TisaleoGPT2Bot

This model is a fine-tuned version of DeepESP/gpt2-spanish on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.5536
• Perplexity: 13564.8888
• Exact Match: 0.6372
• F1 Score: 0.0009
• Coverage Rate: 0.0
• Context Errors: 13888

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: 4
• eval_batch_size: 4
• 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
• lr_scheduler_warmup_steps: 500
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Perplexity	Exact Match	F1 Score	Coverage Rate	Context Errors
1.1224	1.7921	500	0.6410	13079.0079	0.6372	0.0011	0.0	13888
0.479	3.5842	1000	0.5560	13364.5681	0.6372	0.0009	0.0	13888
0.3067	5.3763	1500	0.5441	13318.3878	0.6372	0.0009	0.0	13888
0.2065	7.1685	2000	0.5526	13633.0373	0.6372	0.0009	0.0	13888
0.1561	8.9606	2500	0.5534	13571.6876	0.6372	0.0009	0.0	13888

Framework versions

• Transformers 4.47.1
• Pytorch 2.5.1+cu124
• Datasets 3.2.0
• Tokenizers 0.21.0