Hugging Face's logo

Vishal24
/
BCG-classifier

Text Classification
setfit
Safetensors
sentence-transformers
bert
generated_from_setfit_trainer
Eval Results
Model card Files Community
BCG-classifier / README.md
Vishal24's picture
Vishal24
Push model using huggingface_hub.
1f4e243 verified
preview code
|
raw
history blame contribute delete
34 kB
metadata
base_model: Vishal24/bert-1ds-domain
datasets:
  - Vishal24/BCG_classifier
library_name: setfit
metrics:
  - f1
pipeline_tag: text-classification
tags:
  - setfit
  - sentence-transformers
  - text-classification
  - generated_from_setfit_trainer
widget:
  - text: fair and handsome 100 oil clear face wash
  - text: hazelnut
  - text: aqualohica body mist
  - text: joy body lotion 300 ml
  - text: top of browse listings page
inference: true
model-index:
  - name: SetFit with Vishal24/bert-1ds-domain
    results:
      - task:
          type: text-classification
          name: Text Classification
        dataset:
          name: Vishal24/BCG_classifier
          type: Vishal24/BCG_classifier
          split: test
        metrics:
          - type: f1
            value: 0.9233278955954323
            name: F1

SetFit with Vishal24/bert-1ds-domain

This is a SetFit model trained on the Vishal24/BCG_classifier dataset that can be used for Text Classification. This SetFit model uses Vishal24/bert-1ds-domain as the Sentence Transformer embedding model. A SetFitHead 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
0
  • 'mois'
  • 'time skincare soap'
  • 'paraben free'
1
  • 'tomato ketchup 1kg flipkart'
  • 'sunsilk keratin yogurt shampoo lusciously thick long'
  • 'wow aloevera soap'

Evaluation

Metrics

Label F1
all 0.9233

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("Vishal24/BCG-classifier")
# Run inference
preds = model("hazelnut")

Training Details

Training Set Metrics

Training set Min Median Max
Word count 1 3.4474 19
Label Training Sample Count
0 2252
1 1262

Training Hyperparameters

  • batch_size: (16, 2)
  • num_epochs: (3, 3)
  • max_steps: -1
  • sampling_strategy: oversampling
  • num_iterations: 20
  • 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: False

Training Results

Epoch Step Training Loss Validation Loss
0.0001 1 0.2765 -
0.0057 50 0.2529 -
0.0114 100 0.252 -
0.0171 150 0.2657 -
0.0228 200 0.2735 -
0.0285 250 0.236 -
0.0341 300 0.2366 -
0.0398 350 0.2316 -
0.0455 400 0.185 -
0.0512 450 0.1396 -
0.0569 500 0.2137 -
0.0626 550 0.093 -
0.0683 600 0.1219 -
0.0740 650 0.0974 -
0.0797 700 0.2257 -
0.0854 750 0.0951 -
0.0911 800 0.0994 -
0.0968 850 0.0752 -
0.1024 900 0.0848 -
0.1081 950 0.015 -
0.1138 1000 0.0541 -
0.1195 1050 0.0357 -
0.1252 1100 0.0314 -
0.1309 1150 0.0557 -
0.1366 1200 0.0027 -
0.1423 1250 0.0387 -
0.1480 1300 0.0026 -
0.1537 1350 0.044 -
0.1594 1400 0.0499 -
0.1651 1450 0.001 -
0.1707 1500 0.0007 -
0.1764 1550 0.0008 -
0.1821 1600 0.0009 -
0.1878 1650 0.053 -
0.1935 1700 0.1111 -
0.1992 1750 0.0018 -
0.2049 1800 0.0009 -
0.2106 1850 0.0008 -
0.2163 1900 0.0011 -
0.2220 1950 0.0042 -
0.2277 2000 0.0005 -
0.2334 2050 0.0023 -
0.2390 2100 0.0003 -
0.2447 2150 0.0004 -
0.2504 2200 0.055 -
0.2561 2250 0.0584 -
0.2618 2300 0.06 -
0.2675 2350 0.0004 -
0.2732 2400 0.0022 -
0.2789 2450 0.0005 -
0.2846 2500 0.0014 -
0.2903 2550 0.0008 -
0.2960 2600 0.0004 -
0.3017 2650 0.0118 -
0.3073 2700 0.0892 -
0.3130 2750 0.0004 -
0.3187 2800 0.0061 -
0.3244 2850 0.0601 -
0.3301 2900 0.0003 -
0.3358 2950 0.0007 -
0.3415 3000 0.0006 -
0.3472 3050 0.0002 -
0.3529 3100 0.0002 -
0.3586 3150 0.0005 -
0.3643 3200 0.0003 -
0.3699 3250 0.0002 -
0.3756 3300 0.0008 -
0.3813 3350 0.0002 -
0.3870 3400 0.0513 -
0.3927 3450 0.0003 -
0.3984 3500 0.0002 -
0.4041 3550 0.0006 -
0.4098 3600 0.0005 -
0.4155 3650 0.0003 -
0.4212 3700 0.0002 -
0.4269 3750 0.0002 -
0.4326 3800 0.0005 -
0.4382 3850 0.0001 -
0.4439 3900 0.0002 -
0.4496 3950 0.0001 -
0.4553 4000 0.0003 -
0.4610 4050 0.0001 -
0.4667 4100 0.0595 -
0.4724 4150 0.0002 -
0.4781 4200 0.0001 -
0.4838 4250 0.0002 -
0.4895 4300 0.0001 -
0.4952 4350 0.0002 -
0.5009 4400 0.0001 -
0.5065 4450 0.0001 -
0.5122 4500 0.0002 -
0.5179 4550 0.0001 -
0.5236 4600 0.0014 -
0.5293 4650 0.0001 -
0.5350 4700 0.0001 -
0.5407 4750 0.0002 -
0.5464 4800 0.0001 -
0.5521 4850 0.0419 -
0.5578 4900 0.0001 -
0.5635 4950 0.0001 -
0.5692 5000 0.0001 -
0.5748 5050 0.0001 -
0.5805 5100 0.0001 -
0.5862 5150 0.0001 -
0.5919 5200 0.0001 -
0.5976 5250 0.0001 -
0.6033 5300 0.0001 -
0.6090 5350 0.0001 -
0.6147 5400 0.0 -
0.6204 5450 0.0 -
0.6261 5500 0.0001 -
0.6318 5550 0.0 -
0.6375 5600 0.0001 -
0.6431 5650 0.0001 -
0.6488 5700 0.0006 -
0.6545 5750 0.0001 -
0.6602 5800 0.0001 -
0.6659 5850 0.0001 -
0.6716 5900 0.0001 -
0.6773 5950 0.0001 -
0.6830 6000 0.0002 -
0.6887 6050 0.0002 -
0.6944 6100 0.0001 -
0.7001 6150 0.0001 -
0.7057 6200 0.0001 -
0.7114 6250 0.0 -
0.7171 6300 0.0001 -
0.7228 6350 0.0001 -
0.7285 6400 0.0001 -
0.7342 6450 0.0001 -
0.7399 6500 0.0002 -
0.7456 6550 0.0001 -
0.7513 6600 0.0001 -
0.7570 6650 0.0 -
0.7627 6700 0.0001 -
0.7684 6750 0.0001 -
0.7740 6800 0.0001 -
0.7797 6850 0.0003 -
0.7854 6900 0.0515 -
0.7911 6950 0.0001 -
0.7968 7000 0.0003 -
0.8025 7050 0.0001 -
0.8082 7100 0.0001 -
0.8139 7150 0.0001 -
0.8196 7200 0.0 -
0.8253 7250 0.0001 -
0.8310 7300 0.0 -
0.8367 7350 0.0001 -
0.8423 7400 0.0001 -
0.8480 7450 0.0001 -
0.8537 7500 0.0001 -
0.8594 7550 0.0 -
0.8651 7600 0.0 -
0.8708 7650 0.0 -
0.8765 7700 0.0 -
0.8822 7750 0.0014 -
0.8879 7800 0.0001 -
0.8936 7850 0.0001 -
0.8993 7900 0.0 -
0.9050 7950 0.0001 -
0.9106 8000 0.0002 -
0.9163 8050 0.0001 -
0.9220 8100 0.0 -
0.9277 8150 0.0 -
0.9334 8200 0.0001 -
0.9391 8250 0.0 -
0.9448 8300 0.0001 -
0.9505 8350 0.0004 -
0.9562 8400 0.0001 -
0.9619 8450 0.0 -
0.9676 8500 0.001 -
0.9732 8550 0.0001 -
0.9789 8600 0.0001 -
0.9846 8650 0.0 -
0.9903 8700 0.0 -
0.9960 8750 0.0001 -
1.0017 8800 0.0002 -
1.0074 8850 0.0 -
1.0131 8900 0.0 -
1.0188 8950 0.0 -
1.0245 9000 0.0001 -
1.0302 9050 0.0 -
1.0359 9100 0.0 -
1.0415 9150 0.0 -
1.0472 9200 0.0 -
1.0529 9250 0.0 -
1.0586 9300 0.0 -
1.0643 9350 0.0 -
1.0700 9400 0.0001 -
1.0757 9450 0.0 -
1.0814 9500 0.0 -
1.0871 9550 0.0 -
1.0928 9600 0.0 -
1.0985 9650 0.0 -
1.1042 9700 0.0001 -
1.1098 9750 0.0002 -
1.1155 9800 0.0097 -
1.1212 9850 0.0 -
1.1269 9900 0.0 -
1.1326 9950 0.0001 -
1.1383 10000 0.0 -
1.1440 10050 0.0 -
1.1497 10100 0.0001 -
1.1554 10150 0.0004 -
1.1611 10200 0.0 -
1.1668 10250 0.0 -
1.1725 10300 0.0 -
1.1781 10350 0.0 -
1.1838 10400 0.0001 -
1.1895 10450 0.0 -
1.1952 10500 0.0 -
1.2009 10550 0.0 -
1.2066 10600 0.0 -
1.2123 10650 0.0 -
1.2180 10700 0.0001 -
1.2237 10750 0.0 -
1.2294 10800 0.0 -
1.2351 10850 0.0001 -
1.2408 10900 0.0305 -
1.2464 10950 0.0617 -
1.2521 11000 0.0 -
1.2578 11050 0.0 -
1.2635 11100 0.0 -
1.2692 11150 0.0 -
1.2749 11200 0.0 -
1.2806 11250 0.0 -
1.2863 11300 0.0 -
1.2920 11350 0.0 -
1.2977 11400 0.0 -
1.3034 11450 0.0 -
1.3090 11500 0.0 -
1.3147 11550 0.0 -
1.3204 11600 0.0 -
1.3261 11650 0.0 -
1.3318 11700 0.0 -
1.3375 11750 0.0 -
1.3432 11800 0.0 -
1.3489 11850 0.0 -
1.3546 11900 0.0 -
1.3603 11950 0.0 -
1.3660 12000 0.0 -
1.3717 12050 0.0 -
1.3773 12100 0.0 -
1.3830 12150 0.0 -
1.3887 12200 0.0 -
1.3944 12250 0.0 -
1.4001 12300 0.0 -
1.4058 12350 0.0 -
1.4115 12400 0.0 -
1.4172 12450 0.0 -
1.4229 12500 0.0 -
1.4286 12550 0.0 -
1.4343 12600 0.0 -
1.4400 12650 0.0 -
1.4456 12700 0.0 -
1.4513 12750 0.0 -
1.4570 12800 0.0 -
1.4627 12850 0.0 -
1.4684 12900 0.0 -
1.4741 12950 0.0 -
1.4798 13000 0.0 -
1.4855 13050 0.0 -
1.4912 13100 0.0 -
1.4969 13150 0.0001 -
1.5026 13200 0.0 -
1.5083 13250 0.0 -
1.5139 13300 0.0 -
1.5196 13350 0.0 -
1.5253 13400 0.0 -
1.5310 13450 0.0 -
1.5367 13500 0.0001 -
1.5424 13550 0.0 -
1.5481 13600 0.0 -
1.5538 13650 0.0 -
1.5595 13700 0.0001 -
1.5652 13750 0.0001 -
1.5709 13800 0.0 -
1.5766 13850 0.0001 -
1.5822 13900 0.0 -
1.5879 13950 0.0 -
1.5936 14000 0.0 -
1.5993 14050 0.0 -
1.6050 14100 0.0 -
1.6107 14150 0.0 -
1.6164 14200 0.0 -
1.6221 14250 0.0 -
1.6278 14300 0.0 -
1.6335 14350 0.0 -
1.6392 14400 0.0 -
1.6448 14450 0.0 -
1.6505 14500 0.0 -
1.6562 14550 0.0 -
1.6619 14600 0.0 -
1.6676 14650 0.0 -
1.6733 14700 0.0 -
1.6790 14750 0.0 -
1.6847 14800 0.0 -
1.6904 14850 0.0 -
1.6961 14900 0.0 -
1.7018 14950 0.0 -
1.7075 15000 0.0 -
1.7131 15050 0.0 -
1.7188 15100 0.0 -
1.7245 15150 0.0001 -
1.7302 15200 0.0 -
1.7359 15250 0.0 -
1.7416 15300 0.0002 -
1.7473 15350 0.0 -
1.7530 15400 0.0 -
1.7587 15450 0.0 -
1.7644 15500 0.0 -
1.7701 15550 0.0 -
1.7758 15600 0.0 -
1.7814 15650 0.0 -
1.7871 15700 0.0 -
1.7928 15750 0.0 -
1.7985 15800 0.0 -
1.8042 15850 0.0 -
1.8099 15900 0.0 -
1.8156 15950 0.0 -
1.8213 16000 0.0 -
1.8270 16050 0.0 -
1.8327 16100 0.0 -
1.8384 16150 0.0001 -
1.8441 16200 0.0 -
1.8497 16250 0.0 -
1.8554 16300 0.0 -
1.8611 16350 0.0 -
1.8668 16400 0.0 -
1.8725 16450 0.0 -
1.8782 16500 0.0 -
1.8839 16550 0.0 -
1.8896 16600 0.0 -
1.8953 16650 0.0 -
1.9010 16700 0.0 -
1.9067 16750 0.0 -
1.9124 16800 0.0 -
1.9180 16850 0.0 -
1.9237 16900 0.0 -
1.9294 16950 0.0 -
1.9351 17000 0.0 -
1.9408 17050 0.0 -
1.9465 17100 0.0 -
1.9522 17150 0.0 -
1.9579 17200 0.0 -
1.9636 17250 0.0 -
1.9693 17300 0.0 -
1.9750 17350 0.0 -
1.9806 17400 0.0 -
1.9863 17450 0.0 -
1.9920 17500 0.0 -
1.9977 17550 0.0 -
2.0034 17600 0.0 -
2.0091 17650 0.0 -
2.0148 17700 0.0 -
2.0205 17750 0.0 -
2.0262 17800 0.0 -
2.0319 17850 0.0523 -
2.0376 17900 0.0 -
2.0433 17950 0.0 -
2.0489 18000 0.0 -
2.0546 18050 0.0 -
2.0603 18100 0.0 -
2.0660 18150 0.0 -
2.0717 18200 0.0 -
2.0774 18250 0.0 -
2.0831 18300 0.0 -
2.0888 18350 0.0 -
2.0945 18400 0.0 -
2.1002 18450 0.0 -
2.1059 18500 0.0 -
2.1116 18550 0.0 -
2.1172 18600 0.0 -
2.1229 18650 0.0 -
2.1286 18700 0.0 -
2.1343 18750 0.0 -
2.1400 18800 0.0 -
2.1457 18850 0.0 -
2.1514 18900 0.0 -
2.1571 18950 0.0 -
2.1628 19000 0.0 -
2.1685 19050 0.0 -
2.1742 19100 0.0 -
2.1799 19150 0.0 -
2.1855 19200 0.0 -
2.1912 19250 0.0 -
2.1969 19300 0.0 -
2.2026 19350 0.0 -
2.2083 19400 0.0 -
2.2140 19450 0.0 -
2.2197 19500 0.0 -
2.2254 19550 0.0 -
2.2311 19600 0.0 -
2.2368 19650 0.0 -
2.2425 19700 0.0 -
2.2482 19750 0.0 -
2.2538 19800 0.0 -
2.2595 19850 0.0 -
2.2652 19900 0.0 -
2.2709 19950 0.0 -
2.2766 20000 0.0 -
2.2823 20050 0.0 -
2.2880 20100 0.0 -
2.2937 20150 0.0 -
2.2994 20200 0.0 -
2.3051 20250 0.0 -
2.3108 20300 0.0 -
2.3164 20350 0.0 -
2.3221 20400 0.0 -
2.3278 20450 0.0 -
2.3335 20500 0.0 -
2.3392 20550 0.0 -
2.3449 20600 0.0 -
2.3506 20650 0.0 -
2.3563 20700 0.0 -
2.3620 20750 0.0 -
2.3677 20800 0.0 -
2.3734 20850 0.0 -
2.3791 20900 0.0 -
2.3847 20950 0.0 -
2.3904 21000 0.0 -
2.3961 21050 0.0 -
2.4018 21100 0.0 -
2.4075 21150 0.0 -
2.4132 21200 0.0 -
2.4189 21250 0.0 -
2.4246 21300 0.0 -
2.4303 21350 0.0 -
2.4360 21400 0.0 -
2.4417 21450 0.0 -
2.4474 21500 0.0 -
2.4530 21550 0.0 -
2.4587 21600 0.0 -
2.4644 21650 0.0 -
2.4701 21700 0.0 -
2.4758 21750 0.0 -
2.4815 21800 0.0 -
2.4872 21850 0.0 -
2.4929 21900 0.0 -
2.4986 21950 0.0 -
2.5043 22000 0.0 -
2.5100 22050 0.0 -
2.5157 22100 0.0 -
2.5213 22150 0.0 -
2.5270 22200 0.0 -
2.5327 22250 0.0 -
2.5384 22300 0.0 -
2.5441 22350 0.0 -
2.5498 22400 0.0 -
2.5555 22450 0.0 -
2.5612 22500 0.0 -
2.5669 22550 0.0 -
2.5726 22600 0.0 -
2.5783 22650 0.0 -
2.5839 22700 0.0 -
2.5896 22750 0.0 -
2.5953 22800 0.0 -
2.6010 22850 0.0 -
2.6067 22900 0.0 -
2.6124 22950 0.0 -
2.6181 23000 0.0 -
2.6238 23050 0.0 -
2.6295 23100 0.0 -
2.6352 23150 0.0 -
2.6409 23200 0.0 -
2.6466 23250 0.0 -
2.6522 23300 0.0 -
2.6579 23350 0.0 -
2.6636 23400 0.0 -
2.6693 23450 0.0 -
2.6750 23500 0.0 -
2.6807 23550 0.0 -
2.6864 23600 0.0 -
2.6921 23650 0.0 -
2.6978 23700 0.0 -
2.7035 23750 0.0 -
2.7092 23800 0.0 -
2.7149 23850 0.0 -
2.7205 23900 0.0 -
2.7262 23950 0.0 -
2.7319 24000 0.0 -
2.7376 24050 0.0 -
2.7433 24100 0.0 -
2.7490 24150 0.0 -
2.7547 24200 0.0 -
2.7604 24250 0.0 -
2.7661 24300 0.0 -
2.7718 24350 0.0 -
2.7775 24400 0.0 -
2.7832 24450 0.0 -
2.7888 24500 0.0 -
2.7945 24550 0.0 -
2.8002 24600 0.0 -
2.8059 24650 0.0 -
2.8116 24700 0.0 -
2.8173 24750 0.0 -
2.8230 24800 0.0 -
2.8287 24850 0.0 -
2.8344 24900 0.0 -
2.8401 24950 0.0 -
2.8458 25000 0.0 -
2.8515 25050 0.0 -
2.8571 25100 0.0 -
2.8628 25150 0.0 -
2.8685 25200 0.0 -
2.8742 25250 0.0 -
2.8799 25300 0.0 -
2.8856 25350 0.0 -
2.8913 25400 0.0 -
2.8970 25450 0.0 -
2.9027 25500 0.0 -
2.9084 25550 0.0 -
2.9141 25600 0.0 -
2.9197 25650 0.0 -
2.9254 25700 0.0 -
2.9311 25750 0.0 -
2.9368 25800 0.0 -
2.9425 25850 0.0 -
2.9482 25900 0.0 -
2.9539 25950 0.0 -
2.9596 26000 0.0 -
2.9653 26050 0.0 -
2.9710 26100 0.0 -
2.9767 26150 0.0 -
2.9824 26200 0.0 -
2.9880 26250 0.0 -
2.9937 26300 0.0 -
2.9994 26350 0.0 -

Framework Versions

  • Python: 3.10.12
  • SetFit: 1.0.3
  • Sentence Transformers: 3.3.1
  • Transformers: 4.41.2
  • PyTorch: 2.1.0+cu118
  • Datasets: 2.20.0
  • 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}
}