Gopal2002/Material_Receipt_Report_ZEON

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:

1. Fine-tuning a Sentence Transformer with contrastive learning.
2. 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: 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
0	'_\ni.\nSe\nNew\n~~\ned\nTy\nSw\nNe\nNw\ned\n2:\n\n \n\x0c' 'ne.\n\n \n \n\n \n \n\n \n\nbBo fy20 5 ‘ )\n- wi Pas BOOKING STATION\nstat” SURAT GEIS TRA: BSPORT HOT. LTE, DIMPLE COURT, 2ND FLOOR,\n H.O.: “VIRAJ IMPEX HOUSE”, 47, D' M= -toRoaD, + AT_OW. ER’S RISK oer\n' , a” MUMBAI - 400 009 Tel. : 4076 7676 sianan Gece i al CARGO iS INSUR BY CUSTOMER — PH, : 033-30821697, 22\n{ 1. Consignor’s Name & Address As. ExOme peas Br. Code\ndT ncuer\n
1	"Posatis ils. H\n\n \n\niS\nvs\na (uf\n\noe\n\n \n\n-\n\n \n\nSarichor Pls: q\n\nPea :\n\nITEM /\n\n1. Description/ Received Reject Delivered Retur\n\n \n \n\nSPARE TX. Phat\n\n(MARKETED BY MESAPD\n\nPact eta\n\n \n\nMATERIAL RECEIPT REPORT\n\n \n\n \n \n \n\n \n\nCUM nea\n\n00 LeTlooo 0.000\n\nPAS\n\n \n \n\nELT\n\nJUPLICATE FOR TRANSPORTE?-\nOGPY (EMGISE INVOICE) RECEIVED\n\nMite ariant Eee\n\nPRAM MUIMAFE RCL RE\n\n \n\n \n\nFrys\n\n \n\not\n\nSuds oT\n\n \n \n\npeas\n\nee ase\n\n. Tax Gelig\n\nGrand Tooke\n\ni\n\nRM\n\nRate/Unit\n\nMRR SUBMITTED\nwv\n\nITH PARTY'S INVIGCE\n\nEET RY MO SSO OT Soe ELS\n\nLS.\n\n \n\n \n\n \n\nWee\n\n7; Ae 18\n\nTrcic\n\ni\nSu\n\n~s\n\n“en\n\nnny\n\x0c" "«= ITER /\ncit BDescription/ Received\n\nms\n\n \n \n\n \n\nIces\n\ne to\n\ntea tae\n\nhoimeryh bea\n\nPorccheninernyh Qerkees\n\nRican dec\n\nrarer:\n\nPAD RP eAR eR\n\nMeare\n\n \n\nMATERIAL RECEIPT\n\n \n\nREPORT\n\n \n\nwe ie 7\nhe\n\nSeba.\nbh ETS\n\n \n\nReject Delivered Retur\n\nTESLA y’\n\n \n\n \n\n \n\nLF PIE\n\nTAIT a\n\nSUPLICATE FOR TRANSPORTER\nOGPY (EXGISE INVOICE) RECEIVED\n\noy\n\nf\n\n“soarewe Pk Beak\nree\n\nRAF

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/Material_Receipt_Report_ZEON")
# Run inference
preds = model("SOT Ue

 

         

oH

| ia

I
od

Hi

a

|
To) Sig Pere
a

al |g
&%
5)

wS\
eB
SB
“5
“O
S
€X

Bea

em

Pe eS

se aE a

4 |] | tat [ety

tt pe Ta
&
a

OK

¢

SRLS ia Leh coe

 
 
")

Training Details

Training Set Metrics

Training set	Min	Median	Max
Word count	1	182.1336	1108

Label	Training Sample Count
0	202
1	45

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
0.0007	1	0.2952	-
0.0371	50	0.2253	-
0.0742	100	0.1234	-
0.1114	150	0.0115	-
0.1485	200	0.0036	-
0.1856	250	0.0024	-
0.2227	300	0.0015	-
0.2598	350	0.0011	-
0.2970	400	0.0009	-
0.3341	450	0.0007	-
0.3712	500	0.0011	-
0.4083	550	0.0008	-
0.4454	600	0.0008	-
0.4826	650	0.0007	-
0.5197	700	0.0005	-
0.5568	750	0.0006	-
0.5939	800	0.0005	-
0.6310	850	0.0005	-
0.6682	900	0.0004	-
0.7053	950	0.0003	-
0.7424	1000	0.0004	-
0.7795	1050	0.0005	-
0.8166	1100	0.0004	-
0.8537	1150	0.0004	-
0.8909	1200	0.0005	-
0.9280	1250	0.0004	-
0.9651	1300	0.0003	-
1.0022	1350	0.0003	-
1.0393	1400	0.0003	-
1.0765	1450	0.0004	-
1.1136	1500	0.0003	-
1.1507	1550	0.0004	-
1.1878	1600	0.0004	-
1.2249	1650	0.0004	-
1.2621	1700	0.0003	-
1.2992	1750	0.0003	-
1.3363	1800	0.0003	-
1.3734	1850	0.0003	-
1.4105	1900	0.0003	-
1.4477	1950	0.0002	-
1.4848	2000	0.0003	-
1.5219	2050	0.0003	-
1.5590	2100	0.0003	-
1.5961	2150	0.0002	-
1.6333	2200	0.0003	-
1.6704	2250	0.0004	-
1.7075	2300	0.0004	-
1.7446	2350	0.0003	-
1.7817	2400	0.0002	-
1.8189	2450	0.0002	-
1.8560	2500	0.0003	-
1.8931	2550	0.0002	-
1.9302	2600	0.0003	-
1.9673	2650	0.0003	-

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}
}