CafeBERT_massive_crf_v2

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

• Loss: 5.5963
• Slot P: 0.0077
• Slot R: 0.0082
• Slot F1: 0.0079
• Slot Exact Match: 0.3246
• Intent Acc: 0.8751

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: 128
• eval_batch_size: 128
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 256
• optimizer: Use 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_ratio: 0.06
• num_epochs: 30
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Slot P	Slot R	Slot F1	Slot Exact Match	Intent Acc
No log	1.0	45	15.8863	0.0	0.0	0.0	0.4088	0.0733
69.6998	2.0	90	7.8039	0.0080	0.0076	0.0078	0.3438	0.5032
22.3589	3.0	135	4.5345	0.0091	0.0100	0.0095	0.3227	0.7846
10.4218	4.0	180	4.0667	0.0110	0.0111	0.0111	0.3384	0.8406
6.8199	5.0	225	3.8871	0.0092	0.0100	0.0096	0.3261	0.8623
5.4068	6.0	270	3.9234	0.0106	0.0117	0.0111	0.3212	0.8633
4.2552	7.0	315	4.0332	0.0115	0.0129	0.0122	0.3168	0.8657
3.5197	8.0	360	4.2753	0.0080	0.0088	0.0084	0.3222	0.8647
2.8374	9.0	405	4.6031	0.0099	0.0106	0.0102	0.3256	0.8701
2.2784	10.0	450	4.7992	0.0118	0.0129	0.0123	0.3237	0.8652
2.2784	11.0	495	5.0575	0.0118	0.0129	0.0123	0.3222	0.8652
1.8204	12.0	540	5.1371	0.0088	0.0094	0.0091	0.3266	0.8731
1.5073	13.0	585	5.4768	0.0109	0.0123	0.0116	0.3133	0.8677
1.275	14.0	630	5.5963	0.0077	0.0082	0.0079	0.3246	0.8751

Framework versions

• Transformers 4.55.0
• Pytorch 2.7.0+cu126
• Datasets 3.6.0
• Tokenizers 0.21.4