You can deploy this model and start earning money today!

MiguelRod/
SetFit-MeSH-Biomaterials-Classifier

Name: SetFit-MeSH-Biomaterials-Classifier
Rating: 2.5 (12 reviews)

setfit

safetensors

bert

sentence-transformers

text-classification

generated_from_setfit_trainer

2209.11055

cambridgeltl/SapBERT-from-PubMedBERT-fulltext

finetune

model-index

Views

No views yet

SetFit with cambridgeltl/SapBERT-from-PubMedBERT-fulltext

This is a SetFit model that can be used for Text Classification. This SetFit model uses cambridgeltl/SapBERT-from-PubMedBERT-fulltext as the Sentence Transformer embedding model. A OneVsRestClassifier instance is used for classification.

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

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

Model Details

Model Description

Model Type: SetFit
Sentence Transformer body: cambridgeltl/SapBERT-from-PubMedBERT-fulltext
Classification head: a OneVsRestClassifier instance
Maximum Sequence Length: 512 tokens

Model Sources

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

Evaluation

Metrics

Label	Accuracy
all	0.5241

Uses

Direct Use for Inference

First install the SetFit library:

pip install setfit

Then you can load this model and run inference.

python

from setfit import SetFitModel

# Download from the 🤗 Hub
model = SetFitModel.from_pretrained("setfit_model_id")
# Run inference
preds = model("foreign-body reaction: Chronic inflammation and granuloma formation around irritating foreign bodies.")

Training Details

Training Set Metrics

Training set	Min	Median	Max
Word count	2	30.4473	134

Training Hyperparameters

batch_size: (16, 16)
num_epochs: (1, 1)
max_steps: -1
sampling_strategy: oversampling
num_iterations: 20
body_learning_rate: (2e-05, 2e-05)
head_learning_rate: 2e-05
loss: CosineSimilarityLoss
distance_metric: cosine_distance
margin: 0.25
end_to_end: False
use_amp: False
warmup_proportion: 0.1
l2_weight: 0.01
seed: 42
eval_max_steps: -1
load_best_model_at_end: False

Training Results

Epoch	Step	Training Loss	Validation Loss
0.0003	1	0.2929	-
0.0136	50	0.2377	-
0.0272	100	0.2321	-
0.0408	150	0.2199	-
0.0544	200	0.1726	-
0.0680	250	0.1355	-
0.0816	300	0.1207	-
0.0952	350	0.1138	-
0.1088	400	0.1154	-
0.1223	450	0.095	-
0.1359	500	0.106	-
0.1495	550	0.0913	-
0.1631	600	0.0943	-
0.1767	650	0.0974	-
0.1903	700	0.0923	-
0.2039	750	0.0893	-
0.2175	800	0.0804	-
0.2311	850	0.0849	-
0.2447	900	0.0766	-
0.2583	950	0.0838	-
0.2719	1000	0.0725	-
0.2855	1050	0.073	-
0.2991	1100	0.055	-
0.3127	1150	0.0758	-
0.3263	1200	0.0709	-
0.3399	1250	0.0569	-
0.3535	1300	0.0535	-
0.3670	1350	0.0557	-
0.3806	1400	0.0596	-
0.3942	1450	0.0453	-
0.4078	1500	0.0428	-
0.4214	1550	0.0482	-
0.4350	1600	0.0465	-
0.4486	1650	0.0469	-
0.4622	1700	0.0479	-
0.4758	1750	0.0451	-
0.4894	1800	0.0613	-
0.5030	1850	0.0533	-
0.5166	1900	0.0476	-
0.5302	1950	0.0449	-
0.5438	2000	0.0543	-
0.5574	2050	0.0509	-
0.5710	2100	0.043	-
0.5846	2150	0.0482	-
0.5982	2200	0.0513	-
0.6117	2250	0.0366	-
0.6253	2300	0.0385	-
0.6389	2350	0.0446	-
0.6525	2400	0.0411	-
0.6661	2450	0.037	-
0.6797	2500	0.0321	-
0.6933	2550	0.0468	-
0.7069	2600	0.0331	-
0.7205	2650	0.0315	-
0.7341	2700	0.0435	-
0.7477	2750	0.0394	-
0.7613	2800	0.0381	-
0.7749	2850	0.0418	-
0.7885	2900	0.0347	-
0.8021	2950	0.0468	-
0.8157	3000	0.0352	-
0.8293	3050	0.0416	-
0.8428	3100	0.0354	-
0.8564	3150	0.0329	-
0.8700	3200	0.0359	-
0.8836	3250	0.036	-
0.8972	3300	0.0362	-
0.9108	3350	0.0296	-
0.9244	3400	0.041	-
0.9380	3450	0.0375	-
0.9516	3500	0.0282	-
0.9652	3550	0.0341	-
0.9788	3600	0.0283	-
0.9924	3650	0.0339	-

Framework Versions

Python: 3.11.11
SetFit: 1.1.1
Sentence Transformers: 3.4.1
Transformers: 4.50.0
PyTorch: 2.6.0+cu124
Datasets: 3.4.1
Tokenizers: 0.21.1

Citation

BibTeX

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}
}

MiguelRod/SetFit-MeSH-Biomaterials-Classifier