mt5-small-xasqp-rest16

This model is a fine-tuned version of google/mt5-small on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.4130
• Precision: 0.379
• Recall: 0.4242
• F1: 0.4003

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: 0.0003
• train_batch_size: 16
• eval_batch_size: 16
• 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: 15

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1
No log	1.0	92	0.9849	0.021	0.0324	0.0255
No log	2.0	184	0.6488	0.0706	0.1635	0.0986
No log	3.0	276	0.4987	0.2033	0.2555	0.2264
No log	4.0	368	0.4678	0.1592	0.3203	0.2127
No log	5.0	460	0.4411	0.238	0.356	0.2853
1.8244	6.0	552	0.4291	0.3047	0.368	0.3333
1.8244	7.0	644	0.4419	0.3007	0.3765	0.3343
1.8244	8.0	736	0.4201	0.3382	0.3969	0.3652
1.8244	9.0	828	0.4162	0.3358	0.3901	0.3609
1.8244	10.0	920	0.4233	0.3378	0.385	0.3599
0.3879	11.0	1012	0.4216	0.3645	0.4123	0.3869
0.3879	12.0	1104	0.4139	0.3672	0.4191	0.3914
0.3879	13.0	1196	0.4127	0.3847	0.4208	0.402
0.3879	14.0	1288	0.4152	0.3714	0.4208	0.3946
0.3879	15.0	1380	0.4130	0.379	0.4242	0.4003

Framework versions

• Transformers 4.51.3
• Pytorch 2.6.0+cu124
• Datasets 3.6.0
• Tokenizers 0.21.1