |
| "case_variation": { | |
| "correct": 9, | |
| "total": 9, | |
| "accuracy": 1.0 | |
| }, | |
| "unrelated": { | |
| "correct": 30, | |
| "total": 30, | |
| "accuracy": 1.0 | |
| } | |
| }, | |
| "classification_report": " precision recall f1-score support\n\n no_match 0.99 0.97 0.98 1470\n match 0.96 0.98 0.97 1040\n\n accuracy 0.97 2510\n macro avg 0.97 0.97 0.97 2510\nweighted avg 0.97 0.97 0.97 2510\n", | |
| "model_type": "trained_person_v6_lora_focal", | |
| "test_set_version": "v3", | |
| "test_set_size": 2510 | |
| } |