Hugging Face's logo

firqaaa
/
indo-setfit-bert-base-p1

Text Classification
setfit
Safetensors
sentence-transformers
bert
generated_from_setfit_trainer
Eval Results
Model card Files Community
indo-setfit-bert-base-p1 / README.md
firqaaa's picture
firqaaa
Add SetFit model
5b6b921 verified
preview code
|
raw
history blame
No virus
15.6 kB
metadata
library_name: setfit
tags:
  - setfit
  - sentence-transformers
  - text-classification
  - generated_from_setfit_trainer
metrics:
  - accuracy
widget:
  - text: >-
      Ini adalah kisah tentang dua orang yang tidak selaras dan tidak memiliki
      kesempatan sendirian, tetapi bersama-sama mereka luar biasa.
  - text: >-
      ia tidak percaya pada dirinya sendiri, ia tidak memiliki rasa humor ... ia
      hanya merasa bosan.
  - text: >-
      Keberanian band dalam menghadapi represi resmi sangat menginspirasi,
      terutama bagi para hippie yang telah menua (termasuk saya sendiri).
  - text: film yang cepat, lucu, dan sangat menghibur.
  - text: >-
      film ini mencapai dampak yang sama besar dengan menyimpan
      pemikiran-pemikiran ini tersembunyi seperti halnya film "Quills" yang
      menunjukkannya.
pipeline_tag: text-classification
inference: true
base_model: firqaaa/indo-sentence-bert-base
model-index:
  - name: SetFit with firqaaa/indo-sentence-bert-base
    results:
      - task:
          type: text-classification
          name: Text Classification
        dataset:
          name: Unknown
          type: unknown
          split: test
        metrics:
          - type: accuracy
            value: 0.8
            name: Accuracy

SetFit with firqaaa/indo-sentence-bert-base

This is a SetFit model that can be used for Text Classification. This SetFit model uses firqaaa/indo-sentence-bert-base 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 Sources

Model Labels

Label Examples
positif
  • 'benar-benar lucu'
  • 'gulungan dari sebuah tong tong yang tersesat'
  • ', mereka menemukan rute-rute baru melalui lingkungan yang sudah dikenal'
negatif
  • 'tidak menarik atau berbau tidak sedap'
  • "telah melakukan kesalahan nyaris fatal dengan menjadi apa yang orang Inggris sebut 'terlalu pintar setengah mati'."
  • 'untuk roboh'

Evaluation

Metrics

Label Accuracy
all 0.8

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("firqaaa/indo-setfit-bert-base-p1")
# Run inference
preds = model("film yang cepat, lucu, dan sangat menghibur.")

Training Details

Training Set Metrics

Training set Min Median Max
Word count 1 9.4825 51
Label Training Sample Count
negatif 200
positif 200

Training Hyperparameters

  • batch_size: (32, 32)
  • 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
0.0004 1 0.3079 -
0.0199 50 0.3644 -
0.0398 100 0.2816 -
0.0597 150 0.2254 -
0.0796 200 0.1798 -
0.0995 250 0.0478 -
0.1194 300 0.0049 -
0.1393 350 0.0016 -
0.1592 400 0.0011 -
0.1791 450 0.0005 -
0.1990 500 0.0003 -
0.2189 550 0.0004 -
0.2388 600 0.0003 -
0.2587 650 0.0003 -
0.2786 700 0.0001 -
0.2984 750 0.0002 -
0.3183 800 0.0001 -
0.3382 850 0.0001 -
0.3581 900 0.0001 -
0.3780 950 0.0001 -
0.3979 1000 0.0001 -
0.4178 1050 0.0001 -
0.4377 1100 0.0001 -
0.4576 1150 0.0001 -
0.4775 1200 0.0001 -
0.4974 1250 0.0001 -
0.5173 1300 0.0001 -
0.5372 1350 0.0001 -
0.5571 1400 0.0001 -
0.5770 1450 0.0001 -
0.5969 1500 0.0001 -
0.6168 1550 0.0001 -
0.6367 1600 0.0001 -
0.6566 1650 0.0001 -
0.6765 1700 0.0002 -
0.6964 1750 0.0001 -
0.7163 1800 0.0001 -
0.7362 1850 0.0001 -
0.7561 1900 0.0001 -
0.7760 1950 0.0001 -
0.7959 2000 0.0001 -
0.8158 2050 0.0001 -
0.8357 2100 0.0001 -
0.8556 2150 0.0001 -
0.8754 2200 0.0001 -
0.8953 2250 0.0 -
0.9152 2300 0.0001 -
0.9351 2350 0.0 -
0.9550 2400 0.0 -
0.9749 2450 0.0 -
0.9948 2500 0.0 -
1.0 2513 - 0.2622
1.0147 2550 0.0 -
1.0346 2600 0.0 -
1.0545 2650 0.0 -
1.0744 2700 0.0 -
1.0943 2750 0.0 -
1.1142 2800 0.0 -
1.1341 2850 0.0 -
1.1540 2900 0.0 -
1.1739 2950 0.0 -
1.1938 3000 0.0 -
1.2137 3050 0.0 -
1.2336 3100 0.0 -
1.2535 3150 0.0 -
1.2734 3200 0.0 -
1.2933 3250 0.0 -
1.3132 3300 0.0 -
1.3331 3350 0.0 -
1.3530 3400 0.0 -
1.3729 3450 0.0 -
1.3928 3500 0.0 -
1.4127 3550 0.0 -
1.4326 3600 0.0 -
1.4524 3650 0.0 -
1.4723 3700 0.0 -
1.4922 3750 0.0 -
1.5121 3800 0.0 -
1.5320 3850 0.0 -
1.5519 3900 0.0 -
1.5718 3950 0.0 -
1.5917 4000 0.0 -
1.6116 4050 0.0 -
1.6315 4100 0.0 -
1.6514 4150 0.0 -
1.6713 4200 0.0 -
1.6912 4250 0.0 -
1.7111 4300 0.0 -
1.7310 4350 0.0 -
1.7509 4400 0.0 -
1.7708 4450 0.0 -
1.7907 4500 0.0 -
1.8106 4550 0.0 -
1.8305 4600 0.0 -
1.8504 4650 0.0 -
1.8703 4700 0.0 -
1.8902 4750 0.0 -
1.9101 4800 0.0 -
1.9300 4850 0.0 -
1.9499 4900 0.0 -
1.9698 4950 0.0 -
1.9897 5000 0.0 -
2.0 5026 - 0.269
2.0096 5050 0.0 -
2.0294 5100 0.0 -
2.0493 5150 0.0 -
2.0692 5200 0.0 -
2.0891 5250 0.0 -
2.1090 5300 0.0 -
2.1289 5350 0.0 -
2.1488 5400 0.0 -
2.1687 5450 0.0 -
2.1886 5500 0.0 -
2.2085 5550 0.0 -
2.2284 5600 0.0 -
2.2483 5650 0.0 -
2.2682 5700 0.0 -
2.2881 5750 0.0 -
2.3080 5800 0.0 -
2.3279 5850 0.0 -
2.3478 5900 0.0 -
2.3677 5950 0.0 -
2.3876 6000 0.0 -
2.4075 6050 0.0 -
2.4274 6100 0.0 -
2.4473 6150 0.0 -
2.4672 6200 0.0 -
2.4871 6250 0.0 -
2.5070 6300 0.0 -
2.5269 6350 0.0 -
2.5468 6400 0.0 -
2.5667 6450 0.0 -
2.5865 6500 0.0 -
2.6064 6550 0.0 -
2.6263 6600 0.0 -
2.6462 6650 0.0 -
2.6661 6700 0.0 -
2.6860 6750 0.0 -
2.7059 6800 0.0 -
2.7258 6850 0.0 -
2.7457 6900 0.0 -
2.7656 6950 0.0 -
2.7855 7000 0.0 -
2.8054 7050 0.0 -
2.8253 7100 0.0 -
2.8452 7150 0.0 -
2.8651 7200 0.0 -
2.8850 7250 0.0 -
2.9049 7300 0.0 -
2.9248 7350 0.0 -
2.9447 7400 0.0 -
2.9646 7450 0.0 -
2.9845 7500 0.0 -
3.0 7539 - 0.2744
  • The bold row denotes the saved checkpoint.

Framework Versions

  • Python: 3.10.13
  • SetFit: 1.0.3
  • Sentence Transformers: 2.3.1
  • Transformers: 4.36.2
  • PyTorch: 2.1.2+cu121
  • Datasets: 2.16.1
  • Tokenizers: 0.15.0

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}
}