sincere-skink-211

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

• Loss: 0.2292
• Hamming Loss: 0.0656
• Zero One Loss: 0.3875
• Jaccard Score: 0.3258
• Hamming Loss Optimised: 0.0638
• Hamming Loss Threshold: 0.5944
• Zero One Loss Optimised: 0.3775
• Zero One Loss Threshold: 0.5944
• Jaccard Score Optimised: 0.3152
• Jaccard Score Threshold: 0.2751

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: 2.4994391789078344e-05
• train_batch_size: 4
• eval_batch_size: 4
• seed: 2024
• 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: linear
• num_epochs: 9

Training results

Training Loss	Epoch	Step	Validation Loss	Hamming Loss	Zero One Loss	Jaccard Score	Hamming Loss Optimised	Hamming Loss Threshold	Zero One Loss Optimised	Zero One Loss Threshold	Jaccard Score Optimised	Jaccard Score Threshold
0.2814	1.0	800	0.1793	0.0622	0.5162	0.4817	0.0607	0.4376	0.4425	0.3508	0.3393	0.2558
0.157	2.0	1600	0.1652	0.0575	0.3812	0.3328	0.0565	0.6526	0.3675	0.4638	0.3076	0.3993
0.121	3.0	2400	0.1694	0.0609	0.3962	0.3441	0.0578	0.6697	0.3988	0.3777	0.3055	0.2653
0.0954	4.0	3200	0.1847	0.062	0.3825	0.3320	0.0605	0.7112	0.3812	0.4056	0.3115	0.2369
0.0714	5.0	4000	0.1986	0.0644	0.3838	0.3286	0.0626	0.7004	0.3812	0.4671	0.3120	0.2433
0.0528	6.0	4800	0.2123	0.0651	0.38	0.3202	0.0617	0.8136	0.375	0.6080	0.3155	0.2947
0.0405	7.0	5600	0.2183	0.0649	0.3825	0.3295	0.0601	0.8621	0.3825	0.4639	0.3160	0.2485
0.0332	8.0	6400	0.2266	0.0665	0.3875	0.3282	0.0659	0.4056	0.385	0.4711	0.3176	0.2886
0.0267	9.0	7200	0.2292	0.0656	0.3875	0.3258	0.0638	0.5944	0.3775	0.5944	0.3152	0.2751

Framework versions

• Transformers 4.47.0
• Pytorch 2.5.1+cu124
• Datasets 3.1.0
• Tokenizers 0.21.0