IEETA/MultiClinNER-IT

MultiClinNER IT Models

Clinical NER models for IT, trained with Multi-Head CRF architecture.

Best Model

• Model: IVN-RIN-C64-H1-E60-Arandom-%0.1-P0.2-42
• Best F1: 0.7062
• Branch: main

Usage

# Load the best model (main branch)
from transformers import AutoTokenizer, AutoModelForTokenClassification

model = AutoModelForTokenClassification.from_pretrained("IEETA/MultiClinNER-IT")
tokenizer = AutoTokenizer.from_pretrained("IEETA/MultiClinNER-IT")

# Load a specific model variant
model = AutoModelForTokenClassification.from_pretrained("IEETA/MultiClinNER-IT", revision="BRANCH_NAME")

All Models (14 variants)

Branch	Model	Best?
main	IVN-RIN-C64-H1-E60-Arandom-%0.1-P0.2-42	Yes
IVN-RIN-C64-H1-E60-Arandom-pct0.1-P0.2-123	IVN-RIN-C64-H1-E60-Arandom-%0.1-P0.2-123
IVN-RIN-C64-H1-E60-Arandom-pct0.1-P0.2-456	IVN-RIN-C64-H1-E60-Arandom-%0.1-P0.2-456
IVN-RIN-C64-H1-E60-Arandom-pct0.1-P0.2-999	IVN-RIN-C64-H1-E60-Arandom-%0.1-P0.2-999
IVN-RIN-C64-H1-E60-Arandom-pct0.25-P0.5-123	IVN-RIN-C64-H1-E60-Arandom-%0.25-P0.5-123
IVN-RIN-C64-H1-E60-Arandom-pct0.25-P0.5-42	IVN-RIN-C64-H1-E60-Arandom-%0.25-P0.5-42
IVN-RIN-C64-H1-E60-Arandom-pct0.25-P0.5-456	IVN-RIN-C64-H1-E60-Arandom-%0.25-P0.5-456
IVN-RIN-C64-H1-E60-Aukn-pct0.1-P0.5-123	IVN-RIN-C64-H1-E60-Aukn-%0.1-P0.5-123
IVN-RIN-C64-H1-E60-Aukn-pct0.1-P0.5-42	IVN-RIN-C64-H1-E60-Aukn-%0.1-P0.5-42
IVN-RIN-C64-H1-E60-Aukn-pct0.1-P0.5-456	IVN-RIN-C64-H1-E60-Aukn-%0.1-P0.5-456
IVN-RIN-C64-H1-E60-Aukn-pct0.25-P0.5-123	IVN-RIN-C64-H1-E60-Aukn-%0.25-P0.5-123
IVN-RIN-C64-H1-E60-Aukn-pct0.25-P0.5-42	IVN-RIN-C64-H1-E60-Aukn-%0.25-P0.5-42
IVN-RIN-C64-H1-E60-Aukn-pct0.25-P0.5-456	IVN-RIN-C64-H1-E60-Aukn-%0.25-P0.5-456
IVN-RIN-C64-H1-E60-Aukn-pct0.25-P0.5-999	IVN-RIN-C64-H1-E60-Aukn-%0.25-P0.5-999