mpnet_token_cls_model

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

• Loss: 0.1318
• F1: 0.8327
• Precision: 0.8373
• Recall: 0.8281

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: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Use 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: 4
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	F1	Precision	Recall
0.174	0.1553	1000	0.1766	0.7930	0.8137	0.7734
0.1408	0.3105	2000	0.1469	0.8030	0.8143	0.7920
0.1326	0.4658	3000	0.1313	0.8187	0.8399	0.7985
0.1283	0.6210	4000	0.1308	0.8169	0.8188	0.8150
0.1259	0.7763	5000	0.1270	0.8195	0.8325	0.8069
0.12	0.9315	6000	0.1224	0.8162	0.8272	0.8055
0.1072	1.0868	7000	0.1221	0.8215	0.82	0.8230
0.1068	1.2420	8000	0.1216	0.8208	0.8234	0.8182
0.1022	1.3973	9000	0.1256	0.8234	0.8188	0.8281
0.1034	1.5526	10000	0.1217	0.8267	0.8292	0.8241
0.1051	1.7078	11000	0.1203	0.8288	0.8435	0.8146
0.1011	1.8631	12000	0.1246	0.8299	0.8284	0.8314
0.0917	2.0183	13000	0.1266	0.8248	0.8274	0.8223
0.0887	2.1736	14000	0.1213	0.8261	0.8260	0.8263
0.0863	2.3288	15000	0.1255	0.8272	0.8263	0.8281
0.0897	2.4841	16000	0.1265	0.8210	0.8302	0.8120
0.0835	2.6393	17000	0.1233	0.8299	0.8284	0.8314
0.0833	2.7946	18000	0.1259	0.8341	0.8398	0.8285
0.0829	2.9499	19000	0.1189	0.8328	0.8397	0.8259
0.0704	3.1051	20000	0.1308	0.8302	0.8290	0.8314
0.073	3.2604	21000	0.1273	0.8296	0.8330	0.8263
0.0711	3.4156	22000	0.1335	0.8304	0.8399	0.8212
0.0695	3.5709	23000	0.1325	0.8283	0.8353	0.8215
0.0708	3.7261	24000	0.1316	0.8319	0.8384	0.8255
0.0706	3.8814	25000	0.1318	0.8327	0.8373	0.8281

Framework versions

• Transformers 4.50.0
• Pytorch 2.5.1+cu121
• Datasets 2.21.0
• Tokenizers 0.21.4