finbert-ft-icar-a-mda-direct

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

• Loss: 2.6969
• Accuracy: 0.5238
• Precision: 0.5217
• Recall: 0.5057
• F1: 0.4885

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
1.4149	1.0	83	1.3897	0.3492	0.3407	0.3465	0.2913
1.0795	2.0	166	1.2172	0.3810	0.3806	0.3808	0.2989
0.9891	3.0	249	1.2202	0.3651	0.3639	0.3624	0.3012
0.9453	4.0	332	1.3485	0.3492	0.3462	0.3516	0.2667
0.8793	5.0	415	1.3342	0.3968	0.3583	0.3777	0.3338
0.8093	6.0	498	1.3721	0.3968	0.3802	0.3801	0.3722
0.7	7.0	581	1.5765	0.4444	0.3407	0.4216	0.3419
0.5749	8.0	664	1.5505	0.4444	0.4142	0.4252	0.4114
0.5182	9.0	747	1.6208	0.4444	0.4011	0.4189	0.3966
0.4217	10.0	830	1.6965	0.4603	0.4370	0.4422	0.4360
0.4098	11.0	913	1.7849	0.4603	0.4381	0.4473	0.4350
0.3067	12.0	996	1.9142	0.4444	0.4346	0.4365	0.4166
0.261	13.0	1079	2.0077	0.4286	0.4247	0.4207	0.4049
0.1653	14.0	1162	2.0661	0.4444	0.4452	0.4365	0.4177
0.1148	15.0	1245	2.1456	0.4921	0.4748	0.4740	0.4623
0.1325	16.0	1328	2.2233	0.4603	0.4483	0.4473	0.4345
0.0883	17.0	1411	2.2824	0.4762	0.4614	0.4581	0.4490
0.0906	18.0	1494	2.4244	0.4603	0.4426	0.4448	0.4354
0.0544	19.0	1577	2.4539	0.4603	0.4434	0.4473	0.4343
0.0432	20.0	1660	2.5901	0.4286	0.4310	0.4207	0.4048
0.0542	21.0	1743	2.5440	0.4603	0.4434	0.4473	0.4343
0.0179	22.0	1826	2.6234	0.4603	0.4483	0.4499	0.4324
0.0129	23.0	1909	2.7417	0.5079	0.5085	0.4899	0.4758
0.0284	24.0	1992	2.6969	0.5238	0.5217	0.5057	0.4885
0.0316	25.0	2075	2.8483	0.4921	0.4862	0.4765	0.4615
0.0089	26.0	2158	2.8401	0.4921	0.4748	0.4740	0.4623
0.0205	27.0	2241	2.9610	0.4762	0.4785	0.4632	0.4473
0.0045	28.0	2324	2.9589	0.4921	0.4843	0.4740	0.4624
0.0139	29.0	2407	2.9910	0.4762	0.4642	0.4632	0.4469
0.0113	30.0	2490	3.0800	0.4603	0.4515	0.4499	0.4321
0.0036	31.0	2573	3.1598	0.4762	0.4857	0.4657	0.4453
0.0044	32.0	2656	3.0963	0.4603	0.4482	0.4499	0.4319
0.0117	33.0	2739	3.1607	0.4444	0.4429	0.4365	0.4169
0.0084	34.0	2822	3.1653	0.4762	0.4628	0.4607	0.4484
0.0028	35.0	2905	3.2284	0.4286	0.4285	0.4207	0.4048
0.0025	36.0	2988	3.2340	0.4762	0.4726	0.4607	0.4487
0.0191	37.0	3071	3.2446	0.4444	0.4337	0.4340	0.4195
0.0082	38.0	3154	3.2743	0.4444	0.4414	0.4340	0.4198
0.0111	39.0	3237	3.2742	0.4444	0.4337	0.4340	0.4195

Framework versions

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