ania3000/kubert-morph

trainer_output

This model is a fine-tuned version of ania3000/kubert-from_multilingual on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 2.0092
• Accuracy: 69.0856
• Sentence accuracy: 14.7059

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: 8
• eval_batch_size: 8
• 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
• num_epochs: 25

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Sentence accuracy
No log	1.0	51	2.6089	51.0885	4.9020
No log	2.0	102	2.0624	57.9100	6.8627
No log	3.0	153	1.9063	61.9739	9.8039
No log	4.0	204	1.8096	62.9898	9.8039
No log	5.0	255	1.8159	64.0058	10.7843
No log	6.0	306	1.8145	65.4572	10.7843
No log	7.0	357	1.7694	66.9086	11.7647
No log	8.0	408	1.8472	67.6343	13.7255
No log	9.0	459	1.8234	68.7954	10.7843
1.2156	10.0	510	1.8620	67.6343	16.6667
1.2156	11.0	561	1.9227	68.9405	14.7059
1.2156	12.0	612	1.9168	68.9405	16.6667
1.2156	13.0	663	1.9321	69.2308	15.6863
1.2156	14.0	714	1.9782	68.5051	15.6863
1.2156	15.0	765	2.0042	68.7954	14.7059
1.2156	16.0	816	2.0225	69.0856	13.7255
1.2156	17.0	867	2.0092	69.0856	14.7059

Framework versions

• Transformers 4.57.3
• Pytorch 2.11.0+cu128
• Datasets 4.0.0
• Tokenizers 0.22.2