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:

Fine-tuning a Sentence Transformer with contrastive learning.
Training a classification head with features from the fine-tuned Sentence Transformer.

Model Details

Model Description

Model Type: SetFit
Sentence Transformer body: Vishal24/bert-1ds-domain
Classification head: a SetFitHead instance
Maximum Sequence Length: 512 tokens
Number of Classes: 2 classes
Training Dataset: Vishal24/BCG_classifier

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
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.

python

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

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}
}