finebooks-complexity-regressor-v1

This model is a fine-tuned version of davanstrien/finebooks-page-router-v0 on the davanstrien/finebooks-complexity-distil dataset. It achieves the following results on the evaluation set:

• Loss: 0.0106
• Mse: 0.0106
• Mae: 0.0620
• Spearman: 0.6373
• Pearson: 0.8084

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: 5e-05
• train_batch_size: 32
• eval_batch_size: 64
• 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: cosine
• lr_scheduler_warmup_steps: 0.1
• num_epochs: 20.0

Training results

Training Loss	Epoch	Step	Validation Loss	Mse	Mae	Spearman	Pearson
0.0111	1.0	57	0.0122	0.0122	0.0705	0.5960	0.7759
0.0128	2.0	114	0.0174	0.0174	0.0794	0.5442	0.7385
0.0111	3.0	171	0.0144	0.0144	0.0764	0.6134	0.7944
0.0098	4.0	228	0.0115	0.0115	0.0640	0.6142	0.7920
0.0083	5.0	285	0.0132	0.0132	0.0693	0.5966	0.7756
0.0077	6.0	342	0.0119	0.0119	0.0663	0.6100	0.8131
0.0073	7.0	399	0.0116	0.0116	0.0683	0.6199	0.7883
0.0046	8.0	456	0.0147	0.0147	0.0680	0.5732	0.7318
0.0055	9.0	513	0.0126	0.0126	0.0640	0.6126	0.7768
0.0069	10.0	570	0.0162	0.0162	0.0730	0.5754	0.7412
0.0048	11.0	627	0.0138	0.0138	0.0661	0.6069	0.7479
0.0036	12.0	684	0.0110	0.0110	0.0649	0.6282	0.8009
0.0038	13.0	741	0.0107	0.0107	0.0627	0.6278	0.8061
0.0029	14.0	798	0.0106	0.0106	0.0620	0.6373	0.8084
0.0023	15.0	855	0.0128	0.0128	0.0648	0.6122	0.7656
0.0015	16.0	912	0.0133	0.0133	0.0649	0.6129	0.7599
0.0015	17.0	969	0.0126	0.0126	0.0647	0.6137	0.7714
0.0015	18.0	1026	0.0129	0.0129	0.0646	0.6170	0.7643
0.0024	19.0	1083	0.0130	0.0130	0.0646	0.6176	0.7637
0.0021	20.0	1140	0.0130	0.0130	0.0646	0.6185	0.7639

Framework versions

• Transformers 5.12.1
• Pytorch 2.12.1+cu130
• Datasets 5.0.0
• Tokenizers 0.22.2