nazhan/bge-large-en-v1.5-brahmaputra-qa-lookup-iter-1-2-epoch

SetFit with BAAI/bge-large-en-v1.5

This is a SetFit model trained on the nazhan/qa-lookup-dataset-iter-1 dataset that can be used for Text Classification. This SetFit model uses BAAI/bge-large-en-v1.5 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: BAAI/bge-large-en-v1.5
• Classification head: a LogisticRegression instance
• Maximum Sequence Length: 512 tokens
• Number of Classes: 2 classes
• Training Dataset: nazhan/qa-lookup-dataset-iter-1

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
Lookup	'Get me the list of customers who placed their first order in 2024.' "Filter by products in the 'Gadgets' category and show me their prices." 'Get me the email addresses of customers who have made a purchase.'
qa	'Provide the value of the accrued vacation liability as of June 2023.' 'Show me the value of the courier service charges for November 2023.' "Provide the value of the consulting contract with 'Client N' finalized in 2023."

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("nazhan/bge-large-en-v1.5-brahmaputra-qa-lookup-iter-1-2-epoch")
# Run inference
preds = model("Provide the value of the environmental compliance cost for 2023.")

Training Details

Training Set Metrics

Training set	Min	Median	Max
Word count	8	12.8309	19

Label	Training Sample Count
Lookup	65
qa	71

Training Hyperparameters

• batch_size: (32, 32)
• num_epochs: (2, 2)
• 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
0.0034	1	0.1823	-
0.1701	50	0.0031	-
0.3401	100	0.0012	-
0.5102	150	0.0011	-
0.6803	200	0.0009	-
0.8503	250	0.0008	-
1.0	294	-	0.0004
1.0204	300	0.0008	-
1.1905	350	0.0008	-
1.3605	400	0.0007	-
1.5306	450	0.0006	-
1.7007	500	0.0006	-
1.8707	550	0.0006	-
2.0	588	-	0.0003

• The bold row denotes the saved checkpoint.

Framework Versions

• Python: 3.11.9
• SetFit: 1.1.0.dev0
• Sentence Transformers: 3.0.1
• Transformers: 4.44.2
• PyTorch: 2.4.0+cu121
• Datasets: 2.21.0
• 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}
}