fge-robos-qa-model

This model is a fine-tuned version of mrm8488/bert-base-spanish-wwm-cased-finetuned-spa-squad2-es on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.0088
• Model Preparation Time: 0.0077
• Exact: 55.7377
• F1: 82.8805
• Total: 915
• Hasans Exact: 55.7377
• Hasans F1: 82.8805
• Hasans Total: 915
• Best Exact: 55.7377
• Best Exact Thresh: 0.0
• Best F1: 82.8805
• Best F1 Thresh: 0.0

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: 64
• eval_batch_size: 64
• 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
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Model Preparation Time	Exact	F1	Total	Hasans Exact	Hasans F1	Hasans Total	Best Exact	Best Exact Thresh	Best F1	Best F1 Thresh
No log	1.0	58	0.8963	0.0077	54.2077	81.4281	915	54.2077	81.4281	915	54.2077	0.0	81.4281	0.0
No log	2.0	116	0.9578	0.0077	54.9727	82.3694	915	54.9727	82.3694	915	54.9727	0.0	82.3694	0.0
No log	3.0	174	1.0088	0.0077	55.7377	82.8805	915	55.7377	82.8805	915	55.7377	0.0	82.8805	0.0
No log	4.0	232	1.0865	0.0077	54.4262	81.7459	915	54.4262	81.7459	915	54.4262	0.0	81.7459	0.0
No log	5.0	290	1.2034	0.0077	53.7705	81.5328	915	53.7705	81.5328	915	53.7705	0.0	81.5328	0.0
No log	6.0	348	1.2822	0.0077	54.2077	81.9985	915	54.2077	81.9985	915	54.2077	0.0	81.9985	0.0
No log	7.0	406	1.3357	0.0077	54.2077	81.7294	915	54.2077	81.7294	915	54.2077	0.0	81.7294	0.0
No log	8.0	464	1.3738	0.0077	54.6448	81.9526	915	54.6448	81.9526	915	54.6448	0.0	81.9526	0.0
0.4292	9.0	522	1.4215	0.0077	54.7541	81.7385	915	54.7541	81.7385	915	54.7541	0.0	81.7385	0.0
0.4292	10.0	580	1.4342	0.0077	53.4426	81.3729	915	53.4426	81.3729	915	53.4426	0.0	81.3729	0.0

Framework versions

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