neo-tax/technical-in-nature-classifier-for-projects

SetFit with avsolatorio/GIST-Embedding-v0

This is a SetFit model that can be used for Text Classification. This SetFit model uses avsolatorio/GIST-Embedding-v0 as the Sentence Transformer embedding model. A LogisticRegression instance is used for classification.

The model has been trained using an efficient few-shot learning technique that involves:

1. Fine-tuning a Sentence Transformer with contrastive learning.
2. Training a classification head with features from the fine-tuned Sentence Transformer.

Model Details

Model Description

• Model Type: SetFit
• Sentence Transformer body: avsolatorio/GIST-Embedding-v0
• Classification head: a LogisticRegression instance
• Maximum Sequence Length: 512 tokens
• Number of Classes: 2 classes

Model Sources

• Repository: SetFit on GitHub
• Paper: Efficient Few-Shot Learning Without Prompts
• Blogpost: SetFit: Efficient Few-Shot Learning Without Prompts

Model Labels

Label	Examples
0	"A manufacturing corporation undertakes an initiative to restructure its manufacturing organization by designing an organizational structure that will improve the company's business operations" "Centers on the production of content for the Brief product. This includes tasks related to drafting insights, creating case studies, and publishing social media posts. The project aims to provide valuable and timely information to Kharon's clients, helping them stay informed about global security topics that impact their commercial activities." 'The team is developing a comprehensive marketing strategy to increase brand awareness and customer engagement. This includes creating targeted advertising campaigns, optimizing our social media presence, and collaborating with influencers to promote our products. We will also analyze market trends and consumer behavior to refine our approach.'
1	"Project focused on enhancing the website's functionality, including tasks related to optimizing search functionality and integrating new features such as bookmarks and table of contents for the web reader. The project aims to provide a seamless online experience for customers by improving the efficiency and speed of our website." 'Design and create an innovative drug delivery system for cancer treatment compatible with different types of cancer and different patient profiles while minimizing negative impacts on healthy tissues' 'Develop a new and advanced Natural Language Processing (NLP) algorithm to enhance the capabilities of virtual assistants used in various applications, such as customer service chatbots. This project involved improving the NLP algorithm to be more responsive in the area of complex natural language understanding, including context comprehension, sentiment analysis, and accurate response generation'

Uses

Direct Use for Inference

First install the SetFit library:

pip install setfit

Then you can load this model and run inference.

from setfit import SetFitModel

# Download from the 🤗 Hub
model = SetFitModel.from_pretrained("setfit_model_id")
# Run inference
preds = model("A tire manufacturing company created a new belt to be used as part of tread cooling during the manufacturing process. Such a belt is not commercially available.")

Training Details

Training Set Metrics

Training set	Min	Median	Max
Word count	23	43.5	54

Label	Training Sample Count
0	8
1	16

Training Hyperparameters

• batch_size: (16, 16)
• num_epochs: (3, 3)
• max_steps: -1
• sampling_strategy: oversampling
• num_iterations: 20
• body_learning_rate: (0.0001, 0.0001)
• head_learning_rate: 0.0001
• loss: CosineSimilarityLoss
• distance_metric: cosine_distance
• margin: 0.25
• end_to_end: False
• use_amp: False
• warmup_proportion: 0.1
• seed: 42
• eval_max_steps: -1
• load_best_model_at_end: False

Training Results

Epoch	Step	Training Loss	Validation Loss
0.0167	1	0.2764	-
0.8333	50	0.0014	-
1.6667	100	0.0011	-
2.5	150	0.0011	-

Framework Versions

• Python: 3.9.16
• SetFit: 1.0.3
• Sentence Transformers: 3.0.1
• Transformers: 4.39.0
• PyTorch: 2.3.1
• Datasets: 2.19.2
• Tokenizers: 0.15.2

Citation

BibTeX

@article{https://doi.org/10.48550/arxiv.2209.11055,
    doi = {10.48550/ARXIV.2209.11055},
    url = {https://arxiv.org/abs/2209.11055},
    author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren},
    keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences},
    title = {Efficient Few-Shot Learning Without Prompts},
    publisher = {arXiv},
    year = {2022},
    copyright = {Creative Commons Attribution 4.0 International}
}