metadata

base_model: sentence-transformers/all-MiniLM-L6-v2
library_name: setfit
metrics:
  - accuracy
pipeline_tag: text-classification
tags:
  - setfit
  - sentence-transformers
  - text-classification
  - generated_from_setfit_trainer
widget:
  - text: >-
      What are the key components involved in developing a deep learning model
      for handwritten digit recognition?
  - text: What is the purpose of the message posted by the CR?
  - text: >-
      How can researchers create and maintain public repositories for
      reproducible research?
  - text: >-
      What are the key components involved in developing a deep learning model
      for handwritten digit recognition?
  - text: >-
      How do you prioritize and delegate tasks to ensure efficient collaboration
      and feedback?
inference: true
model-index:
  - name: SetFit with sentence-transformers/all-MiniLM-L6-v2
    results:
      - task:
          type: text-classification
          name: Text Classification
        dataset:
          name: Unknown
          type: unknown
          split: test
        metrics:
          - type: accuracy
            value: 0.5
            name: Accuracy

SetFit with sentence-transformers/all-MiniLM-L6-v2

This is a SetFit model that can be used for Text Classification. This SetFit model uses sentence-transformers/all-MiniLM-L6-v2 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:

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: sentence-transformers/all-MiniLM-L6-v2
Classification head: a LogisticRegression instance
Maximum Sequence Length: 256 tokens
Number of Classes: 4 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
lexical	'What are the key considerations when choosing an optimization method for a complex problem?' 'What are the challenges of being a remote mentor or sponsor?' 'How do researchers typically obtain information on the ranking of machine learning conferences?'
semantic	'What are common issues that users may encounter when accessing a platform that uses JumpCloud for authentication?' 'What are the key components involved in developing a deep learning model for handwritten digit recognition?' 'How can machine learning and data enrichment be used to improve business outcomes in various industries?'
very_semantic	"What are people's opinions on a particular topic?" 'What are the key considerations when proposing names for a project or initiative?' 'What are the key considerations for successful collaboration between industry and academia in research and development projects?'
very_lexical	'How can one track and store keys in a Flink operator?' 'What role do companies like Solvay play in addressing key societal challenges through their business strategies and operations?' 'What is the purpose of the scoring methodology in determining RAI maturity?'

Evaluation

Metrics

Label	Accuracy
all	0.5

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("yaniseuranova/setfit-rag-hybrid-search-query-router-test")
# Run inference
preds = model("What is the purpose of the message posted by the CR?")

Training Details

Training Set Metrics

Training set	Min	Median	Max
Word count	8	14.4138	24

Label	Training Sample Count
lexical	32
semantic	21
very_lexical	10
very_semantic	24

Training Hyperparameters

batch_size: (8, 8)
num_epochs: (3, 3)
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
0.0015	1	0.268	-
0.0736	50	0.2649	-
0.1473	100	0.3352	-
0.2209	150	0.2516	-
0.2946	200	0.2438	-
0.3682	250	0.1808	-
0.4418	300	0.2365	-
0.5155	350	0.1337	-
0.5891	400	0.2263	-
0.6627	450	0.1936	-
0.7364	500	0.0612	-
0.8100	550	0.1664	-
0.8837	600	0.0987	-
0.9573	650	0.0736	-
1.0	679	-	0.2288
1.0309	700	0.0568	-
1.1046	750	0.0765	-
1.1782	800	0.1193	-
1.2518	850	0.199	-
1.3255	900	0.2734	-
1.3991	950	0.194	-
1.4728	1000	0.1085	-
1.5464	1050	0.1496	-
1.6200	1100	0.1673	-
1.6937	1150	0.2225	-
1.7673	1200	0.0503	-
1.8409	1250	0.1531	-
1.9146	1300	0.2287	-
1.9882	1350	0.1187	-
2.0	1358	-	0.2055
2.0619	1400	0.0546	-
2.1355	1450	0.2072	-
2.2091	1500	0.1208	-
2.2828	1550	0.0837	-
2.3564	1600	0.0405	-
2.4300	1650	0.1334	-
2.5037	1700	0.1458	-
2.5773	1750	0.2189	-
2.6510	1800	0.0561	-
2.7246	1850	0.1656	-
2.7982	1900	0.1351	-
2.8719	1950	0.1826	-
2.9455	2000	0.1905	-
3.0	2037	-	0.2273

The bold row denotes the saved checkpoint.

Framework Versions

Python: 3.10.12
SetFit: 1.0.3
Sentence Transformers: 2.6.1
Transformers: 4.39.0
PyTorch: 2.3.1+cu121
Datasets: 2.18.0
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}
}