metadata

library_name: setfit
tags:
  - setfit
  - sentence-transformers
  - text-classification
  - generated_from_setfit_trainer
metrics:
  - accuracy
widget:
  - text: >-
      I am hoping one of you fine cocktail connoisseurs can help me out. I am
      predominately a whisk(e)y/bourbon drinker. I love rye and the spice notes
      in it. I don’t find myself drinking much whiskey in summertime just due to
      my lack of imagination with it. I lean towards Jalapeño margaritas or
      Mezcal drinks to get the spicy notes. Does anyone have or recommend a good
      “spicy” Whiskey version or summertime cocktail? Thanks in advance to all
      of you!
  - text: >-
      Something simple for the glorious weather we are having.  Irish Mule. JJ
      Corry The Hanson,Ginger Beer and a Lime. I found the ginger notes from The
      Hanson really complemented the Ginger Beer and lime. Sláinte #whiskey
      #cocktail #jjcorry #ginger #beer #lime #tasty #drink #drinkaware #sláinte
  - text: >-
      Trying the Jameson Black Barrel today. Wow! Decidedly different from their
      regular whiskey and I like it!!!!! So smooth. I might have found a second
      favorite ( Basil-Hayden bourbon being my first. Crazy when the cheaper
      stuff is actually better than some of the more expensive. Although it
      wasn’t “ cheap”, it was more than the regular Jameson but affordable
  - text: >-
      lol don’t we all. What is your favorite drink? I’m ok crown royal with
      peach and sweet tea lol my friend got me on it
  - text: >-
      a hot toddy is a generalized Midwestern us drink.... it's used mostly as a
      medicine... recipes very but the general one that I know of at least is
      hot tea, whiskey, and honey
pipeline_tag: text-classification
inference: true
base_model: sentence-transformers/paraphrase-mpnet-base-v2

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:

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/paraphrase-mpnet-base-v2
Classification head: a LogisticRegression instance
Maximum Sequence Length: 512 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
1	'Jim bean fire is my fireball replacement 10x better than fireball.' "I drank a local Bourbon last night to celebrate National Bourbon day. I've been trying to get into whiskey for a few years now and am starting to appreciate it more than I used to. But I just discovered Islay Scotch and am in love. I think i'd had Scotch Whisky in the past and didn't think much of it. And I'm just at Ardbeg and Laphroig so far" 'makers mark is damn good whisky though makes want a whisky sour now'
0	"I've always liked Jack Daniels mixed with tea, punch, lemonade or coke. Great drink on a hot summers day." "I usually use rye as my preference (I like the slightly more spicy flavor), but to be honest I'll use either depending on what is available and have also made a good one using a smokey peated whisky. Good use of Wild Turkey which is a great whiskey. Worth going that little bit further and getting their 101 for bourbon or rye as the extra proof makes it so good for cocktails. Luxardo cherries are also so worth the money. A variation on the cocktail I love that was inspired by an Amsterdam restaurant is to use popcorn flavoured syrup and chocolate bitters. The chocolate and the popcorn really work well together." 'lol don’t we all. What is your favorite drink? I’m ok crown royal with peach and sweet tea lol my friend got me on it'

Label

Examples

'Jim bean fire is my fireball replacement 10x better than fireball.'
"I drank a local Bourbon last night to celebrate National Bourbon day. I've been trying to get into whiskey for a few years now and am starting to appreciate it more than I used to. But I just discovered Islay Scotch and am in love. I think i'd had Scotch Whisky in the past and didn't think much of it. And I'm just at Ardbeg and Laphroig so far"
'makers mark is damn good whisky though makes want a whisky sour now'

"I've always liked Jack Daniels mixed with tea, punch, lemonade or coke. Great drink on a hot summers day."
"I usually use rye as my preference (I like the slightly more spicy flavor), but to be honest I'll use either depending on what is available and have also made a good one using a smokey peated whisky. Good use of Wild Turkey which is a great whiskey. Worth going that little bit further and getting their 101 for bourbon or rye as the extra proof makes it so good for cocktails. Luxardo cherries are also so worth the money. A variation on the cocktail I love that was inspired by an Amsterdam restaurant is to use popcorn flavoured syrup and chocolate bitters. The chocolate and the popcorn really work well together."
'lol don’t we all. What is your favorite drink? I’m ok crown royal with peach and sweet tea lol my friend got me on it'

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("bhaskars113/whiskey-recipe-model")
# Run inference
preds = model("lol don’t we all. What is your favorite drink? I’m ok crown royal with peach and sweet tea lol my friend got me on it")

Training Details

Training Set Metrics

Training set	Min	Median	Max
Word count	12	57.8438	152

Label	Training Sample Count
0	16
1	16

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
seed: 42
eval_max_steps: -1
load_best_model_at_end: False

Training Results

Epoch	Step	Training Loss	Validation Loss
0.0125	1	0.1981	-
0.625	50	0.0005	-

Framework Versions

Python: 3.10.12
SetFit: 1.0.3
Sentence Transformers: 2.5.1
Transformers: 4.38.2
PyTorch: 2.1.0+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}
}