nepali-gec-binary-detector

DipeshChaudhary/nepali-gec-binary-detector

nepali-gec-binary-detector is a token classification model from DipeshChaudhary. Use it when you need labels on individual words, such as names. It is set up for transformers.

should probably proofread and complete it, then remove this comment. --

Downloads · 30 days

2% of all-time downloads

All-time downloads

496

Public

Parameters

124M

19.4 GB on disk

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.safetensors496 MB · 99%

At a glance

Task: Token Classification
Library: transformers
Model type: roberta
Access: Public
Created: Nov 9, 2025
Updated: Nov 9, 2025
SHA: bdb89e9d

Try a prompt

Base models

IRIIS-RESEARCH/RoBERTa_Nepali_125M

Model card

From the Hugging Face model README

nepali-gec-binary-detector

This model is a fine-tuned version of IRIIS-RESEARCH/RoBERTa_Nepali_125M on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 0.0407
Accuracy: 0.9874
Precision: 0.9338
Recall: 0.8248
F1: 0.8759
Sentence Accuracy: 0.8828

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 2e-06
train_batch_size: 1024
eval_batch_size: 1024
seed: 42
optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 500
num_epochs: 3
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1	Sentence Accuracy
0.1917	0.0787	1000	0.1144	0.9668	0.8991	0.4344	0.5858	0.7347
0.1083	0.1574	2000	0.0958	0.9708	0.8986	0.5197	0.6586	0.7625
0.0927	0.2361	3000	0.0791	0.9745	0.8839	0.6084	0.7207	0.7772
0.0807	0.3149	4000	0.0686	0.9780	0.8904	0.6760	0.7685	0.8015
0.0726	0.3936	5000	0.0623	0.9800	0.8910	0.7187	0.7956	0.8179
0.0671	0.4723	6000	0.0588	0.9813	0.9001	0.7369	0.8104	0.8286
0.0637	0.5510	7000	0.0559	0.9823	0.9017	0.7551	0.8219	0.8366
0.0608	0.6297	8000	0.0542	0.9830	0.9128	0.7587	0.8287	0.8434
0.0588	0.7084	9000	0.0520	0.9836	0.9135	0.7705	0.8359	0.8489
0.0569	0.7872	10000	0.0507	0.9841	0.9191	0.7748	0.8408	0.8532
0.0552	0.8659	11000	0.0493	0.9845	0.9221	0.7793	0.8447	0.8568
0.0543	0.9446	12000	0.0483	0.9849	0.9226	0.7866	0.8492	0.8597
0.053	1.0233	13000	0.0479	0.9851	0.9242	0.7901	0.8519	0.8623
0.0517	1.1020	14000	0.0467	0.9854	0.9207	0.7990	0.8556	0.8644
0.0511	1.1807	15000	0.0461	0.9856	0.9251	0.7992	0.8575	0.8671
0.0504	1.2594	16000	0.0452	0.9858	0.9205	0.8082	0.8607	0.8683
0.0497	1.3382	17000	0.0448	0.9860	0.9244	0.8073	0.8619	0.8700
0.0492	1.4169	18000	0.0443	0.9862	0.9289	0.8058	0.8630	0.8718
0.0485	1.4956	19000	0.0438	0.9863	0.9284	0.8098	0.8651	0.8731
0.0483	1.5743	20000	0.0436	0.9864	0.9292	0.8111	0.8662	0.8743
0.0477	1.6530	21000	0.0428	0.9866	0.9272	0.8156	0.8678	0.8751
0.0476	1.7317	22000	0.0431	0.9866	0.9333	0.8109	0.8678	0.8764
0.047	1.8105	23000	0.0426	0.9867	0.9311	0.8154	0.8694	0.8771
0.0466	1.8892	24000	0.0424	0.9868	0.9338	0.8144	0.8700	0.8782
0.0463	1.9679	25000	0.0421	0.9869	0.9326	0.8172	0.8711	0.8788
0.0459	2.0466	26000	0.0420	0.9870	0.9333	0.8178	0.8718	0.8794
0.0459	2.1253	27000	0.0416	0.9871	0.9308	0.8218	0.8729	0.8800
0.0455	2.2040	28000	0.0414	0.9871	0.9314	0.8223	0.8735	0.8803
0.0453	2.2827	29000	0.0414	0.9871	0.9327	0.8217	0.8737	0.8809
0.0452	2.3615	30000	0.0412	0.9872	0.9330	0.8223	0.8742	0.8813
0.045	2.4402	31000	0.0411	0.9872	0.9315	0.8245	0.8747	0.8815
0.045	2.5189	32000	0.0410	0.9873	0.9330	0.8237	0.8749	0.8819
0.0447	2.5976	33000	0.0409	0.9873	0.9328	0.8248	0.8755	0.8823
0.0448	2.6763	34000	0.0409	0.9873	0.9344	0.8234	0.8754	0.8825
0.0447	2.7550	35000	0.0407	0.9873	0.9330	0.8252	0.8758	0.8825
0.0445	2.8338	36000	0.0408	0.9873	0.9345	0.8237	0.8756	0.8827
0.0444	2.9125	37000	0.0407	0.9874	0.9336	0.8249	0.8759	0.8828
0.0446	2.9912	38000	0.0407	0.9874	0.9338	0.8248	0.8759	0.8828

Framework versions

Transformers 4.57.1
Pytorch 2.8.0+cu128
Datasets 4.4.1
Tokenizers 0.22.1

Task: Token Classification
Library: transformers
Type: roberta
Created: Nov 9, 2025
Updated: Nov 9, 2025