Hugging Face's logo

Funnyworld1412
/
ABSA_review_game_genshin-aspect

Text Classification
setfit
Safetensors
sentence-transformers
bert
absa
generated_from_setfit_trainer
Model card Files Community
Edit model card

SetFit Aspect Model

This is a SetFit model that can be used for Aspect Based Sentiment Analysis (ABSA). A LogisticRegression instance is used for classification. In particular, this model is in charge of filtering aspect span candidates.

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.

This model was trained within the context of a larger system for ABSA, which looks like so:

  1. Use a spaCy model to select possible aspect span candidates.
  2. Use this SetFit model to filter these possible aspect span candidates.
  3. Use a SetFit model to classify the filtered aspect span candidates.

Model Details

Model Description

Model Sources

Model Labels

Label Examples
aspect
  • 'story:saranku developer harus menciptakan sebuah story yang sangat menarik, agar tidak kehilangan para player karena masalahnya banyak player yg tidak bertahan lama karena repetitif dan monoton tiap update, size makin gede doang yg isinya cuma chest baru itupun sampah, puzzle yg makin lama makin rumit tapi chest nya sampah, story kebanyakan npc teyvat story utama punya mc dilupain gak difokusin , map kalo udah kosong ya nyampah bikin size gede doang. main 3 tahun rasanya monoton, perkembangan buruk'
  • 'reward:tolong ditambah lagi reward untuk gachanya, untuk player lama kesulitan mendapatkan primo karena sudah tidak ada lagi quest dan eksplorasi juga sudah 100 . dasar developer kapitalis, game ini makin lama makin monoton dan tidak ramah untuk player lama yang kekurangan bahan untuk gacha karakter'
  • 'event:cuman saran jangan terlalu pelit.. biar para player gak kabur sama game sebelah hadiah event quest di perbaiki.... udah nunggu event lama lama hadiah cuman gitu gitu aja... sampek event selesai primogemnya buat 10 pull gacha gak cukup.... tingakat kesulitan beda hadiah sama saja... lama lama yang main pada kabur kalok terlalu pelit.. dan 1 lagi jariang mohon di perbaiki untuk server indonya trimaksih'
no aspect
  • 'saranku developer:saranku developer harus menciptakan sebuah story yang sangat menarik, agar tidak kehilangan para player karena masalahnya banyak player yg tidak bertahan lama karena repetitif dan monoton tiap update, size makin gede doang yg isinya cuma chest baru itupun sampah, puzzle yg makin lama makin rumit tapi chest nya sampah, story kebanyakan npc teyvat story utama punya mc dilupain gak difokusin , map kalo udah kosong ya nyampah bikin size gede doang. main 3 tahun rasanya monoton, perkembangan buruk'
  • 'story:saranku developer harus menciptakan sebuah story yang sangat menarik, agar tidak kehilangan para player karena masalahnya banyak player yg tidak bertahan lama karena repetitif dan monoton tiap update, size makin gede doang yg isinya cuma chest baru itupun sampah, puzzle yg makin lama makin rumit tapi chest nya sampah, story kebanyakan npc teyvat story utama punya mc dilupain gak difokusin , map kalo udah kosong ya nyampah bikin size gede doang. main 3 tahun rasanya monoton, perkembangan buruk'
  • 'player:saranku developer harus menciptakan sebuah story yang sangat menarik, agar tidak kehilangan para player karena masalahnya banyak player yg tidak bertahan lama karena repetitif dan monoton tiap update, size makin gede doang yg isinya cuma chest baru itupun sampah, puzzle yg makin lama makin rumit tapi chest nya sampah, story kebanyakan npc teyvat story utama punya mc dilupain gak difokusin , map kalo udah kosong ya nyampah bikin size gede doang. main 3 tahun rasanya monoton, perkembangan buruk'

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 AbsaModel

# Download from the 🤗 Hub
model = AbsaModel.from_pretrained(
    "Funnyworld1412/ABSA_review_game_genshin-aspect",
    "Funnyworld1412/ABSA_review_game_genshin-polarity",
)
# Run inference
preds = model("The food was great, but the venue is just way too busy.")

Training Details

Training Set Metrics

Training set Min Median Max
Word count 4 49.9079 94
Label Training Sample Count
no aspect 2281
aspect 477

Training Hyperparameters

  • batch_size: (4, 4)
  • num_epochs: (1, 1)
  • max_steps: -1
  • sampling_strategy: oversampling
  • num_iterations: 10
  • 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 0.25 -
0.0036 50 0.331 -
0.0073 100 0.5002 -
0.0109 150 0.2904 -
0.0145 200 0.3791 -
0.0181 250 0.2253 -
0.0218 300 0.1909 -
0.0254 350 0.2504 -
0.0290 400 0.1241 -
0.0326 450 0.1021 -
0.0363 500 0.0985 -
0.0399 550 0.3831 -
0.0435 600 0.1841 -
0.0471 650 0.2487 -
0.0508 700 0.1573 -
0.0544 750 0.0499 -
0.0580 800 0.2214 -
0.0616 850 0.1427 -
0.0653 900 0.3544 -
0.0689 950 0.042 -
0.0725 1000 0.2918 -
0.0761 1050 0.0134 -
0.0798 1100 0.1933 -
0.0834 1150 0.0115 -
0.0870 1200 0.2393 -
0.0906 1250 0.2625 -
0.0943 1300 0.1496 -
0.0979 1350 0.1417 -
0.1015 1400 0.2111 -
0.1051 1450 0.2158 -
0.1088 1500 0.1378 -
0.1124 1550 0.0988 -
0.1160 1600 0.1183 -
0.1197 1650 0.324 -
0.1233 1700 0.3722 -
0.1269 1750 0.1696 -
0.1305 1800 0.2893 -
0.1342 1850 0.198 -
0.1378 1900 0.2854 -
0.1414 1950 0.3339 -
0.1450 2000 0.0783 -
0.1487 2050 0.014 -
0.1523 2100 0.0205 -
0.1559 2150 0.0151 -
0.1595 2200 0.3783 -
0.1632 2250 0.381 -
0.1668 2300 0.144 -
0.1704 2350 0.0023 -
0.1740 2400 0.1903 -
0.1777 2450 0.0033 -
0.1813 2500 0.0039 -
0.1849 2550 0.0019 -
0.1885 2600 0.0565 -
0.1922 2650 0.1551 -
0.1958 2700 0.0729 -
0.1994 2750 0.0272 -
0.2030 2800 0.495 -
0.2067 2850 0.0396 -
0.2103 2900 0.2288 -
0.2139 2950 0.0077 -
0.2175 3000 0.0642 -
0.2212 3050 0.0037 -
0.2248 3100 0.2447 -
0.2284 3150 0.0097 -
0.2321 3200 0.0011 -
0.2357 3250 0.1254 -
0.2393 3300 0.0046 -
0.2429 3350 0.0127 -
0.2466 3400 0.0093 -
0.2502 3450 0.0005 -
0.2538 3500 0.0022 -
0.2574 3550 0.0005 -
0.2611 3600 0.0002 -
0.2647 3650 0.0231 -
0.2683 3700 0.0016 -
0.2719 3750 0.1945 -
0.2756 3800 0.002 -
0.2792 3850 0.0235 -
0.2828 3900 0.006 -
0.2864 3950 0.0003 -
0.2901 4000 0.007 -
0.2937 4050 0.0227 -
0.2973 4100 0.1794 -
0.3009 4150 0.2629 -
0.3046 4200 0.3005 -
0.3082 4250 0.1974 -
0.3118 4300 0.001 -
0.3154 4350 0.0123 -
0.3191 4400 0.0027 -
0.3227 4450 0.0002 -
0.3263 4500 0.0005 -
0.3299 4550 0.0002 -
0.3336 4600 0.0007 -
0.3372 4650 0.0332 -
0.3408 4700 0.052 -
0.3445 4750 0.0103 -
0.3481 4800 0.0067 -
0.3517 4850 0.0003 -
0.3553 4900 0.0008 -
0.3590 4950 0.0088 -
0.3626 5000 0.0002 -
0.3662 5050 0.0111 -
0.3698 5100 0.0836 -
0.3735 5150 0.0001 -
0.3771 5200 0.2398 -
0.3807 5250 0.0002 -
0.3843 5300 0.1435 -
0.3880 5350 0.0001 -
0.3916 5400 0.0296 -
0.3952 5450 0.0003 -
0.3988 5500 0.1126 -
0.4025 5550 0.0009 -
0.4061 5600 0.0055 -
0.4097 5650 0.0031 -
0.4133 5700 0.1929 -
0.4170 5750 0.0002 -
0.4206 5800 0.2565 -
0.4242 5850 0.0002 -
0.4278 5900 0.0033 -
0.4315 5950 0.0011 -
0.4351 6000 0.0001 -
0.4387 6050 0.0004 -
0.4423 6100 0.0003 -
0.4460 6150 0.1076 -
0.4496 6200 0.0011 -
0.4532 6250 0.0034 -
0.4569 6300 0.0176 -
0.4605 6350 0.2883 -
0.4641 6400 0.0 -
0.4677 6450 0.0172 -
0.4714 6500 0.0014 -
0.4750 6550 0.0571 -
0.4786 6600 0.0287 -
0.4822 6650 0.1461 -
0.4859 6700 0.2333 -
0.4895 6750 0.1468 -
0.4931 6800 0.0005 -
0.4967 6850 0.0039 -
0.5004 6900 0.0004 -
0.5040 6950 0.0008 -
0.5076 7000 0.0004 -
0.5112 7050 0.0005 -
0.5149 7100 0.001 -
0.5185 7150 0.0041 -
0.5221 7200 0.0157 -
0.5257 7250 0.0228 -
0.5294 7300 0.0002 -
0.5330 7350 0.0004 -
0.5366 7400 0.0081 -
0.5402 7450 0.0004 -
0.5439 7500 0.1227 -
0.5475 7550 0.0001 -
0.5511 7600 0.0006 -
0.5547 7650 0.0003 -
0.5584 7700 0.0475 -
0.5620 7750 0.1848 -
0.5656 7800 0.0007 -
0.5693 7850 0.001 -
0.5729 7900 0.0002 -
0.5765 7950 0.0018 -
0.5801 8000 0.0009 -
0.5838 8050 0.0019 -
0.5874 8100 0.0001 -
0.5910 8150 0.0012 -
0.5946 8200 0.0536 -
0.5983 8250 0.0943 -
0.6019 8300 0.006 -
0.6055 8350 0.0019 -
0.6091 8400 0.0 -
0.6128 8450 0.0004 -
0.6164 8500 0.0 -
0.6200 8550 0.2588 -
0.6236 8600 0.0001 -
0.6273 8650 0.0084 -
0.6309 8700 0.0001 -
0.6345 8750 0.4123 -
0.6381 8800 0.073 -
0.6418 8850 0.0 -
0.6454 8900 0.1361 -
0.6490 8950 0.0249 -
0.6526 9000 0.0003 -
0.6563 9050 0.0018 -
0.6599 9100 0.0115 -
0.6635 9150 0.1789 -
0.6672 9200 0.0001 -
0.6708 9250 0.0006 -
0.6744 9300 0.002 -
0.6780 9350 0.0 -
0.6817 9400 0.0042 -
0.6853 9450 0.0003 -
0.6889 9500 0.0105 -
0.6925 9550 0.0 -
0.6962 9600 0.0285 -
0.6998 9650 0.0002 -
0.7034 9700 0.0 -
0.7070 9750 0.001 -
0.7107 9800 0.0641 -
0.7143 9850 0.0096 -
0.7179 9900 0.0001 -
0.7215 9950 0.0003 -
0.7252 10000 0.3666 -
0.7288 10050 0.0001 -
0.7324 10100 0.0001 -
0.7360 10150 0.0001 -
0.7397 10200 0.2526 -
0.7433 10250 0.0286 -
0.7469 10300 0.0001 -
0.7505 10350 0.004 -
0.7542 10400 0.0 -
0.7578 10450 0.0237 -
0.7614 10500 0.0012 -
0.7650 10550 0.0001 -
0.7687 10600 0.0223 -
0.7723 10650 0.0349 -
0.7759 10700 0.033 -
0.7796 10750 0.0005 -
0.7832 10800 0.0001 -
0.7868 10850 0.0001 -
0.7904 10900 0.0002 -
0.7941 10950 0.0005 -
0.7977 11000 0.0003 -
0.8013 11050 0.0 -
0.8049 11100 0.0348 -
0.8086 11150 0.0 -
0.8122 11200 0.0001 -
0.8158 11250 0.0 -
0.8194 11300 0.0 -
0.8231 11350 0.0 -
0.8267 11400 0.0002 -
0.8303 11450 0.0002 -
0.8339 11500 0.0112 -
0.8376 11550 0.0099 -
0.8412 11600 0.0 -
0.8448 11650 0.0 -
0.8484 11700 0.045 -
0.8521 11750 0.138 -
0.8557 11800 0.0283 -
0.8593 11850 0.0001 -
0.8629 11900 0.0 -
0.8666 11950 0.0751 -
0.8702 12000 0.0002 -
0.8738 12050 0.0 -
0.8774 12100 0.0001 -
0.8811 12150 0.0948 -
0.8847 12200 0.0896 -
0.8883 12250 0.1255 -
0.8920 12300 0.0001 -
0.8956 12350 0.0 -
0.8992 12400 0.1456 -
0.9028 12450 0.0079 -
0.9065 12500 0.0 -
0.9101 12550 0.0 -
0.9137 12600 0.0002 -
0.9173 12650 0.0047 -
0.9210 12700 0.1701 -
0.9246 12750 0.0423 -
0.9282 12800 0.0001 -
0.9318 12850 0.0969 -
0.9355 12900 0.0001 -
0.9391 12950 0.0 -
0.9427 13000 0.0 -
0.9463 13050 0.0301 -
0.9500 13100 0.0066 -
0.9536 13150 0.0 -
0.9572 13200 0.0 -
0.9608 13250 0.0 -
0.9645 13300 0.0 -
0.9681 13350 0.0008 -
0.9717 13400 0.0255 -
0.9753 13450 0.0 -
0.9790 13500 0.0908 -
0.9826 13550 0.0826 -
0.9862 13600 0.0 -
0.9898 13650 0.0247 -
0.9935 13700 0.0 -
0.9971 13750 0.0546 -

Framework Versions

  • Python: 3.10.13
  • SetFit: 1.0.3
  • Sentence Transformers: 3.0.1
  • spaCy: 3.7.5
  • Transformers: 4.36.2
  • PyTorch: 2.1.2
  • Datasets: 2.19.2
  • 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}
}
Downloads last month
0
Safetensors
Model size
124M params
Tensor type
F32
·
Inference Examples
Text Classification
Inference API (serverless) has been turned off for this model.