NaniBer/my_setfit_model

SetFit with sentence-transformers/paraphrase-mpnet-base-v2

This is a SetFit model that can be used for Text Classification. This SetFit model uses sentence-transformers/paraphrase-mpnet-base-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:

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/paraphrase-mpnet-base-v2
• Classification head: a LogisticRegression instance
• Maximum Sequence Length: 512 tokens
• Number of Classes: 23 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
27	'that is what retardation looks like' "Don't kiss your doorbell! Or anyone else's for that matter..." 'Hello everyone. Im from Toronto as well. Can call and visit in personal if needed.'
2	"Troll, bro. They know they're saying stupid shit. The motherfucker does nothing but stink up libertarian subs talking shit" 'Ok, then what the actual fuck is your plan?' 'Stupidly stubborn / stubbornly stupid'
4	'Pay you for what, just standing there? Done.' "Sometimes life actually hands you lemons. We're just lucky that we have a proverbial phrase that gives us an idea of what we can do with them." "true I am a troll, but fortunately for me I'm not emotionally invested in it."
3	'Dirty Southern Wankers' 'Shit, I guess I accidentally bought a Pay-Per-View boxing match' 'The republicans are the military. You are an idiot.'
0	'Awesome! I’m a cradle [RELIGION], so really interesting to hear your experience. Thanks for sharing.' 'What a wonderful world' 'Twilight... STILL a better love story than The Last Jedi!'
6	"All sounds possible except the key, I can't see how it was missed in the first search. " 'What does FPTP have to do with the referendum?' 'Maybe that’s what happened to the great white at Houston zoo'
10	"This isn't really wholesome"
16	'I read on a different post that he died shortly after of internal injuries.' 'I miss them being alive'
7	'I think the 90 day rule applies to increases over 5%?' 'So this means the people who have debt can see those that don’t. Am I sensing an easier target for muggings and such?'
1	"Aww... she'll probably come around eventually, I'm sure she was just jealous of [NAME]... I mean, what woman wouldn't be! lol " 'And then they say, “HAHAHAHHA IT WAS RIGHT THERE WOW!”' 'just noticed, lol. damn pervert foreigners.'
25	'my brain hurts...' 'Pretty sure I’ve seen this. He swings away with the harness he is wearing. Still looks painful but I think he lives' 'sorry [NAME]! 😘😘😘'
15	'Super, thanks' 'Thank you friend' 'Yes I heard abt the f bombs! That has to be why. Thanks for your reply:) until then hubby and I will anxiously wait 😝'
18	'I love Rocket Love and Blasted. I just wonder who the songs were written for because these are all reference tracks except Acura Intergul'
26	"OmG pEyToN iSn'T gOoD eNoUgH tO hElP uS iN tHe PlAyOfFs! Dumbass Broncos fans circa December 2015."
17	'Happy to be able to help.'
8	'We need more boards and to create a bit more space for [NAME]. Then we’ll be good.'
5	"R/sleeptrain Might be time for some sleep training. Take a look and try to feel out what's right for your family."
14	'To make her feel threatened'
13	'Very interesting. Thx' 'This...has 9k upvotes. Wow.'
20	"It's true though. He either gets no shirt and freezes to death or wears a stupid looking butchers cape. I hope he gets something better next season"
12	"I just shit my pants. Then walk away. Embarrassing enough he won't press or follow you." 'i got a bump and a bald spot. i feel dumb <3'
9	'He was off by 5 minutes, not impressed. '
24	'Apologies, I take it all back as I’ve just seen his latest effort'

Evaluation

Metrics

Label	Accuracy
all	0.3

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("Cheers, sololander!")

Training Details

Training Set Metrics

Training set	Min	Median	Max
Word count	2	10.6	27

Label	Training Sample Count
0	10
1	3
2	5
3	4
4	5
5	1
6	4
7	2
8	1
9	1
10	1
12	2
13	2
14	1
15	5
16	2
17	1
18	1
20	1
24	1
25	3
26	1
27	33

Training Hyperparameters

• batch_size: (32, 32)
• num_epochs: (1, 1)
• 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
• l2_weight: 0.01
• seed: 42
• eval_max_steps: -1
• load_best_model_at_end: False

Training Results

Epoch	Step	Training Loss	Validation Loss
0.0047	1	0.2551	-
0.2358	50	0.2056	-
0.4717	100	0.0522	-
0.7075	150	0.0206	-
0.9434	200	0.0154	-

Framework Versions

• Python: 3.12.12
• SetFit: 1.1.3
• Sentence Transformers: 5.1.2
• Transformers: 4.57.1
• PyTorch: 2.8.0+cu126
• Datasets: 4.0.0
• Tokenizers: 0.22.1

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}
}