Open-source classifier for classifying bank transactions and statements - Free deployment to achieve accurate classification of French bank transaction statements

Home

Bank Transactions Statements Classification

Developed by moctarsmal

Bank transaction classifier fine-tuned based on Flaubert small French model

Text Classification

Transformers

Open Source License:MIT #Financial Text Classification #French Pretraining Fine-tuning #Transaction Flow Analysis

Downloads 85

Release Time : 10/11/2023

Model Overview

This model is specifically designed for classifying bank transactions, fine-tuned based on the Flaubert small French model, achieving 76.83% accuracy on the evaluation set.

Model Features

High Accuracy Classification

Achieves 76.83% accuracy on bank transaction classification tasks

French Language Optimization

Fine-tuned based on the French pretrained model Flaubert

Lightweight Model

Uses a small model architecture suitable for resource-constrained environments

Model Capabilities

Bank Transaction Classification

French Text Processing

Financial Text Analysis

Use Cases

Banking Operations

Automatic Transaction Classification

Automatically classifies bank transactions into predefined categories

76.83% accuracy

Financial Data Analysis

Supports automated classification and analysis of financial data

🚀 bank-transactions-statements-classification

This model is designed for classifying bank transaction statements. It's a fine - tuned version of an existing model, offering improved performance on relevant tasks.

🚀 Quick Start

This model is a fine - tuned version of flaubert/flaubert_small_cased on the None dataset. It achieves the following results on the evaluation set:

Loss: 1.0458
Accuracy: 0.7683
F1 Macro: 0.7945
F1 Weighted: 0.7635

📚 Documentation

Training and evaluation data

More information needed

Model description

More information needed

Intended uses & limitations

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 2e - 05
train_batch_size: 32
eval_batch_size: 32
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon = 1e - 08
lr_scheduler_type: linear
num_epochs: 15

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1 Macro	F1 Weighted
No log	0.29	50	3.6955	0.1012	0.0238	0.0880
No log	0.58	100	3.2965	0.2150	0.0569	0.1598
No log	0.87	150	3.1122	0.2530	0.0833	0.1889
No log	1.16	200	2.6838	0.3622	0.1800	0.3051
No log	1.45	250	2.5128	0.3808	0.1938	0.3139
No log	1.74	300	2.1573	0.4913	0.3241	0.4522
No log	2.03	350	2.0208	0.5220	0.3910	0.4832
No log	2.33	400	2.0454	0.5053	0.4090	0.4613
No log	2.62	450	1.7601	0.5599	0.4682	0.5303
3.1338	2.91	500	1.6837	0.5965	0.5489	0.5736
3.1338	3.2	550	1.6337	0.5885	0.5744	0.5609
3.1338	3.49	600	1.4553	0.6491	0.6219	0.6322
3.1338	3.78	650	1.4483	0.6531	0.6441	0.6345
3.1338	4.07	700	1.4108	0.6625	0.6810	0.6522
3.1338	4.36	750	1.3241	0.6924	0.6999	0.6769
3.1338	4.65	800	1.3254	0.6824	0.6960	0.6703
3.1338	4.94	850	1.3349	0.6937	0.6952	0.6759
3.1338	5.23	900	1.2264	0.7057	0.7157	0.6931
3.1338	5.52	950	1.3012	0.6891	0.7061	0.6748
1.6259	5.81	1000	1.2756	0.7071	0.7224	0.6925
1.6259	6.1	1050	1.1432	0.7317	0.7440	0.7267
1.6259	6.4	1100	1.2014	0.7290	0.7434	0.7161
1.6259	6.69	1150	1.1029	0.7483	0.7656	0.7367
1.6259	6.98	1200	1.1643	0.7310	0.7470	0.7227
1.6259	7.27	1250	1.1112	0.7477	0.7561	0.7371
1.6259	7.56	1300	1.1662	0.7350	0.7668	0.7254
1.6259	7.85	1350	1.0756	0.7577	0.7823	0.7530
1.6259	8.14	1400	1.1390	0.7403	0.7657	0.7318
1.6259	8.43	1450	1.1555	0.7437	0.7637	0.7377
1.092	8.72	1500	1.1086	0.7437	0.7686	0.7384
1.092	9.01	1550	1.0789	0.7510	0.7780	0.7427
1.092	9.3	1600	1.0613	0.7543	0.7823	0.7492
1.092	9.59	1650	1.0750	0.7477	0.7701	0.7382
1.092	9.88	1700	1.1412	0.7423	0.7772	0.7349
1.092	10.17	1750	1.0580	0.7617	0.7918	0.7549
1.092	10.47	1800	1.0667	0.7670	0.7856	0.7580
1.092	10.76	1850	1.1344	0.7403	0.7757	0.7332
1.092	11.05	1900	1.0808	0.7603	0.7944	0.7571
1.092	11.34	1950	1.0367	0.7690	0.7932	0.7655
0.9029	11.63	2000	1.0921	0.7577	0.7861	0.7504
0.9029	11.92	2050	1.0833	0.7603	0.7912	0.7541
0.9029	12.21	2100	1.0523	0.7716	0.7968	0.7662
0.9029	12.5	2150	1.0467	0.7683	0.7939	0.7614
0.9029	12.79	2200	1.0515	0.7703	0.7987	0.7667
0.9029	13.08	2250	1.0604	0.7696	0.8020	0.7654
0.9029	13.37	2300	1.0900	0.7716	0.8002	0.7663
0.9029	13.66	2350	1.0348	0.7743	0.8009	0.7686
0.9029	13.95	2400	1.0495	0.7656	0.7929	0.7610
0.9029	14.24	2450	1.0411	0.7670	0.7956	0.7624
0.7924	14.53	2500	1.0458	0.7683	0.7945	0.7635
0.7924	14.83	2550	1.0401	0.7696	0.7982	0.7649

Framework versions

Transformers 4.34.0
Pytorch 2.0.1+cu118
Datasets 2.14.5
Tokenizers 0.14.1

📄 License

This model is released under the MIT license.

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご