SetFit with BAAI/bge-small-en-v1.5

This is a SetFit model that can be used for Text Classification. This SetFit model uses BAAI/bge-small-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:

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: BAAI/bge-small-en-v1.5
Classification head: a LogisticRegression instance
Maximum Sequence Length: 512 tokens
Number of Classes: 3 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
2	'ORIGINAL FOR RECIPIENT\n\nM/S JAIDURGA CONSTRUCTION\n\nAT-BUDAKATA , PO- GADAMUNDA MOBILE NO. : 9438452293/9861569621/9178245171\nHIRAKUD, DIST: SAMBALPUR E MAIL - dkppni@rediffmail.com\n\nSS\n(ISSUEDUNDER RULE 46 OF GST/OGST RULE,2017)\nINVOICE NO. A8tHRA/BT-1/20\nINVO DA 6.12.\n\n \n \n\n \n \n \n \n \n\n \n \n\n \n \n\nHIRAKUD POWER\n
1	'A GSTIN: 21AAACH1201R12Z\n_ HINDALCO INDUSTRIES LIMITED State Code: 21 - Odisha\n\nHIRAKUD POWER, HIRAKUD-768 016,DIST.:SAMBALPUR (ODISHA) GST RaligeiaN ibe:\n\nSambalpur\nPHONE: 0663-2481365, FAX: 0663-2481942 GST Commissionerate -Cuttack\n\n \n\nPURCHASE ORDER\n\nVendor Code: 6301 P.OINo: P/PO/SRV/2122/0480 _ Date:08-SEP-2021\nM/s GOVINDAM Revise No: Date:\nOrder Type: PURCHASE ORDER\n\nNEAR ASHAPALI SCHOOL Effective From 01/09/2021 To 31/08/2022\nGM, COLLEGE ROAD Pilon Bante: 4\nSAM IR, ODISHA, IN 768001 Transportation arrangement\n\niaiiatipraiiinaaiisaaas Ship to Location HIRAKUD - POWER\nEmail: sauravkedia&4@gmail.com Cartier\nFax’ () Currency INR\nContact: Saurav Kedia (+91) 9338017181 Hindalco Contact Person: AMARESH MISHRA,\n\nJ GSTIN: 21AHLPK8955E3Z6 —State:21 - Odisha Email of Contact Person: amaresh.m@adityabiria.com\n\n \n\n“You will be raising vaikd Tax invoice as per GST Laws for the supply covered by the instant\nPO and comply with all the GST rules and regufations as notified and/or to be notified in fulure\nincluding fing of return . payment of taxes etc. Failure to comply with any of the provisions of\nGST Laws will lead to cancellation of this order and / or subject to any other action as\nmanagement will deem fit."\n\n \n\n“The parties undertake and warrant to each other that they have not offered, given or agreed to CHANDRA MAKTHALA\ngive (and that they will not offer, give or agree to give) to any person any untawfut gift or Plant Materials Head\nconsideration of any kind as an inducement or reward for doing or forbearing to do anything in Hindalco Industries Limited\n‘elation to the obtaining of this Contract or the performance by the parties of their abligations Hirakud\n\nunder this Contract The parties warrant that they have in place, and undertake thal they will\ncomply with. policies and procedures to avoid the risk of bribery fas set out in any legistation in\nthe applicable jurisdiction to the respective Party) and fraud with its organization and in\n‘connection with its dealings wih other parties.”\n\n{you have any complain regarding the value violation, please mail to our Unit Value Standard Committee “hindaleohirakud. UVSC@adityabirla.com*\nAny kind of plastic packaging, plastic enclouser , bags etc must be avoided as per govt norms.\n\n \n\nRegd. office:\n\nHINDALCO INDUSTRIES LIMITED ni\n\nAhura Centre, 1st Floor, 8 Wing, Mahakali Caves Road Andheri(East), Mumbai 400093, india, ‘age 2 of 13\nTel No: 91 22 6691 7000, Email: hindalco@adityabirla com, Website: www adityabirla.com\n\nCorporate Identity No: L27020MH1958PLCO11238\n\x0c' ' \n\no\n\nGSTIN: 21AAACH1201R1ZZ\n\n-HIINDALCO INDUSTRIES LIMITED state Code: 21 - Ouisha\nBl ureau0 POWER, HIRAKUD-76® O16DIST:SAMBALPUR (ODISHA) St Rangaldvon\n\nPHONE: 0663-2481365, FAX: 0663-2481342 GST Commissionerate -Cuttack\n\n \n\nPURCHASE ORDER\n‘Vendor Code: U138 P.OINe P/PO/SRV/2122/0781 Date: 31-0EC-2021\nM/s UTKAL CONSTRUCTION Revise No Date:\n\nOrder Type: PURCHASE ORDER\n\nAT-MALIPADA Effective From 01/01/2022 To 31/12/2022\nPOTD! Price Basis\nSAMBALPUR, ORISSA, IN 768 016 Transportation arrangement\n\nShip to Location HIRAKUD - POWER\nEmail: antanyanipanda7?@gmail.com Carer\nFax: () Currency PINR\nContact: ANTARYAM! PANDA (+91) 9861236394 Hindalco Contact Person: SIODHARTH KUNDA,\n\nGSTIN; 21ANPPPS5428H1ZU —State:21 - Odisha Email of Contact Person: sidharth kunda@adityabirla.com\n\nRet: MANPOWER LTS\n\n \n\n \n\n \n\nOrder Unitof Rate/Urit value\nS
0	'O.61.9023\n\n \n\nVW\n\x0c' 'pos 2 wee la 4\noh RRC om Bs q\n\nRodiAdd\n\nae apace Sg\n\n \n\n \n\x0c' 'SERVICE BILL\n\nVendor: JAIDURGA CONSTRUCTION (J123 )\n\n \n\nPO. No. incentive & Penality Tas WO Value Total Value\n\n \n\nP/PO/SRV/1820/1161 © 24397.98 0.00 36493104 419329.09\n\n \n\n \n\n

Evaluation

Metrics

Label	Accuracy
all	1.0

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("Gopal2002/SERVICE_ZEON")
# Run inference
preds = model("ci lll ll Me

sy sa

< : Chet > 24.09 2? Lib Gs)
CTS 25M AS ow

           

pind
eo, YERIPHERLAL— —Sigh\
ATE. OUST EER ING re: vw
a cerpuer a Wa,
76.09.19 02 ym f tauren aoe ae ay Mi hid i c o) + :
f 24-09 «19 O° wo. “do -— mae oD wie - Ae 2 AC
” . a pie 1ay 4 qT
ie Oi. SOEs = = ple ak nyse a
29-09: 190 2W - ~ as 20 -¥. 44 oF ww +r An as
reearccene to 4. we Xs OL Ke get oe
HOt XK. 49 e cal de my.440 ini ed
o2xX 19.0 2mm “Ar a. gx. 440 De fe le ot
sy eam Pot A le eggoem po
0A. X. 2 Ale. Sa Wid ’ elt
o4 ‘4 0 2mm — 4: joe OR AA WE ay ea si
if c es

Hae.. 44 OL Wun.
( for.y. 14 0 2am. t

—4
Ae —4
a a QC
ope HOt wep A !
oq ke $4 0 2 Wh to —4e
40x. 44 © Lm qt ~—Ae Ye

15 xX a 0 7, WH “-

Lhe
|
&

4 1 A

6 &

 

\ \ yy
SA Lho®
i
a=
Ge

‘Q
“,

|

 
")

Training Details

Training Set Metrics

Training set	Min	Median	Max
Word count	2	284.6628	699

Label	Training Sample Count
0	30
1	24
2	32

Training Hyperparameters

batch_size: (32, 32)
num_epochs: (2, 2)
max_steps: -1
sampling_strategy: oversampling
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
0.0065	1	0.2818	-
0.3268	50	0.0374	-
0.6536	100	0.0053	-
0.9804	150	0.003	-
1.3072	200	0.0028	-
1.6340	250	0.0029	-
1.9608	300	0.0032	-

Framework Versions

Python: 3.10.12
SetFit: 1.0.3
Sentence Transformers: 2.2.2
Transformers: 4.35.2
PyTorch: 2.1.0+cu121
Datasets: 2.16.1
Tokenizers: 0.15.0

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}
}

Gopal2002
/

SERVICE_ZEON