Upload README.md with huggingface_hub
Browse files
README.md
ADDED
|
@@ -0,0 +1,59 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
tags:
|
| 3 |
+
- named-entity-recognition
|
| 4 |
+
- clinical-nlp
|
| 5 |
+
- multiclinner
|
| 6 |
+
- multi-head-crf
|
| 7 |
+
- token-classification
|
| 8 |
+
language:
|
| 9 |
+
- en
|
| 10 |
+
license: apache-2.0
|
| 11 |
+
---
|
| 12 |
+
|
| 13 |
+
# MultiClinNER EN Models
|
| 14 |
+
|
| 15 |
+
Clinical NER models for EN, trained with Multi-Head CRF architecture.
|
| 16 |
+
|
| 17 |
+
## Best Model
|
| 18 |
+
|
| 19 |
+
- **Model**: `microsoft-BiomedNLP-PubMedBERT-large-uncased-abstract-C64-H3-E30-Arandom-%0.25-P0.5-42`
|
| 20 |
+
- **Best F1**: 0.7368
|
| 21 |
+
- **Branch**: `main`
|
| 22 |
+
|
| 23 |
+
## Usage
|
| 24 |
+
|
| 25 |
+
```python
|
| 26 |
+
# Load the best model (main branch)
|
| 27 |
+
from transformers import AutoTokenizer, AutoModelForTokenClassification
|
| 28 |
+
|
| 29 |
+
model = AutoModelForTokenClassification.from_pretrained("IEETA/MultiClinNER-EN")
|
| 30 |
+
tokenizer = AutoTokenizer.from_pretrained("IEETA/MultiClinNER-EN")
|
| 31 |
+
|
| 32 |
+
# Load a specific model variant
|
| 33 |
+
model = AutoModelForTokenClassification.from_pretrained("IEETA/MultiClinNER-EN", revision="BRANCH_NAME")
|
| 34 |
+
```
|
| 35 |
+
|
| 36 |
+
## All Models (20 variants)
|
| 37 |
+
|
| 38 |
+
| Branch | Model | Best? |
|
| 39 |
+
|--------|-------|-------|
|
| 40 |
+
| `main` | `microsoft-BiomedNLP-PubMedBERT-large-uncased-abstract-C64-H3-E30-Arandom-%0.25-P0.5-42` | **Yes** |
|
| 41 |
+
| `microsoft-BiomedNLP-PubMedBERT-large-uncased-abstract-C64-H3-E30-Anone-pct0.2-P0.5-123` | `microsoft-BiomedNLP-PubMedBERT-large-uncased-abstract-C64-H3-E30-Anone-%0.2-P0.5-123` | |
|
| 42 |
+
| `microsoft-BiomedNLP-PubMedBERT-large-uncased-abstract-C64-H3-E30-Anone-pct0.2-P0.5-42` | `microsoft-BiomedNLP-PubMedBERT-large-uncased-abstract-C64-H3-E30-Anone-%0.2-P0.5-42` | |
|
| 43 |
+
| `microsoft-BiomedNLP-PubMedBERT-large-uncased-abstract-C64-H3-E30-Anone-pct0.2-P0.5-456` | `microsoft-BiomedNLP-PubMedBERT-large-uncased-abstract-C64-H3-E30-Anone-%0.2-P0.5-456` | |
|
| 44 |
+
| `microsoft-BiomedNLP-PubMedBERT-large-uncased-abstract-C64-H3-E30-Anone-pct0.2-P0.5-999` | `microsoft-BiomedNLP-PubMedBERT-large-uncased-abstract-C64-H3-E30-Anone-%0.2-P0.5-999` | |
|
| 45 |
+
| `microsoft-BiomedNLP-PubMedBERT-large-uncased-abstract-C64-H3-E30-Arandom-pct0.1-P0.5-123` | `microsoft-BiomedNLP-PubMedBERT-large-uncased-abstract-C64-H3-E30-Arandom-%0.1-P0.5-123` | |
|
| 46 |
+
| `microsoft-BiomedNLP-PubMedBERT-large-uncased-abstract-C64-H3-E30-Arandom-pct0.1-P0.5-42` | `microsoft-BiomedNLP-PubMedBERT-large-uncased-abstract-C64-H3-E30-Arandom-%0.1-P0.5-42` | |
|
| 47 |
+
| `microsoft-BiomedNLP-PubMedBERT-large-uncased-abstract-C64-H3-E30-Arandom-pct0.1-P0.5-456` | `microsoft-BiomedNLP-PubMedBERT-large-uncased-abstract-C64-H3-E30-Arandom-%0.1-P0.5-456` | |
|
| 48 |
+
| `microsoft-BiomedNLP-PubMedBERT-large-uncased-abstract-C64-H3-E30-Arandom-pct0.1-P0.5-999` | `microsoft-BiomedNLP-PubMedBERT-large-uncased-abstract-C64-H3-E30-Arandom-%0.1-P0.5-999` | |
|
| 49 |
+
| `microsoft-BiomedNLP-PubMedBERT-large-uncased-abstract-C64-H3-E30-Arandom-pct0.25-P0.5-123` | `microsoft-BiomedNLP-PubMedBERT-large-uncased-abstract-C64-H3-E30-Arandom-%0.25-P0.5-123` | |
|
| 50 |
+
| `microsoft-BiomedNLP-PubMedBERT-large-uncased-abstract-C64-H3-E30-Arandom-pct0.25-P0.5-456` | `microsoft-BiomedNLP-PubMedBERT-large-uncased-abstract-C64-H3-E30-Arandom-%0.25-P0.5-456` | |
|
| 51 |
+
| `microsoft-BiomedNLP-PubMedBERT-large-uncased-abstract-C64-H3-E30-Arandom-pct0.25-P0.5-999` | `microsoft-BiomedNLP-PubMedBERT-large-uncased-abstract-C64-H3-E30-Arandom-%0.25-P0.5-999` | |
|
| 52 |
+
| `microsoft-BiomedNLP-PubMedBERT-large-uncased-abstract-C64-H3-E30-Aukn-pct0.25-P0.2-123` | `microsoft-BiomedNLP-PubMedBERT-large-uncased-abstract-C64-H3-E30-Aukn-%0.25-P0.2-123` | |
|
| 53 |
+
| `microsoft-BiomedNLP-PubMedBERT-large-uncased-abstract-C64-H3-E30-Aukn-pct0.25-P0.2-42` | `microsoft-BiomedNLP-PubMedBERT-large-uncased-abstract-C64-H3-E30-Aukn-%0.25-P0.2-42` | |
|
| 54 |
+
| `microsoft-BiomedNLP-PubMedBERT-large-uncased-abstract-C64-H3-E30-Aukn-pct0.25-P0.2-456` | `microsoft-BiomedNLP-PubMedBERT-large-uncased-abstract-C64-H3-E30-Aukn-%0.25-P0.2-456` | |
|
| 55 |
+
| `microsoft-BiomedNLP-PubMedBERT-large-uncased-abstract-C64-H3-E30-Aukn-pct0.25-P0.2-999` | `microsoft-BiomedNLP-PubMedBERT-large-uncased-abstract-C64-H3-E30-Aukn-%0.25-P0.2-999` | |
|
| 56 |
+
| `microsoft-BiomedNLP-PubMedBERT-large-uncased-abstract-C64-H3-E30-Aukn-pct0.25-P0.5-123` | `microsoft-BiomedNLP-PubMedBERT-large-uncased-abstract-C64-H3-E30-Aukn-%0.25-P0.5-123` | |
|
| 57 |
+
| `microsoft-BiomedNLP-PubMedBERT-large-uncased-abstract-C64-H3-E30-Aukn-pct0.25-P0.5-42` | `microsoft-BiomedNLP-PubMedBERT-large-uncased-abstract-C64-H3-E30-Aukn-%0.25-P0.5-42` | |
|
| 58 |
+
| `microsoft-BiomedNLP-PubMedBERT-large-uncased-abstract-C64-H3-E30-Aukn-pct0.25-P0.5-456` | `microsoft-BiomedNLP-PubMedBERT-large-uncased-abstract-C64-H3-E30-Aukn-%0.25-P0.5-456` | |
|
| 59 |
+
| `microsoft-BiomedNLP-PubMedBERT-large-uncased-abstract-C64-H3-E30-Aukn-pct0.25-P0.5-999` | `microsoft-BiomedNLP-PubMedBERT-large-uncased-abstract-C64-H3-E30-Aukn-%0.25-P0.5-999` | |
|