Fine-tuned LoRA Classifier on MobileBERT

This is a fine-tuned LoRA (Low-Rank Adaptation) classifier on MobileBERT, designed for intent classification across multiple categories. MobileBERT is a compact and efficient version of BERT_LARGE, optimized for resource-constrained devices while maintaining robust performance. The model has been fine-tuned to classify intents such as information queries, navigation, purchase, and more.

Model Details

Model Description

This model is based on MobileBERT (uncased_L-24_H-128_B-512_A-4_F-4_OPT) and fine-tuned using LoRA for intent classification. The fine-tuning process adapts the model to predict intents across 8 categories:

  • information_intent
  • yelp_intent
  • navigation_intent
  • travel_intent
  • purchase_intent
  • weather_intent
  • translation_intent
  • unknown

Model Sources

Citation

If you use this model, please cite it as:

@misc{mozilla_mobilebert_lora_intent,
  title       = {Fine-tuned LoRA Classifier on MobileBERT},
  author      = {Mozilla},
  year        = {2024},
  url         = {https://huggingface.co/Mozilla/mobilebert-uncased-finetuned-LoRA-intent-classifier},
  license     = {Apache-2.0}
}
Downloads last month
31
Safetensors
Model size
24.6M params
Tensor type
F32
ยท
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Space using Mozilla/mobilebert-uncased-finetuned-LoRA-intent-classifier 1