metadata
library_name: setfit
tags:
- setfit
- absa
- sentence-transformers
- text-classification
- generated_from_setfit_trainer
metrics:
- accuracy
widget:
- text: >-
genshin impact, grafik nya udah bagus:pengalaman yang aku rasakan saat
main genshin impact, grafik nya udah bagus, sesuai dengan ukurannya yang
besar, tapi ada hal yang nyeselin saat aku main genshin impact, ada bug
layar hp aku suka gerak gerak sendiri saat aku baru baru download genshin
impact itu layarnya gak gerak sendiri, pengalaman saya main genshin impact
sekarang ini gak nyaman karena ada bug layar gerak sendiri. mohon
bantuannya cognnosphere pte. ltd.
- text: >-
grafiknya juga keren karakternya cakep:gamenya sangat bagus sama grafiknya
juga keren karakternya cakep
- text: >-
aja tidak ada fitur skip story apalagi:genshin impact game kikir saya
sudah main 3 tahun masih gitu2 aja hadiah ulang tahun sama imlek hadiahnya
biasa2 aja tidak ada fitur skip story apalagi story nya bikin ngantuk jadi
makin boring main ini, mc bisu kebanyakan paimon yang banyak bicaranya
berisik lagi tuh
- text: >-
,mulai dari konten yang disajikan sampai:overall game nya bagus,mulai dari
konten yang disajikan sampai design karakter nya,namun yang disayangkan
adalah performa gameplay nya untuk hp kelas low end karena saya mengalami
force close setiap kali mulai selesai quest,jadi mohon agar developer nya
memperhatikan masalah ini
- text: >-
story mantul, map luas bgt,:game paling debes yg pernah gwe temuin ampe
saat ini. gameplay seru, story mantul, map luas bgt, grapik jangan di
tanya salutlah ama hoyoverse. coba klo hoyoverse lebih ngurusin ni game
bakalan jadi lebih seru lagi d
pipeline_tag: text-classification
inference: false
SetFit Polarity 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 classifying aspect polarities.
The model has been trained using an efficient few-shot learning technique that involves:
- Fine-tuning a Sentence Transformer with contrastive learning.
- 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:
- Use a spaCy model to select possible aspect span candidates.
- Use a SetFit model to filter these possible aspect span candidates.
- Use this SetFit model to classify the filtered aspect span candidates.
Model Details
Model Description
- Model Type: SetFit
- Classification head: a LogisticRegression instance
- spaCy Model: id_core_news_trf
- SetFitABSA Aspect Model: Funnyworld1412/ABSA_review_game_genshin-aspect
- SetFitABSA Polarity Model: Funnyworld1412/ABSA_review_game_genshin-polarity
- Maximum Sequence Length: 8192 tokens
- Number of Classes: 2 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 |
---|---|
Negative |
|
Positive |
|
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 | 6 | 46.7275 | 98 |
Label | Training Sample Count |
---|---|
konflik | 0 |
negatif | 0 |
netral | 0 |
positif | 0 |
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 |
---|---|---|---|
0.0004 | 1 | 0.2547 | - |
0.0210 | 50 | 0.2787 | - |
0.0419 | 100 | 0.002 | - |
0.0629 | 150 | 0.2062 | - |
0.0839 | 200 | 0.2148 | - |
0.1048 | 250 | 0.209 | - |
0.1258 | 300 | 0.1926 | - |
0.1468 | 350 | 0.2244 | - |
0.1677 | 400 | 0.0034 | - |
0.1887 | 450 | 0.2523 | - |
0.2096 | 500 | 0.0027 | - |
0.2306 | 550 | 0.001 | - |
0.2516 | 600 | 0.0016 | - |
0.2725 | 650 | 0.0011 | - |
0.2935 | 700 | 0.2077 | - |
0.3145 | 750 | 0.0025 | - |
0.3354 | 800 | 0.0014 | - |
0.3564 | 850 | 0.0011 | - |
0.3774 | 900 | 0.0028 | - |
0.3983 | 950 | 0.0004 | - |
0.4193 | 1000 | 0.0005 | - |
0.4403 | 1050 | 0.0011 | - |
0.4612 | 1100 | 0.0011 | - |
0.4822 | 1150 | 0.0007 | - |
0.5031 | 1200 | 0.0009 | - |
0.5241 | 1250 | 0.0161 | - |
0.5451 | 1300 | 0.0013 | - |
0.5660 | 1350 | 0.0003 | - |
0.5870 | 1400 | 0.0003 | - |
0.6080 | 1450 | 0.0005 | - |
0.6289 | 1500 | 0.0004 | - |
0.6499 | 1550 | 0.0003 | - |
0.6709 | 1600 | 0.0004 | - |
0.6918 | 1650 | 0.0005 | - |
0.7128 | 1700 | 0.0005 | - |
0.7338 | 1750 | 0.0003 | - |
0.7547 | 1800 | 0.0013 | - |
0.7757 | 1850 | 0.0004 | - |
0.7966 | 1900 | 0.0006 | - |
0.8176 | 1950 | 0.0003 | - |
0.8386 | 2000 | 0.0003 | - |
0.8595 | 2050 | 0.0005 | - |
0.8805 | 2100 | 0.0003 | - |
0.9015 | 2150 | 0.0005 | - |
0.9224 | 2200 | 0.0002 | - |
0.9434 | 2250 | 0.0003 | - |
0.9644 | 2300 | 0.0003 | - |
0.9853 | 2350 | 0.0002 | - |
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}
}