README.md

metadata

tags:
  - setfit
  - sentence-transformers
  - text-classification
  - generated_from_setfit_trainer
widget:
  - text: I need some modern design assets
  - text: Reduce the letter spacing
  - text: Group the menu items
  - text: Create a photo of a delicious pizza
  - text: Mirror the icon horizontally
metrics:
  - accuracy
pipeline_tag: text-classification
library_name: setfit
inference: true
base_model: nomic-ai/nomic-embed-text-v1.5
model-index:
  - name: SetFit with nomic-ai/nomic-embed-text-v1.5
    results:
      - task:
          type: text-classification
          name: Text Classification
        dataset:
          name: Unknown
          type: unknown
          split: test
        metrics:
          - type: accuracy
            value: 0.4163860830527497
            name: Accuracy

SetFit with nomic-ai/nomic-embed-text-v1.5

This is a SetFit model that can be used for Text Classification. This SetFit model uses nomic-ai/nomic-embed-text-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: nomic-ai/nomic-embed-text-v1.5
• Classification head: a LogisticRegression instance
• Maximum Sequence Length: 8192 tokens
• Number of Classes: 63 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
0	'Add a corporate presentation background' 'Insert a modern icon set for the design'
1	'Show me some background options for my design' 'I need some modern design assets'
2	'Add a footer with contact information' 'Insert a descriptive caption for the image'
3	'Align the shapes to the top' 'Right align the buttons'
4	'Make the image slide from left' 'Make the image bounce in'
5	'Add a spiral animation to the page' 'Make the page bounce in from the top'
6	'Take out the cat and put a dog in the image' "Add glasses to the person's face"
7	'Delete the test element' 'Remove the unused icon'
8	'How do I change the font size?' 'How can I duplicate elements?'
9	'Distribute the buttons around the center image' 'Arrange the shapes in a circle pattern'
10	'Create a variation of this design' 'Make a copy for editing'
11	'Duplicate the shape and change its color' 'Copy the image and make it smaller'
12	'Copy the text to the last page' 'Duplicate the footer to page 2'
13	'Fix the text layout' 'Improve the overall typography'
14	'Mirror the icon horizontally' 'Flip the shape vertically'
15	'Create a photo of a beautiful garden' 'Create a photo of a delicious pizza'
16	'Generate an Instagram post for a birthday' 'Generate a Facebook post for a birthday'
17	'Group the icons together' 'Group the menu items'
18	'Position the image at the top right' 'Move the image to the center'
19	'Add a dreamy effect' 'Make the image pop with color'
20	'Suggest some shape designs' 'Suggest some button designs'
21	'Redo the filter effect' 'Bring back the original position'
22	'Remove the background from the icon' 'Remove the background from the logo'
23	'Delete the watermark' 'Remove the graffiti'
24	'Change the decorative element' 'Change the background image'
25	'Update the body text' 'Change the subtitle'
26	'Remove all filters from the image' 'Restore the original brightness'
27	'Reduce the shape size' 'Increase the image size'
28	'Change to poster size' 'Resize to Facebook post dimensions'
29	'Rotate the element 60 degrees' 'Rotate the image 90 degrees'
30	'Randomly place the icons on the page' 'Randomly place the shapes on the page'
31	'Select the decorative elements' 'Choose the heading text'
32	'Make the background transparent' 'Change to a dark background'
33	'Change the blend mode to screen' 'Set the blend mode to darken'
34	'Add a gaussian blur' 'Blur the background elements'
35	'Add a gradient border' 'Change the border opacity'
36	'Make the image more brilliant' 'Increase the brightness contrast'
37	'Send the background to the back' 'Send the watermark to the back'
38	'Make the image more dramatic' 'Make the image more defined'
39	'Make the image square' 'Crop the image to a circle'
40	'Add a directional shadow' 'Create a floating shadow'
41	'Change the color to silver' 'Change the button color to purple'
42	'Increase the body text size' 'Increase the heading size'
43	'Add strikethrough to the text' 'Add strikethrough to the label'
44	'Use a serif font for the heading' 'Change to a clean font'
45	'Brighten the light areas' 'Enhance the bright spots'
46	'Set the image as full-page background' 'Make the image the background layer'
47	'Reduce the letter spacing' 'Add more space between letters'
48	'Make the text more compact' 'Reduce the line height'
49	'Increase the transparency of the element' 'Make the element semi-transparent'
50	'Add spacing between text blocks' 'Make the paragraphs more compact'
51	'Reduce the color richness' 'Reduce the color vibrancy'
52	'Make the shadows more prominent' 'Increase the shadow intensity'
53	'Increase the image resolution' 'Make the details more sharp'
54	'Right align the caption' 'Center the button text'
55	'Create a decorative border' 'Create a border around the text'
56	'Make the text flow in an arc' 'Create text that follows a circle'
57	'Convert the text to a list format' 'Make the text into a numbered sequence'
58	'Add a dramatic text shadow' 'Add a colored glow to the text'
59	'Make the image more amber' 'Increase the warm color balance'
60	'Open the file upload dialog' 'I need to upload a logo'
61	'Go back to the original position' 'Go back to the original text'
62	'Separate the combined elements' 'Ungroup the merged elements'

Evaluation

Metrics

Label	Accuracy
all	0.4164

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("Group the menu items")

Training Details

Training Set Metrics

Training set	Min	Median	Max
Word count	3	5.2302	11

Label	Training Sample Count
0	2
1	2
2	2
3	2
4	2
5	2
6	2
7	2
8	2
9	2
10	2
11	2
12	2
13	2
14	2
15	2
16	2
17	2
18	2
19	2
20	2
21	2
22	2
23	2
24	2
25	2
26	2
27	2
28	2
29	2
30	2
31	2
32	2
33	2
34	2
35	2
36	2
37	2
38	2
39	2
40	2
41	2
42	2
43	2
44	2
45	2
46	2
47	2
48	2
49	2
50	2
51	2
52	2
53	2
54	2
55	2
56	2
57	2
58	2
59	2
60	2
61	2
62	2

Training Hyperparameters

• batch_size: (64, 64)
• 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
0.0041	1	0.1097	-
0.2041	50	0.0477	-
0.4082	100	0.0106	-
0.6122	150	0.0045	-
0.8163	200	0.0032	-

Framework Versions

• Python: 3.12.11
• SetFit: 1.1.3
• Sentence Transformers: 5.1.0
• Transformers: 4.54.1
• PyTorch: 2.7.1
• Datasets: 4.0.0
• Tokenizers: 0.21.4

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}
}