richie-ghost/setfit-sentence-t5-large-MentalHealth-Topic-Check

SetFit with sentence-transformers/sentence-t5-large

This is a SetFit model that can be used for Text Classification. This SetFit model uses sentence-transformers/sentence-t5-large 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: sentence-transformers/sentence-t5-large
• Classification head: a LogisticRegression instance
• Maximum Sequence Length: 256 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
True	'Exploring historical landmarks in Europe' 'How to create an effective resume' 'Exercises to improve core strength'
False	'Feeling sad or empty for long periods without any specific reason' 'Dealing with the emotional impact of chronic illness' 'Understanding and coping with panic attacks'

Evaluation

Metrics

Label	Accuracy
all	1.0

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("richie-ghost/setfit-sentence-t5-large-MentalHealth-Topic-Check")
# Run inference
preds = model("Understanding stock market trends")

Training Details

Training Set Metrics

Training set	Min	Median	Max
Word count	4	6.4583	11

Label	Training Sample Count
True	22
False	26

Training Hyperparameters

• batch_size: (16, 16)
• num_epochs: (8, 8)
• max_steps: -1
• sampling_strategy: oversampling
• body_learning_rate: (2e-05, 1e-05)
• head_learning_rate: 0.01
• 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: True

Training Results

Epoch	Step	Training Loss	Validation Loss
0.0132	1	0.2504	-
0.6579	50	0.2115	-
1.0	76	-	0.122
1.3158	100	0.0335	-
1.9737	150	0.007	-
2.0	152	-	0.0352
2.6316	200	0.0037	-
3.0	228	-	0.0229
3.2895	250	0.0019	-
3.9474	300	0.0016	-
4.0	304	-	0.0199
4.6053	350	0.001	-
5.0	380	-	0.0162
5.2632	400	0.0009	-
5.9211	450	0.0006	-
6.0	456	-	0.0141
6.5789	500	0.0008	-
7.0	532	-	0.0133
7.2368	550	0.0008	-
7.8947	600	0.0006	-
8.0	608	-	0.0154

• The bold row denotes the saved checkpoint.

Framework Versions

• Python: 3.10.12
• SetFit: 1.0.3
• Sentence Transformers: 2.7.0
• Transformers: 4.40.0
• PyTorch: 2.2.1+cu121
• Datasets: 2.19.0
• Tokenizers: 0.19.1

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}
}