stsb-quora-distilroberta-linear-merge
A demonstration model produced by the native model-merging feature (CrossEncoder.merge)
added to Sentence Transformers. It is a
showcase merge, not a task-tuned release.
It merges these two checkpoints with the linear method (weights [0.5, 0.5]):
How it was created
from sentence_transformers import CrossEncoder
merged = CrossEncoder.merge(
models=["cross-encoder/stsb-distilroberta-base", "cross-encoder/quora-distilroberta-base"],
weights=[0.5, 0.5],
method="linear",
output_path="stsb-quora-distilroberta-linear-merge",
dtype="float16",
)
Usage
from sentence_transformers import CrossEncoder
model = CrossEncoder("yjoonjang/stsb-quora-distilroberta-linear-merge")
scores = model.predict([
("A man is eating food.", "A man is eating a meal."),
("How do I learn Python?", "What is the best way to learn Python?"),
])
print(scores)
License
This is a derivative of the two base models above; their licenses apply. See each base model's card for terms (note that some bases — e.g. the SPLADE cocondenser models — are non-commercial).
- Downloads last month
- 19
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support