mdeberta-semeval25_narratives_fold3

This model is a fine-tuned version of microsoft/mdeberta-v3-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 4.2072
• Precision Samples: 0.3215
• Recall Samples: 0.7807
• F1 Samples: 0.4308
• Precision Macro: 0.6748
• Recall Macro: 0.4927
• F1 Macro: 0.2710
• Precision Micro: 0.2999
• Recall Micro: 0.7380
• F1 Micro: 0.4264
• Precision Weighted: 0.4569
• Recall Weighted: 0.7380
• F1 Weighted: 0.3669

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: 32
• eval_batch_size: 32
• seed: 42
• 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: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Precision Samples	Recall Samples	F1 Samples	Precision Macro	Recall Macro	F1 Macro	Precision Micro	Recall Micro	F1 Micro	Precision Weighted	Recall Weighted	F1 Weighted
5.6486	1.0	19	5.3336	0.3483	0.1721	0.1761	0.9485	0.1103	0.0885	0.2808	0.1513	0.1966	0.8757	0.1513	0.0961
5.1546	2.0	38	5.1500	0.2167	0.4798	0.2793	0.8504	0.2332	0.1144	0.2138	0.4244	0.2843	0.6587	0.4244	0.1537
4.8699	3.0	57	4.9488	0.2564	0.5453	0.3324	0.7640	0.2641	0.1615	0.2495	0.4649	0.3247	0.5390	0.4649	0.2268
4.5166	4.0	76	4.6712	0.2862	0.6829	0.3830	0.7666	0.3645	0.1962	0.2768	0.6089	0.3806	0.5521	0.6089	0.2853
4.6349	5.0	95	4.4841	0.3041	0.7471	0.4094	0.7060	0.4314	0.2415	0.2890	0.6900	0.4074	0.4760	0.6900	0.3405
4.4166	6.0	114	4.3419	0.3141	0.7724	0.4210	0.7013	0.4581	0.2433	0.2912	0.7232	0.4153	0.4724	0.7232	0.3455
3.9808	7.0	133	4.2831	0.3158	0.7755	0.4256	0.6779	0.4803	0.2690	0.2942	0.7306	0.4195	0.4555	0.7306	0.3618
4.051	8.0	152	4.2851	0.3114	0.7416	0.4174	0.6723	0.4709	0.2650	0.2961	0.7048	0.4170	0.4528	0.7048	0.3579
4.0759	9.0	171	4.2030	0.3236	0.7830	0.4340	0.6801	0.4942	0.2758	0.3009	0.7417	0.4281	0.4611	0.7417	0.3710
4.3531	10.0	190	4.2072	0.3215	0.7807	0.4308	0.6748	0.4927	0.2710	0.2999	0.7380	0.4264	0.4569	0.7380	0.3669

Framework versions

• Transformers 4.46.0
• Pytorch 2.3.1
• Datasets 2.21.0
• Tokenizers 0.20.1