metadata

base_model: ProsusAI/finbert
tags:
  - generated_from_trainer
metrics:
  - accuracy
  - f1
  - precision
  - recall
model-index:
  - name: finBert
    results: []

finBert

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

Loss: 0.3623
Accuracy: 0.868
F1: 0.8666
Precision: 0.8671
Recall: 0.8685

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: 5e-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
lr_scheduler_warmup_steps: 100
num_epochs: 3
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall
1.2152	0.0820	50	0.6140	0.7004	0.6545	0.7339	0.7020
0.5357	0.1639	100	0.4736	0.8265	0.8216	0.8270	0.8268
0.4421	0.2459	150	0.4189	0.8501	0.8466	0.8494	0.8505
0.4591	0.3279	200	0.4177	0.8390	0.8336	0.8417	0.8395
0.4382	0.4098	250	0.3876	0.8549	0.8537	0.8538	0.8550
0.4002	0.4918	300	0.4063	0.8434	0.8454	0.8503	0.8433
0.4085	0.5738	350	0.3818	0.8576	0.8574	0.8572	0.8579
0.3692	0.6557	400	0.3700	0.8608	0.8592	0.8598	0.8611
0.3686	0.7377	450	0.3730	0.8610	0.8604	0.8600	0.8613
0.3966	0.8197	500	0.3760	0.8566	0.8573	0.8584	0.8569
0.3729	0.9016	550	0.3533	0.8673	0.8667	0.8664	0.8675
0.3605	0.9836	600	0.3582	0.8632	0.8637	0.8645	0.8632
0.3471	1.0656	650	0.3671	0.8644	0.8643	0.8646	0.8645
0.3397	1.1475	700	0.3632	0.8658	0.8658	0.8656	0.8660
0.2977	1.2295	750	0.3737	0.8631	0.8619	0.8620	0.8634
0.3047	1.3115	800	0.3609	0.8656	0.8664	0.8673	0.8657
0.2817	1.3934	850	0.3816	0.8682	0.8661	0.8675	0.8685
0.2667	1.4754	900	0.3734	0.8690	0.8680	0.8686	0.8692
0.2809	1.5574	950	0.3523	0.8702	0.8706	0.8710	0.8703
0.2964	1.6393	1000	0.3621	0.8709	0.8693	0.8705	0.8711
0.2993	1.7213	1050	0.3502	0.8731	0.8728	0.8729	0.8732
0.3153	1.8033	1100	0.3532	0.8723	0.8703	0.8718	0.8726
0.3089	1.8852	1150	0.3581	0.8721	0.8710	0.8711	0.8724
0.3027	1.9672	1200	0.3513	0.8725	0.8712	0.8719	0.8727
0.2294	2.0492	1250	0.3673	0.8714	0.8715	0.8715	0.8715
0.2321	2.1311	1300	0.3630	0.8716	0.8718	0.8727	0.8716
0.2003	2.2131	1350	0.3951	0.872	0.8714	0.8711	0.8722
0.1948	2.2951	1400	0.3912	0.8724	0.8721	0.8719	0.8726
0.1939	2.3770	1450	0.3871	0.8733	0.8734	0.8734	0.8735
0.1897	2.4590	1500	0.3937	0.8738	0.8732	0.8730	0.8741
0.1951	2.5410	1550	0.3915	0.8711	0.8702	0.8704	0.8713
0.193	2.6230	1600	0.3902	0.8732	0.8729	0.8727	0.8734
0.1982	2.7049	1650	0.3903	0.8734	0.8734	0.8733	0.8736
0.1836	2.7869	1700	0.3929	0.8734	0.8732	0.8731	0.8736
0.2112	2.8689	1750	0.3927	0.8742	0.8738	0.8738	0.8743
0.1937	2.9508	1800	0.3899	0.8738	0.8739	0.8741	0.8740

Framework versions

Transformers 4.42.4
Pytorch 2.3.1+cu121
Tokenizers 0.19.1