tlocvsdyspneaTask-unsup-Qwen3-8B-datav3

This model is a fine-tuned version of ferrazzipietro/unsup-Qwen3-8B-datav3 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.2780
• F1 Micro: 0.7023
• F1 Macro: 0.3545
• F1 Weighted: 0.7670

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: 8
• eval_batch_size: 16
• seed: 42
• distributed_type: multi-GPU
• gradient_accumulation_steps: 8
• total_train_batch_size: 64
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-07 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss	F1 Micro	F1 Macro	F1 Weighted
1.7241	0.1067	2	1.7247	0.0	0.0	0.0
1.7241	0.2133	4	1.7052	0.0	0.0	0.0
1.6512	0.32	6	1.6111	0.0	0.0	0.0
1.6512	0.4267	8	1.4920	0.0	0.0	0.0
1.4595	0.5333	10	1.4084	0.0	0.0	0.0
1.4595	0.64	12	1.3623	0.5798	0.0176	0.6584
1.4595	0.7467	14	1.3383	0.4183	0.2734	0.4595
1.3499	0.8533	16	1.3242	0.6031	0.3366	0.6050
1.3499	0.96	18	1.3148	0.4261	0.2767	0.4660
1.3208	1.0533	20	1.3107	0.6693	0.3256	0.7212
1.3208	1.16	22	1.3027	0.5953	0.3728	0.7463
1.3208	1.2667	24	1.2951	0.6868	0.4511	0.7421
1.2644	1.3733	26	1.2907	0.5428	0.3585	0.6892
1.2644	1.48	28	1.2874	0.6984	0.3504	0.7606
1.2348	1.5867	30	1.2857	0.6965	0.3587	0.7709
1.2348	1.6933	32	1.2841	0.6401	0.3838	0.7784
1.2348	1.8	34	1.2825	0.6926	0.3578	0.7689
1.2104	1.9067	36	1.2813	0.5506	0.3617	0.6942
1.2104	2.0	38	1.2803	0.7062	0.3554	0.7689
1.2345	2.1067	40	1.2796	0.7043	0.3550	0.7680
1.2345	2.2133	42	1.2791	0.6459	0.3836	0.7812
1.2345	2.32	44	1.2790	0.5058	0.3356	0.6161
1.2058	2.4267	46	1.2784	0.6537	0.3787	0.7806
1.2058	2.5333	48	1.2783	0.6984	0.3559	0.7678
1.2274	2.64	50	1.2781	0.7004	0.3541	0.7660
1.2274	2.7467	52	1.2780	0.7023	0.3545	0.7670
1.2274	2.8533	54	1.2780	0.7023	0.3545	0.7670

Framework versions

• PEFT 0.18.1
• Transformers 4.51.0
• Pytorch 2.8.0+cu128
• Datasets 3.6.0
• Tokenizers 0.21.0