modelsent_test

This model is a fine-tuned version of albert/albert-base-v2 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.2379
• Accuracy: 0.9261
• F1: 0.9261
• Precision: 0.9261
• Recall: 0.9261
• Accuracy Label Negative: 0.9242
• Accuracy Label Positive: 0.9278

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: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• 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
• lr_scheduler_warmup_steps: 500
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall	Accuracy Label Negative	Accuracy Label Positive
0.5403	0.2442	100	0.5274	0.7611	0.7596	0.7728	0.7611	0.8535	0.6746
0.2673	0.4884	200	0.2806	0.8980	0.8980	0.8994	0.8980	0.9230	0.8746
0.247	0.7326	300	0.2610	0.9029	0.9024	0.9074	0.9029	0.8434	0.9586
0.2357	0.9768	400	0.2560	0.9084	0.9084	0.9096	0.9084	0.9318	0.8864
0.2094	1.2198	500	0.3127	0.9090	0.9089	0.9123	0.9090	0.9508	0.8698
0.1695	1.4640	600	0.2298	0.9188	0.9187	0.9189	0.9188	0.9053	0.9314
0.2024	1.7082	700	0.2218	0.9206	0.9206	0.9214	0.9206	0.9394	0.9030
0.1155	1.9524	800	0.2061	0.9236	0.9236	0.9236	0.9236	0.9192	0.9278
0.1361	2.1954	900	0.2299	0.9218	0.9218	0.9226	0.9218	0.9407	0.9041
0.1235	2.4396	1000	0.2668	0.9212	0.9212	0.9246	0.9212	0.9634	0.8817
0.084	2.6838	1100	0.2733	0.9218	0.9218	0.9240	0.9218	0.9545	0.8911
0.1326	2.9280	1200	0.2395	0.9249	0.9249	0.9249	0.9249	0.9192	0.9302

Framework versions

• Transformers 4.49.0
• Pytorch 2.5.1+cu124
• Datasets 3.3.2
• Tokenizers 0.21.0