Zero-Shot Text Classification — convmatch

Multi-scale CNN encoder over pretrained embeddings (no transformer at inference).

This model encodes texts and candidate labels into a shared embedding space using BERT, enabling classification into arbitrary categories without retraining for new labels.

Training Details

Parameter	Value
Base model	bert-base-uncased
Model variant	convmatch
Training steps	1000
Batch size	2
Learning rate	2e-05
Trainable params	24,948,992
Training time	84.1s

Dataset

Trained on polodealvarado/zeroshot-classification.

Evaluation Results

Metric	Score
Precision	0.7531
Recall	0.9922
F1 Score	0.8563

Usage

from models.convmatch import ConvMatchModel

model = ConvMatchModel.from_pretrained("polodealvarado/convmatch")

predictions = model.predict(
    texts=["The stock market crashed yesterday."],
    labels=[["Finance", "Sports", "Biology", "Economy"]],
)
print(predictions)
# [{"text": "...", "scores": {"Finance": 0.98, "Economy": 0.85, ...}}]