mdeberta-qasina-LR-2e-05-BS-16-ML-384 – AI Model by mhdafifan | AlphaNeural AI

You can deploy this model and start earning money today!

mhdafifan/
mdeberta-qasina-LR-2e-05-BS-16-ML-384

like

0

question-answering

generated_from_trainer

microsoft/mdeberta-v3-base

endpoints_compatible

Views

No views yet

mdeberta-qasina-LR-2e-05-BS-16-ML-384

This model is a fine-tuned version of microsoft/mdeberta-v3-base on the QASiNa dataset. It achieves the following results on the evaluation set:

Loss: 1.8233
F1: 56.5003
Exact Match: 41.5094

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: 16
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_ratio: 0.2
num_epochs: 7

Training results

Training Loss	Epoch	Step	Validation Loss	F1	Exact Match
No log	1.0	46	5.0143	6.7253	1.8868
No log	2.0	92	2.6005	36.303	25.4717
No log	3.0	138	1.8435	52.2863	39.6226
No log	4.0	184	1.7622	55.2724	39.6226
No log	5.0	230	1.8261	55.9947	40.566
No log	6.0	276	1.8186	58.1579	43.3962
No log	7.0	322	1.8233	56.5003	41.5094

Framework versions

Transformers 4.57.1
Pytorch 2.9.1+cu128
Datasets 4.4.1
Tokenizers 0.22.1