finbert-ft-icar-a-v0.11

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

• Loss: 0.8809
• Accuracy: 0.8847
• Precision: 0.8742
• Recall: 0.8438
• F1: 0.8566

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: 3e-06
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 3
• total_train_batch_size: 3
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
2.2028	1.0	704	0.7380	0.7902	0.8014	0.7093	0.7293
1.6961	2.0	1408	0.8784	0.8280	0.8359	0.7532	0.7712
1.5125	3.0	2112	0.9310	0.8318	0.8383	0.7574	0.7740
1.1895	4.0	2816	0.9300	0.8374	0.8490	0.7531	0.7737
0.9869	5.0	3520	0.8736	0.8582	0.8511	0.8102	0.8255
0.719	6.0	4224	0.8385	0.8620	0.8524	0.8140	0.8292
0.5451	7.0	4928	0.8900	0.8752	0.8793	0.8253	0.8451
0.4815	8.0	5632	0.9242	0.8715	0.8577	0.8359	0.8452
0.3935	9.0	6336	0.8809	0.8847	0.8742	0.8438	0.8566
0.2789	10.0	7040	0.9693	0.8696	0.8542	0.8227	0.8355
0.1994	11.0	7744	1.0208	0.8733	0.8620	0.8396	0.8492
0.2142	12.0	8448	1.0157	0.8771	0.8714	0.8307	0.8463
0.1583	13.0	9152	1.0452	0.8733	0.8649	0.8276	0.8420
0.1566	14.0	9856	1.0171	0.8790	0.8766	0.8333	0.8500
0.1489	15.0	10560	1.0512	0.8809	0.8843	0.8271	0.8465
0.1127	16.0	11264	1.0802	0.8677	0.8488	0.8337	0.8404
0.0828	17.0	11968	1.0955	0.8752	0.8622	0.8338	0.8459
0.0833	18.0	12672	1.1333	0.8752	0.8531	0.8415	0.8467
0.0655	19.0	13376	1.1919	0.8771	0.8811	0.8283	0.8468
0.0799	20.0	14080	1.1709	0.8828	0.8799	0.8377	0.8543
0.0778	21.0	14784	1.2235	0.8733	0.8633	0.8283	0.8424
0.0431	22.0	15488	1.3049	0.8677	0.8607	0.8203	0.8364
0.038	23.0	16192	1.2969	0.8790	0.8760	0.8288	0.8454
0.081	24.0	16896	1.3389	0.8790	0.8738	0.8301	0.8462

Framework versions

• Transformers 4.52.4
• Pytorch 2.6.0+cu124
• Datasets 4.4.1
• Tokenizers 0.21.2