sanigec-phase1

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

• Loss: 1.2194
• Detection F1: 0.6631
• Correction Accuracy: 0.9765

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: 32
• eval_batch_size: 128
• seed: 0
• gradient_accumulation_steps: 2
• total_train_batch_size: 64
• 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: cosine
• lr_scheduler_warmup_steps: 0.06
• num_epochs: 2.0

Training results

Training Loss	Epoch	Step	Validation Loss	Detection F1	Correction Accuracy
1.2737	0.0949	2500	1.2979	0.5029	0.9692
1.2399	0.1898	5000	1.2654	0.5270	0.9717
1.2310	0.2847	7500	1.2547	0.5397	0.9724
1.2229	0.3796	10000	1.2508	0.5879	0.9729
1.2143	0.4746	12500	1.2477	0.5704	0.9733
1.2158	0.5695	15000	1.2432	0.6009	0.9736
1.2086	0.6644	17500	1.2378	0.6042	0.9739
1.2075	0.7593	20000	1.2348	0.6143	0.9743
1.2045	0.8542	22500	1.2325	0.6165	0.9746
1.2036	0.9491	25000	1.2309	0.6170	0.9749
1.1978	1.0440	27500	1.2278	0.6337	0.9752
1.1943	1.1389	30000	1.2279	0.6477	0.9753
1.1949	1.2338	32500	1.2252	0.6500	0.9756
1.1915	1.3287	35000	1.2241	0.6438	0.9759
1.1902	1.4236	37500	1.2225	0.6487	0.9760
1.1896	1.5186	40000	1.2219	0.6554	0.9762
1.1902	1.6135	42500	1.2204	0.6603	0.9762
1.1895	1.7084	45000	1.2201	0.6614	0.9764
1.1869	1.8033	47500	1.2196	0.6624	0.9764
1.1865	1.8982	50000	1.2192	0.6622	0.9765
1.1894	1.9931	52500	1.2194	0.6631	0.9765
1.1894	2.0	52682	1.2194	0.6631	0.9765

Framework versions

• Transformers 5.12.1
• Pytorch 2.5.1+cu121
• Datasets 5.0.0
• Tokenizers 0.22.2