You can deploy this model and start earning money today!

Funnyworld1412/
ABSA_review_game_genshin-aspect

Name: ABSA_review_game_genshin-aspect
Availability: InStock

sentence-transformers

text-classification

generated_from_setfit_trainer

2209.11055

Views

No views yet

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:

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 this SetFit model to filter these possible aspect span candidates.
Use a 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: 512 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

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.

python

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

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}
}