nci-scorer-nci-only-staging

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

• Loss: 0.0141
• Nci Mae: 0.2892
• Nci Acc Within 1: 0.9475
• Nci Labelled Share: 0.8551

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: 32
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED 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: 0.1
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Nci Mae	Nci Acc Within 1	Nci Labelled Share
No log	0.6780	200	0.0332	0.6290	0.8858	0.8550
No log	1.3559	400	0.0210	0.4261	0.9190	0.8550
0.0585	2.0339	600	0.0184	0.3806	0.9302	0.8550
0.0585	2.7119	800	0.0170	0.3440	0.9372	0.8550
0.0181	3.3898	1000	0.0191	0.3441	0.9285	0.8550
0.0181	4.0678	1200	0.0159	0.3232	0.9413	0.8550
0.0181	4.7458	1400	0.0157	0.3172	0.9432	0.8550
0.0137	5.4237	1600	0.0157	0.3149	0.9413	0.8550
0.0137	6.1017	1800	0.0149	0.3059	0.9455	0.8550
0.0114	6.7797	2000	0.0149	0.3038	0.9458	0.8550
0.0114	7.4576	2200	0.0150	0.3025	0.9459	0.8550
0.0114	8.1356	2400	0.0146	0.3004	0.9476	0.8550
0.0101	8.8136	2600	0.0143	0.2991	0.9473	0.8550
0.0101	9.4915	2800	0.0142	0.2969	0.9488	0.8550
0.0101	10.0	2950	0.0143	0.2969	0.9484	0.8550

Framework versions

• Transformers 5.15.1
• Pytorch 2.13.0+cu130
• Datasets 4.8.5
• Tokenizers 0.22.2