Text Classification
Model2Vec
English
poster-sentry
document-classification
scientific-posters
multimodal
poster-detection
machine-actionable
FAIR-data
posters-science
quality-control
Instructions to use fairdataihub/poster-sentry with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Model2Vec
How to use fairdataihub/poster-sentry with Model2Vec:
from model2vec import StaticModel model = StaticModel.from_pretrained("fairdataihub/poster-sentry") embeddings = model.encode(["It's dangerous to go alone!", "It's a secret to everybody."]) print(embeddings.shape) - Notebooks
- Google Colab
- Kaggle
config.json: v1.3.0 license-screened retrain numbers
Browse files- config.json +7 -7
config.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
| 1 |
{
|
| 2 |
"model_type": "poster-sentry",
|
| 3 |
-
"version": "1.
|
| 4 |
"architecture": "stacked_logistic_regression_multimodal",
|
| 5 |
"embedding_backbone": "minishlab/potion-base-32M",
|
| 6 |
"embedding_dim": 512,
|
|
@@ -15,17 +15,17 @@
|
|
| 15 |
"head_file": "models/poster_sentry_head.npz",
|
| 16 |
"scaler": "StandardScaler",
|
| 17 |
"classifier": "LogisticRegression",
|
| 18 |
-
"accuracy": 0.
|
| 19 |
"accuracy_ci95": [
|
| 20 |
-
0.
|
| 21 |
-
0.
|
| 22 |
],
|
| 23 |
-
"oof_accuracy": 0.
|
| 24 |
-
"training_samples":
|
| 25 |
"label_provenance": "human-validated (three-reviewer survey with blinded adjudication)",
|
| 26 |
"library_name": "model2vec",
|
| 27 |
"pdf_backend": "pdfplumber + pypdfium2",
|
| 28 |
"stage1": "LogisticRegression over the 512-d text embedding; poster probability becomes the text_score feature",
|
| 29 |
"final_classifier_features": 31,
|
| 30 |
-
"corpus_poster_pct": 80.
|
| 31 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"model_type": "poster-sentry",
|
| 3 |
+
"version": "1.3.0",
|
| 4 |
"architecture": "stacked_logistic_regression_multimodal",
|
| 5 |
"embedding_backbone": "minishlab/potion-base-32M",
|
| 6 |
"embedding_dim": 512,
|
|
|
|
| 15 |
"head_file": "models/poster_sentry_head.npz",
|
| 16 |
"scaler": "StandardScaler",
|
| 17 |
"classifier": "LogisticRegression",
|
| 18 |
+
"accuracy": 0.9313,
|
| 19 |
"accuracy_ci95": [
|
| 20 |
+
0.9055,
|
| 21 |
+
0.9504
|
| 22 |
],
|
| 23 |
+
"oof_accuracy": 0.9391,
|
| 24 |
+
"training_samples": 3298,
|
| 25 |
"label_provenance": "human-validated (three-reviewer survey with blinded adjudication)",
|
| 26 |
"library_name": "model2vec",
|
| 27 |
"pdf_backend": "pdfplumber + pypdfium2",
|
| 28 |
"stage1": "LogisticRegression over the 512-d text embedding; poster probability becomes the text_score feature",
|
| 29 |
"final_classifier_features": 31,
|
| 30 |
+
"corpus_poster_pct": 80.5
|
| 31 |
}
|