router-mmBERT-base-3e-5-batch32

This model is a fine-tuned version of jhu-clsp/mmBERT-base on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.6479
• Accuracy: 0.6212
• Precision: 0.6206
• Recall: 0.6212
• F1: 0.6194

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: 3e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• 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
• num_epochs: 2

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
0.7063	0.0232	50	0.7010	0.5262	0.5706	0.5262	0.4763
0.7301	0.0465	100	0.6931	0.5765	0.5767	0.5765	0.5766
0.6609	0.0697	150	0.7301	0.4848	0.5680	0.4848	0.3430
0.6551	0.0929	200	0.7030	0.5395	0.6050	0.5395	0.4140
0.6929	0.1162	250	0.6861	0.5417	0.5651	0.5417	0.5212
0.666	0.1394	300	0.6632	0.6102	0.6094	0.6102	0.6080
0.6708	0.1626	350	0.7310	0.5671	0.6685	0.5671	0.4676
0.6805	0.1859	400	0.6653	0.5897	0.5944	0.5897	0.5891
0.6947	0.2091	450	0.7120	0.5157	0.5950	0.5157	0.4295
0.6248	0.2323	500	0.6713	0.5953	0.6370	0.5953	0.5453
0.6988	0.2556	550	0.6682	0.5848	0.6411	0.5848	0.5193
0.6843	0.2788	600	0.7089	0.5174	0.6174	0.5174	0.4230
0.7576	0.3020	650	0.7033	0.5130	0.6062	0.5130	0.4165
0.6426	0.3253	700	0.6584	0.5964	0.5974	0.5964	0.5966
0.6772	0.3485	750	0.6716	0.5682	0.5997	0.5682	0.5476
0.6944	0.3717	800	0.6593	0.6168	0.6175	0.6168	0.6170
0.6491	0.3950	850	0.6546	0.6113	0.6293	0.6113	0.5861
0.6491	0.4182	900	0.6616	0.5958	0.5986	0.5958	0.5958
0.6337	0.4414	950	0.6718	0.6085	0.6116	0.6085	0.6084
0.6789	0.4647	1000	0.6569	0.6140	0.6272	0.6140	0.5940
0.7058	0.4879	1050	0.6596	0.6041	0.6039	0.6041	0.6040
0.7226	0.5112	1100	0.6544	0.5991	0.6087	0.5991	0.5786
0.6477	0.5344	1150	0.6754	0.5511	0.5936	0.5511	0.5167
0.6556	0.5576	1200	0.6955	0.5461	0.5982	0.5461	0.5023
0.6757	0.5809	1250	0.6546	0.6030	0.6027	0.6030	0.6028
0.6885	0.6041	1300	0.6620	0.5919	0.6076	0.5919	0.5853
0.6325	0.6273	1350	0.6538	0.6057	0.6221	0.6057	0.5803
0.6451	0.6506	1400	0.6691	0.5693	0.6067	0.5693	0.5448
0.6791	0.6738	1450	0.6569	0.5980	0.6005	0.5980	0.5981
0.6814	0.6970	1500	0.6572	0.6074	0.6106	0.6074	0.6073
0.6363	0.7203	1550	0.6777	0.5748	0.5983	0.5748	0.5613
0.6725	0.7435	1600	0.6482	0.6173	0.6175	0.6173	0.6133
0.6086	0.7667	1650	0.6557	0.6052	0.6080	0.6052	0.6052
0.6532	0.7900	1700	0.6549	0.6080	0.6295	0.6080	0.5788
0.6432	0.8132	1750	0.6547	0.6085	0.6108	0.6085	0.5997
0.6259	0.8364	1800	0.6517	0.6091	0.6100	0.6091	0.6093
0.6557	0.8597	1850	0.6462	0.6151	0.6154	0.6151	0.6106
0.6279	0.8829	1900	0.6455	0.6118	0.6125	0.6118	0.6062
0.6886	0.9061	1950	0.6548	0.6124	0.6157	0.6124	0.6122
0.6182	0.9294	2000	0.6472	0.6184	0.6177	0.6184	0.6171
0.6937	0.9526	2050	0.6459	0.6113	0.6200	0.6113	0.5950
0.6525	0.9758	2100	0.6493	0.6146	0.6204	0.6146	0.6021
0.6354	0.9991	2150	0.6480	0.6146	0.6139	0.6146	0.6136
0.7016	1.0223	2200	0.6455	0.6251	0.6285	0.6251	0.6169
0.6634	1.0455	2250	0.6465	0.6157	0.6203	0.6157	0.6049
0.6256	1.0688	2300	0.6448	0.6201	0.6194	0.6201	0.6183
0.6227	1.0920	2350	0.6451	0.6195	0.6226	0.6195	0.6113
0.6402	1.1152	2400	0.6571	0.6085	0.6250	0.6085	0.5838
0.6157	1.1385	2450	0.6561	0.6057	0.6061	0.6057	0.6059
0.6129	1.1617	2500	0.6549	0.6201	0.6250	0.6201	0.6097
0.6632	1.1849	2550	0.6468	0.6140	0.6133	0.6140	0.6127
0.6002	1.2082	2600	0.6535	0.6074	0.6076	0.6074	0.6075
0.6406	1.2314	2650	0.6536	0.6024	0.6016	0.6024	0.6013
0.6015	1.2546	2700	0.6603	0.5997	0.6049	0.5997	0.5988
0.6212	1.2779	2750	0.6595	0.6251	0.6283	0.6251	0.6172
0.6146	1.3011	2800	0.6656	0.5875	0.6011	0.5875	0.5819
0.6407	1.3243	2850	0.6646	0.6063	0.6090	0.6063	0.6063
0.6172	1.3476	2900	0.6722	0.5964	0.6072	0.5964	0.5927
0.5796	1.3708	2950	0.6527	0.6201	0.6197	0.6201	0.6173
0.6513	1.3941	3000	0.6570	0.6080	0.6072	0.6080	0.6070
0.6471	1.4173	3050	0.6524	0.6245	0.6301	0.6245	0.6139
0.6176	1.4405	3100	0.6563	0.6289	0.6367	0.6289	0.6168
0.5867	1.4638	3150	0.6567	0.6218	0.6233	0.6218	0.6157
0.6221	1.4870	3200	0.6566	0.6102	0.6095	0.6102	0.6094
0.5836	1.5102	3250	0.6544	0.6063	0.6058	0.6063	0.6059
0.6173	1.5335	3300	0.6542	0.6041	0.6042	0.6041	0.6041
0.5963	1.5567	3350	0.6557	0.6234	0.6276	0.6234	0.6142
0.6362	1.5799	3400	0.6521	0.6223	0.6240	0.6223	0.6161
0.6366	1.6032	3450	0.6492	0.6234	0.6267	0.6234	0.6153
0.6035	1.6264	3500	0.6525	0.6074	0.6085	0.6074	0.6076
0.6701	1.6496	3550	0.6485	0.6223	0.6243	0.6223	0.6156
0.6376	1.6729	3600	0.6483	0.6207	0.6201	0.6207	0.6186
0.5751	1.6961	3650	0.6474	0.6223	0.6229	0.6223	0.6178
0.6204	1.7193	3700	0.6492	0.6201	0.6194	0.6201	0.6187
0.6822	1.7426	3750	0.6488	0.6190	0.6183	0.6190	0.6175
0.6743	1.7658	3800	0.6477	0.6223	0.6220	0.6223	0.6196
0.6085	1.7890	3850	0.6474	0.6251	0.6257	0.6251	0.6207
0.5896	1.8123	3900	0.6482	0.6195	0.6188	0.6195	0.6182
0.6382	1.8355	3950	0.6472	0.6218	0.6212	0.6218	0.6197
0.6346	1.8587	4000	0.6478	0.6212	0.6206	0.6212	0.6192
0.5711	1.8820	4050	0.6482	0.6218	0.6211	0.6218	0.6201
0.6398	1.9052	4100	0.6483	0.6223	0.6217	0.6223	0.6206
0.5947	1.9284	4150	0.6480	0.6207	0.6200	0.6207	0.6191
0.7037	1.9517	4200	0.6480	0.6218	0.6211	0.6218	0.6201
0.5602	1.9749	4250	0.6478	0.6223	0.6217	0.6223	0.6206
0.6186	1.9981	4300	0.6479	0.6212	0.6206	0.6212	0.6194

Framework versions

• Transformers 4.57.1
• Pytorch 2.8.0+cu128
• Datasets 4.2.0
• Tokenizers 0.22.1