edos_taskA_llama_allyears_lora2

This model is a fine-tuned version of meta-llama/Llama-3.2-1B on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.2633
• Accuracy: 0.9273
• F1 Macro: 0.9002

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: 0.0001
• train_batch_size: 16
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• 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: cosine
• lr_scheduler_warmup_ratio: 0.06
• num_epochs: 1.2
• label_smoothing_factor: 0.05

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1 Macro
0.9769	0.1068	100	0.4569	0.8135	0.7665
0.8868	0.2137	200	0.3733	0.8525	0.7959
0.8317	0.3205	300	0.3601	0.873	0.8254
0.7603	0.4274	400	0.3358	0.8845	0.8321
0.8641	0.5342	500	0.3237	0.8925	0.8481
0.6726	0.6410	600	0.2946	0.908	0.8730
0.6893	0.7479	700	0.2917	0.908	0.8621
0.6251	0.8547	800	0.2781	0.916	0.8855
0.636	0.9615	900	0.2657	0.9275	0.8979
0.499	1.0684	1000	0.2631	0.927	0.8991
0.4897	1.1752	1100	0.2596	0.928	0.9000

Framework versions

• PEFT 0.17.1
• Transformers 4.56.2
• Pytorch 2.8.0+cu126
• Datasets 4.1.1
• Tokenizers 0.22.0