gonzalobenegas/gpn-star-maize-v1

fire_1/0.05/128/64/True/defined.phastCons.percentile-75_0.05_0.001/small/0.1/42/50000/True/True/0.1

This model is a fine-tuned version of on the results/dataset/128/64/True/defined.phastCons.percentile-75_0.05_0.001 dataset. It achieves the following results on the evaluation set:

• Loss: 0.1115

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: 0.0001
• train_batch_size: 32
• eval_batch_size: 128
• seed: 42
• distributed_type: multi-GPU
• num_devices: 8
• gradient_accumulation_steps: 2
• total_train_batch_size: 512
• total_eval_batch_size: 1024
• 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: cosine
• lr_scheduler_warmup_steps: 2500
• training_steps: 50000

Training results

Training Loss	Epoch	Step	Validation Loss
0.1298	1.9450	5000	0.1270
0.1192	3.8897	10000	0.1170
0.1164	5.8345	15000	0.1148
0.1149	7.7792	20000	0.1128
0.1137	9.7240	25000	0.1121
0.1131	11.6687	30000	0.1120
0.1126	13.6135	35000	0.1124
0.1124	15.5583	40000	0.1117
0.1124	17.5030	45000	0.1112
0.1122	19.4478	50000	0.1116

Framework versions

• Transformers 4.56.2
• Pytorch 2.8.0+cu128
• Datasets 4.1.1
• Tokenizers 0.22.1