Hugging Face's logo

diwank
/
hn-upvote-classifier

Text Classification
setfit
Safetensors
sentence-transformers
new
generated_from_setfit_trainer
custom_code
Model card Files Community
hn-upvote-classifier / README.md
diwank's picture
diwank
Push model using huggingface_hub.
fa37eec verified
preview code
|
raw
history blame
41.7 kB
metadata
base_model: Alibaba-NLP/gte-base-en-v1.5
datasets:
  - diwank/hn-upvote-data
library_name: setfit
metrics:
  - accuracy
pipeline_tag: text-classification
tags:
  - setfit
  - sentence-transformers
  - text-classification
  - generated_from_setfit_trainer
widget:
  - text: "Title: Pixar’s Rules of\_Storytelling\nSource: b'aerogrammestudio.com'"
  - text: |-
      Title: What I've learned about Open Source community over 30 years
      Source: b''
  - text: |-
      Title: My Python code is a neural network
      Source: b''
  - text: |-
      Title: The telltale words that could identify generative AI text
      Source: b''
  - text: |-
      Title: What I've learned about Open Source community over 30 years
      Source: b''
inference: true

SetFit with Alibaba-NLP/gte-base-en-v1.5

This is a SetFit model trained on the diwank/hn-upvote-data dataset that can be used for Text Classification. This SetFit model uses Alibaba-NLP/gte-base-en-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 Sources

Model Labels

Label Examples
0
  • "Title: The telltale words that could identify generative AI text\nSource: b''"
  • "Title: What I've learned about Open Source community over 30 years\nSource: b''"
  • "Title: My Python code is a neural network\nSource: b''"
1
  • "Title: Rat Park Experiment: A New Theory of Addiction\nSource: b'sub.garrytan.com'"
  • "Title: Thinking the unthinkable\nSource: b'anarchistsoccermom.blogspot.com'"
  • "Title: Realtime Analysis of the Oroville Dam Disaster\nSource: b'github.com'"

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("diwank/hn-upvote-classifier")
# Run inference
preds = model("Title: My Python code is a neural network
Source: b''")

Training Details

Training Set Metrics

Training set Min Median Max
Word count 3 10.2389 20
Label Training Sample Count
0 3302
1 1114

Training Hyperparameters

  • batch_size: (256, 16)
  • num_epochs: (1, 16)
  • max_steps: -1
  • sampling_strategy: undersampling
  • body_learning_rate: (4e-05, 2e-05)
  • head_learning_rate: 0.01
  • loss: CosineSimilarityLoss
  • distance_metric: cosine_distance
  • margin: 0.25
  • end_to_end: True
  • use_amp: True
  • warmup_proportion: 0.05
  • l2_weight: 0.2
  • seed: 42
  • eval_max_steps: -1
  • load_best_model_at_end: True

Training Results

Epoch Step Training Loss Validation Loss
0.0000 1 0.1861 -
0.0017 50 0.1334 -
0.0035 100 0.0344 -
0.0052 150 0.0048 -
0.0070 200 0.0027 -
0.0087 250 0.002 -
0.0104 300 0.0016 -
0.0122 350 0.0011 -
0.0139 400 0.001 -
0.0157 450 0.0009 -
0.0174 500 0.0008 -
0.0191 550 0.0006 -
0.0209 600 0.0006 -
0.0226 650 0.0006 -
0.0244 700 0.0005 -
0.0261 750 0.0005 -
0.0278 800 0.0004 -
0.0296 850 0.0004 -
0.0313 900 0.0004 -
0.0331 950 0.0003 -
0.0348 1000 0.0003 -
0.0365 1050 0.0003 -
0.0383 1100 0.0002 -
0.0400 1150 0.0002 -
0.0418 1200 0.0002 -
0.0435 1250 0.0002 -
0.0452 1300 0.0002 -
0.0470 1350 0.0002 -
0.0487 1400 0.0002 -
0.0505 1450 0.0001 -
0.0522 1500 0.0001 -
0.0539 1550 0.0001 -
0.0557 1600 0.0001 -
0.0574 1650 0.0001 -
0.0592 1700 0.0001 -
0.0609 1750 0.0001 -
0.0626 1800 0.0001 -
0.0644 1850 0.0001 -
0.0661 1900 0.0001 -
0.0679 1950 0.0001 -
0.0696 2000 0.0001 -
0.0713 2050 0.0001 -
0.0731 2100 0.0001 -
0.0748 2150 0.0001 -
0.0766 2200 0.0001 -
0.0783 2250 0.0001 -
0.0800 2300 0.0001 -
0.0818 2350 0.0001 -
0.0835 2400 0.0001 -
0.0853 2450 0.0001 -
0.0870 2500 0.0001 -
0.0887 2550 0.0001 -
0.0905 2600 0.0001 -
0.0922 2650 0.0001 -
0.0940 2700 0.0 -
0.0957 2750 0.0001 -
0.0974 2800 0.0001 -
0.0992 2850 0.0001 -
0.1009 2900 0.0001 -
0.1027 2950 0.0001 -
0.1044 3000 0.0001 0.0
0.1061 3050 0.0001 -
0.1079 3100 0.0001 -
0.1096 3150 0.0001 -
0.1114 3200 0.0001 -
0.1131 3250 0.0 -
0.1148 3300 0.0 -
0.1166 3350 0.0 -
0.1183 3400 0.0 -
0.1201 3450 0.0 -
0.1218 3500 0.0 -
0.1235 3550 0.0 -
0.1253 3600 0.0 -
0.1270 3650 0.0 -
0.1287 3700 0.0 -
0.1305 3750 0.0 -
0.1322 3800 0.0 -
0.1340 3850 0.0 -
0.1357 3900 0.0 -
0.1374 3950 0.0 -
0.1392 4000 0.0 -
0.1409 4050 0.0 -
0.1427 4100 0.0 -
0.1444 4150 0.0 -
0.1461 4200 0.0 -
0.1479 4250 0.0 -
0.1496 4300 0.0 -
0.1514 4350 0.0 -
0.1531 4400 0.0 -
0.1548 4450 0.0 -
0.1566 4500 0.0 -
0.1583 4550 0.0 -
0.1601 4600 0.0 -
0.1618 4650 0.0 -
0.1635 4700 0.0 -
0.1653 4750 0.0 -
0.1670 4800 0.0 -
0.1688 4850 0.0 -
0.1705 4900 0.0 -
0.1722 4950 0.0 -
0.1740 5000 0.0 -
0.1757 5050 0.0 -
0.1775 5100 0.0 -
0.1792 5150 0.0 -
0.1809 5200 0.0 -
0.1827 5250 0.0 -
0.1844 5300 0.0 -
0.1862 5350 0.0 -
0.1879 5400 0.0 -
0.1896 5450 0.0 -
0.1914 5500 0.0 -
0.1931 5550 0.0 -
0.1949 5600 0.0 -
0.1966 5650 0.0 -
0.1983 5700 0.0 -
0.2001 5750 0.0 -
0.2018 5800 0.0 -
0.2036 5850 0.0 -
0.2053 5900 0.0 -
0.2070 5950 0.0 -
0.2088 6000 0.0 0.0
0.2105 6050 0.0 -
0.2123 6100 0.0 -
0.2140 6150 0.0 -
0.2157 6200 0.0 -
0.2175 6250 0.0 -
0.2192 6300 0.0 -
0.2210 6350 0.0 -
0.2227 6400 0.0 -
0.2244 6450 0.0 -
0.2262 6500 0.0 -
0.2279 6550 0.0 -
0.2297 6600 0.0 -
0.2314 6650 0.0 -
0.2331 6700 0.0 -
0.2349 6750 0.0 -
0.2366 6800 0.0 -
0.2384 6850 0.0 -
0.2401 6900 0.0 -
0.2418 6950 0.0 -
0.2436 7000 0.0 -
0.2453 7050 0.0 -
0.2471 7100 0.0 -
0.2488 7150 0.0 -
0.2505 7200 0.0 -
0.2523 7250 0.0 -
0.2540 7300 0.0 -
0.2558 7350 0.0 -
0.2575 7400 0.0 -
0.2592 7450 0.0 -
0.2610 7500 0.0 -
0.2627 7550 0.0 -
0.2645 7600 0.0 -
0.2662 7650 0.0 -
0.2679 7700 0.0 -
0.2697 7750 0.0 -
0.2714 7800 0.0 -
0.2732 7850 0.0 -
0.2749 7900 0.0 -
0.2766 7950 0.0 -
0.2784 8000 0.0 -
0.2801 8050 0.0 -
0.2819 8100 0.0 -
0.2836 8150 0.0 -
0.2853 8200 0.0 -
0.2871 8250 0.0 -
0.2888 8300 0.0 -
0.2906 8350 0.0 -
0.2923 8400 0.0 -
0.2940 8450 0.0 -
0.2958 8500 0.0 -
0.2975 8550 0.0 -
0.2993 8600 0.0 -
0.3010 8650 0.0 -
0.3027 8700 0.0 -
0.3045 8750 0.0 -
0.3062 8800 0.0 -
0.3080 8850 0.0 -
0.3097 8900 0.0 -
0.3114 8950 0.0 -
0.3132 9000 0.0 0.0
0.3149 9050 0.0 -
0.3167 9100 0.0 -
0.3184 9150 0.0 -
0.3201 9200 0.0 -
0.3219 9250 0.0 -
0.3236 9300 0.0 -
0.3254 9350 0.0 -
0.3271 9400 0.0 -
0.3288 9450 0.0 -
0.3306 9500 0.0 -
0.3323 9550 0.0 -
0.3341 9600 0.0 -
0.3358 9650 0.0 -
0.3375 9700 0.0 -
0.3393 9750 0.0 -
0.3410 9800 0.0 -
0.3428 9850 0.0 -
0.3445 9900 0.0 -
0.3462 9950 0.0 -
0.3480 10000 0.0 -
0.3497 10050 0.0 -
0.3515 10100 0.0 -
0.3532 10150 0.0 -
0.3549 10200 0.0 -
0.3567 10250 0.0 -
0.3584 10300 0.0 -
0.3602 10350 0.0 -
0.3619 10400 0.0 -
0.3636 10450 0.0 -
0.3654 10500 0.0 -
0.3671 10550 0.0 -
0.3688 10600 0.0 -
0.3706 10650 0.0 -
0.3723 10700 0.0 -
0.3741 10750 0.0 -
0.3758 10800 0.0 -
0.3775 10850 0.0 -
0.3793 10900 0.0 -
0.3810 10950 0.0 -
0.3828 11000 0.0 -
0.3845 11050 0.0 -
0.3862 11100 0.0 -
0.3880 11150 0.0 -
0.3897 11200 0.0 -
0.3915 11250 0.0 -
0.3932 11300 0.0 -
0.3949 11350 0.0 -
0.3967 11400 0.0 -
0.3984 11450 0.0 -
0.4002 11500 0.0 -
0.4019 11550 0.0 -
0.4036 11600 0.0 -
0.4054 11650 0.0 -
0.4071 11700 0.0 -
0.4089 11750 0.0 -
0.4106 11800 0.0 -
0.4123 11850 0.0 -
0.4141 11900 0.0 -
0.4158 11950 0.0 -
0.4176 12000 0.0 0.0
0.4193 12050 0.0 -
0.4210 12100 0.0 -
0.4228 12150 0.0 -
0.4245 12200 0.0 -
0.4263 12250 0.0 -
0.4280 12300 0.0 -
0.4297 12350 0.0 -
0.4315 12400 0.0 -
0.4332 12450 0.0 -
0.4350 12500 0.0 -
0.4367 12550 0.0 -
0.4384 12600 0.0 -
0.4402 12650 0.0 -
0.4419 12700 0.0 -
0.4437 12750 0.0 -
0.4454 12800 0.0 -
0.4471 12850 0.0 -
0.4489 12900 0.0 -
0.4506 12950 0.0 -
0.4524 13000 0.0 -
0.4541 13050 0.0 -
0.4558 13100 0.0 -
0.4576 13150 0.0 -
0.4593 13200 0.0 -
0.4611 13250 0.0 -
0.4628 13300 0.0 -
0.4645 13350 0.0 -
0.4663 13400 0.0 -
0.4680 13450 0.0 -
0.4698 13500 0.0 -
0.4715 13550 0.0 -
0.4732 13600 0.0 -
0.4750 13650 0.0 -
0.4767 13700 0.0 -
0.4785 13750 0.0 -
0.4802 13800 0.0 -
0.4819 13850 0.0 -
0.4837 13900 0.0 -
0.4854 13950 0.0 -
0.4872 14000 0.0 -
0.4889 14050 0.0 -
0.4906 14100 0.0 -
0.4924 14150 0.0 -
0.4941 14200 0.0 -
0.4959 14250 0.0 -
0.4976 14300 0.0 -
0.4993 14350 0.0 -
0.5011 14400 0.0 -
0.5028 14450 0.0 -
0.5046 14500 0.0 -
0.5063 14550 0.0 -
0.5080 14600 0.0 -
0.5098 14650 0.0 -
0.5115 14700 0.0 -
0.5133 14750 0.0 -
0.5150 14800 0.0 -
0.5167 14850 0.0 -
0.5185 14900 0.0 -
0.5202 14950 0.0 -
0.5220 15000 0.0 0.0
0.5237 15050 0.0 -
0.5254 15100 0.0 -
0.5272 15150 0.0 -
0.5289 15200 0.0 -
0.5307 15250 0.0 -
0.5324 15300 0.0 -
0.5341 15350 0.0 -
0.5359 15400 0.0 -
0.5376 15450 0.0 -
0.5394 15500 0.0 -
0.5411 15550 0.0 -
0.5428 15600 0.0 -
0.5446 15650 0.0 -
0.5463 15700 0.0 -
0.5481 15750 0.0 -
0.5498 15800 0.0 -
0.5515 15850 0.0 -
0.5533 15900 0.0 -
0.5550 15950 0.0 -
0.5568 16000 0.0 -
0.5585 16050 0.0 -
0.5602 16100 0.0 -
0.5620 16150 0.0 -
0.5637 16200 0.0 -
0.5655 16250 0.0 -
0.5672 16300 0.0 -
0.5689 16350 0.0 -
0.5707 16400 0.0 -
0.5724 16450 0.0 -
0.5742 16500 0.0 -
0.5759 16550 0.0 -
0.5776 16600 0.0 -
0.5794 16650 0.0 -
0.5811 16700 0.0 -
0.5829 16750 0.0 -
0.5846 16800 0.0 -
0.5863 16850 0.0 -
0.5881 16900 0.0 -
0.5898 16950 0.0 -
0.5916 17000 0.0 -
0.5933 17050 0.0 -
0.5950 17100 0.0 -
0.5968 17150 0.0 -
0.5985 17200 0.0 -
0.6003 17250 0.0 -
0.6020 17300 0.0 -
0.6037 17350 0.0 -
0.6055 17400 0.0 -
0.6072 17450 0.0 -
0.6089 17500 0.0 -
0.6107 17550 0.0 -
0.6124 17600 0.0 -
0.6142 17650 0.0 -
0.6159 17700 0.0 -
0.6176 17750 0.0 -
0.6194 17800 0.0 -
0.6211 17850 0.0 -
0.6229 17900 0.0 -
0.6246 17950 0.0 -
0.6263 18000 0.0 0.0
0.6281 18050 0.0 -
0.6298 18100 0.0 -
0.6316 18150 0.0 -
0.6333 18200 0.0 -
0.6350 18250 0.0 -
0.6368 18300 0.0 -
0.6385 18350 0.0 -
0.6403 18400 0.0 -
0.6420 18450 0.0 -
0.6437 18500 0.0 -
0.6455 18550 0.0 -
0.6472 18600 0.0 -
0.6490 18650 0.0 -
0.6507 18700 0.0 -
0.6524 18750 0.0 -
0.6542 18800 0.0 -
0.6559 18850 0.0 -
0.6577 18900 0.0 -
0.6594 18950 0.0 -
0.6611 19000 0.0 -
0.6629 19050 0.0 -
0.6646 19100 0.0 -
0.6664 19150 0.0 -
0.6681 19200 0.0 -
0.6698 19250 0.0 -
0.6716 19300 0.0 -
0.6733 19350 0.0 -
0.6751 19400 0.0 -
0.6768 19450 0.0 -
0.6785 19500 0.0 -
0.6803 19550 0.0 -
0.6820 19600 0.0 -
0.6838 19650 0.0 -
0.6855 19700 0.0 -
0.6872 19750 0.0 -
0.6890 19800 0.0 -
0.6907 19850 0.0 -
0.6925 19900 0.0 -
0.6942 19950 0.0 -
0.6959 20000 0.0 -
0.6977 20050 0.0 -
0.6994 20100 0.0 -
0.7012 20150 0.0 -
0.7029 20200 0.0 -
0.7046 20250 0.0 -
0.7064 20300 0.0 -
0.7081 20350 0.0 -
0.7099 20400 0.0 -
0.7116 20450 0.0 -
0.7133 20500 0.0 -
0.7151 20550 0.0 -
0.7168 20600 0.0 -
0.7186 20650 0.0 -
0.7203 20700 0.0 -
0.7220 20750 0.0 -
0.7238 20800 0.0 -
0.7255 20850 0.0 -
0.7273 20900 0.0 -
0.7290 20950 0.0 -
0.7307 21000 0.0 0.0
0.7325 21050 0.0 -
0.7342 21100 0.0 -
0.7360 21150 0.0 -
0.7377 21200 0.0 -
0.7394 21250 0.0 -
0.7412 21300 0.0 -
0.7429 21350 0.0 -
0.7447 21400 0.0 -
0.7464 21450 0.0 -
0.7481 21500 0.0 -
0.7499 21550 0.0 -
0.7516 21600 0.0 -
0.7534 21650 0.0 -
0.7551 21700 0.0 -
0.7568 21750 0.0 -
0.7586 21800 0.0 -
0.7603 21850 0.0 -
0.7621 21900 0.0 -
0.7638 21950 0.0 -
0.7655 22000 0.0 -
0.7673 22050 0.0 -
0.7690 22100 0.0 -
0.7708 22150 0.0 -
0.7725 22200 0.0 -
0.7742 22250 0.0 -
0.7760 22300 0.0 -
0.7777 22350 0.0 -
0.7795 22400 0.0 -
0.7812 22450 0.0 -
0.7829 22500 0.0 -
0.7847 22550 0.0 -
0.7864 22600 0.0 -
0.7882 22650 0.0 -
0.7899 22700 0.0 -
0.7916 22750 0.0 -
0.7934 22800 0.0 -
0.7951 22850 0.0 -
0.7969 22900 0.0 -
0.7986 22950 0.0 -
0.8003 23000 0.0 -
0.8021 23050 0.0 -
0.8038 23100 0.0 -
0.8056 23150 0.0 -
0.8073 23200 0.0 -
0.8090 23250 0.0 -
0.8108 23300 0.0 -
0.8125 23350 0.0 -
0.8143 23400 0.0 -
0.8160 23450 0.0 -
0.8177 23500 0.0 -
0.8195 23550 0.0 -
0.8212 23600 0.0 -
0.8230 23650 0.0 -
0.8247 23700 0.0 -
0.8264 23750 0.0 -
0.8282 23800 0.0 -
0.8299 23850 0.0 -
0.8317 23900 0.0 -
0.8334 23950 0.0 -
0.8351 24000 0.0 0.0
0.8369 24050 0.0 -
0.8386 24100 0.0 -
0.8404 24150 0.0 -
0.8421 24200 0.0 -
0.8438 24250 0.0 -
0.8456 24300 0.0 -
0.8473 24350 0.0 -
0.8491 24400 0.0 -
0.8508 24450 0.0 -
0.8525 24500 0.0 -
0.8543 24550 0.0 -
0.8560 24600 0.0 -
0.8577 24650 0.0 -
0.8595 24700 0.0 -
0.8612 24750 0.0 -
0.8630 24800 0.0 -
0.8647 24850 0.0 -
0.8664 24900 0.0 -
0.8682 24950 0.0 -
0.8699 25000 0.0 -
0.8717 25050 0.0 -
0.8734 25100 0.0 -
0.8751 25150 0.0 -
0.8769 25200 0.0 -
0.8786 25250 0.0 -
0.8804 25300 0.0 -
0.8821 25350 0.0 -
0.8838 25400 0.0 -
0.8856 25450 0.0 -
0.8873 25500 0.0 -
0.8891 25550 0.0 -
0.8908 25600 0.0 -
0.8925 25650 0.0 -
0.8943 25700 0.0 -
0.8960 25750 0.0 -
0.8978 25800 0.0 -
0.8995 25850 0.0 -
0.9012 25900 0.0 -
0.9030 25950 0.0 -
0.9047 26000 0.0 -
0.9065 26050 0.0 -
0.9082 26100 0.0 -
0.9099 26150 0.0 -
0.9117 26200 0.0 -
0.9134 26250 0.0 -
0.9152 26300 0.0 -
0.9169 26350 0.0 -
0.9186 26400 0.0 -
0.9204 26450 0.0 -
0.9221 26500 0.0 -
0.9239 26550 0.0 -
0.9256 26600 0.0 -
0.9273 26650 0.0 -
0.9291 26700 0.0 -
0.9308 26750 0.0 -
0.9326 26800 0.0 -
0.9343 26850 0.0 -
0.9360 26900 0.0 -
0.9378 26950 0.0 -
0.9395 27000 0.0 0.0
0.9413 27050 0.0 -
0.9430 27100 0.0 -
0.9447 27150 0.0 -
0.9465 27200 0.0 -
0.9482 27250 0.0 -
0.9500 27300 0.0 -
0.9517 27350 0.0 -
0.9534 27400 0.0 -
0.9552 27450 0.0 -
0.9569 27500 0.0 -
0.9587 27550 0.0 -
0.9604 27600 0.0 -
0.9621 27650 0.0 -
0.9639 27700 0.0 -
0.9656 27750 0.0 -
0.9674 27800 0.0 -
0.9691 27850 0.0 -
0.9708 27900 0.0 -
0.9726 27950 0.0 -
0.9743 28000 0.0 -
0.9761 28050 0.0 -
0.9778 28100 0.0 -
0.9795 28150 0.0 -
0.9813 28200 0.0 -
0.9830 28250 0.0 -
0.9848 28300 0.0 -
0.9865 28350 0.0 -
0.9882 28400 0.0 -
0.9900 28450 0.0 -
0.9917 28500 0.0 -
0.9935 28550 0.0 -
0.9952 28600 0.0 -
0.9969 28650 0.0 -
0.9987 28700 0.0 -
  • The bold row denotes the saved checkpoint.

Framework Versions

  • Python: 3.10.14
  • SetFit: 1.0.3
  • Sentence Transformers: 3.0.1
  • Transformers: 4.41.2
  • PyTorch: 2.3.1+cu121
  • 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}
}