Instructions to use jcfossati/ForSureLLM with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- sentence-transformers
How to use jcfossati/ForSureLLM with sentence-transformers:
from sentence_transformers import SentenceTransformer model = SentenceTransformer("jcfossati/ForSureLLM") sentences = [ "The weather is lovely today.", "It's so sunny outside!", "He drove to the stadium." ] embeddings = model.encode(sentences) similarities = model.similarity(embeddings, embeddings) print(similarities.shape) # [3, 3] - Notebooks
- Google Colab
- Kaggle
Upload forsurellm-int8.onnx
Browse files---
language:
- fr
- en
tags:
- text-classification
- yes-no
- onnx
- distillation
license: mit
---
# ForSureLLM — yes/no/unknown classifier
Source : https://github.com/jcfossati/ForSureLLM
Basic analysis of english and french language for yes/no detection.
When knowing the question and waiting a yes/no answer, asking a frontier LLM is overkill using to much resources, and latency is high for a limited action.
## Stats
Distilled via KL-divergence on soft labels.
MiniLM-L12-v2 multilingual backbone, fine-tuned + int8 quantized.
- Accuracy: 91.4%
- ECE: 0.007 (calibrated)
- Latency: 2.5ms CPU
- Size: 113 MB
## Usage
- Interaction between an application and the user featuring free-form text input.
- Chatbot asking user and needed a yes/no answer
- forsurellm-int8.onnx +3 -0
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e2faa2a8329fe09a17aa5ae7a8943fef375805fec1153bef9c39e67d03b5a306
|
| 3 |
+
size 118268722
|