rlcc-aroma-upsample_replacement-absa-None

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

• Loss: 2.2234
• Accuracy: 0.7634
• F1 Macro: 0.6921
• Precision Macro: 0.6916
• Recall Macro: 0.6932
• Total Tf: [313, 97, 1133, 97]

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-05
• train_batch_size: 64
• eval_batch_size: 64
• seed: 42
• optimizer: Use adamw_torch 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: 51
• num_epochs: 25

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1 Macro	Precision Macro	Recall Macro	Total Tf
1.1142	1.0	52	1.0920	0.6220	0.4931	0.4924	0.4979	[255, 155, 1075, 155]
0.9901	2.0	104	1.1432	0.6634	0.5654	0.5637	0.5704	[272, 138, 1092, 138]
0.7632	3.0	156	1.0810	0.7049	0.6224	0.6336	0.6433	[289, 121, 1109, 121]
0.5227	4.0	208	1.1417	0.7439	0.6678	0.6699	0.6845	[305, 105, 1125, 105]
0.3542	5.0	260	1.1937	0.7463	0.6687	0.6683	0.6747	[306, 104, 1126, 104]
0.2719	6.0	312	1.4854	0.7146	0.6309	0.6367	0.6601	[293, 117, 1113, 117]
0.1594	7.0	364	1.4395	0.7463	0.6703	0.6703	0.6764	[306, 104, 1126, 104]
0.1511	8.0	416	1.6307	0.7390	0.6641	0.6625	0.6820	[303, 107, 1123, 107]
0.1036	9.0	468	1.6144	0.7683	0.6932	0.6974	0.6915	[315, 95, 1135, 95]
0.1029	10.0	520	1.7708	0.7683	0.6992	0.6976	0.7064	[315, 95, 1135, 95]
0.0859	11.0	572	2.0319	0.7439	0.6678	0.6756	0.6932	[305, 105, 1125, 105]
0.062	12.0	624	1.8619	0.7634	0.6901	0.6926	0.6887	[313, 97, 1133, 97]
0.0489	13.0	676	2.0523	0.7610	0.6878	0.6874	0.6892	[312, 98, 1132, 98]
0.0409	14.0	728	2.2591	0.7439	0.6664	0.6645	0.6741	[305, 105, 1125, 105]
0.0226	15.0	780	2.2234	0.7634	0.6921	0.6916	0.6932	[313, 97, 1133, 97]

Framework versions

• Transformers 4.47.0
• Pytorch 2.5.1+cu121
• Datasets 3.2.0
• Tokenizers 0.21.0