fold_0

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

• Loss: 0.7257
• Unbiased Gwet Ac1: 0.2005
• Unbiased Kripp Alpha: 0.4220
• Qwk: 0.4537
• Mse: 0.7257
• Rmse: 0.8519

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: 2e-05
• train_batch_size: 64
• eval_batch_size: 64
• seed: 42
• 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: 100

Training results

Training Loss	Epoch	Step	Validation Loss	Unbiased Gwet Ac1	Unbiased Kripp Alpha	Qwk	Mse	Rmse
No log	1.0	2	9.3224	0.0	-0.8522	0.0	9.3224	3.0533
No log	2.0	4	8.4963	0.0	-0.8522	0.0	8.4963	2.9148
No log	3.0	6	6.8286	0.0	-0.8522	0.0	6.8286	2.6132
No log	4.0	8	5.6993	-0.0046	-0.7051	0.0316	5.6993	2.3873
No log	5.0	10	4.6042	0.0090	-0.7520	0.0115	4.6042	2.1457
No log	6.0	12	3.8284	0.0030	-0.7687	0.0039	3.8284	1.9566
No log	7.0	14	3.1647	0.0	-0.7771	0.0	3.1647	1.7790
No log	8.0	16	2.7221	0.0	-0.7771	0.0	2.7221	1.6499
No log	9.0	18	2.3650	-0.0062	-0.1858	0.1200	2.3650	1.5378
No log	10.0	20	1.9570	0.0040	-0.2959	0.0316	1.9570	1.3989
No log	11.0	22	1.5558	0.0040	-0.2959	0.0316	1.5558	1.2473
No log	12.0	24	1.2613	0.0040	-0.2959	0.0316	1.2613	1.1231
No log	13.0	26	1.1105	0.0040	-0.2959	0.0316	1.1105	1.0538
No log	14.0	28	1.1261	0.0040	-0.2959	0.0316	1.1261	1.0612
No log	15.0	30	0.9340	0.0287	-0.2019	0.0629	0.9340	0.9664
No log	16.0	32	0.7898	0.1477	0.3779	0.3711	0.7898	0.8887
No log	17.0	34	0.7760	0.1637	0.3636	0.3483	0.7760	0.8809
No log	18.0	36	0.9018	0.1017	0.0431	0.1820	0.9018	0.9496
No log	19.0	38	0.7286	0.1959	0.4185	0.4151	0.7286	0.8536
No log	20.0	40	0.5712	0.1956	0.4242	0.4338	0.5712	0.7558
No log	21.0	42	0.9187	0.1151	0.3136	0.3674	0.9187	0.9585
No log	22.0	44	0.9177	0.1130	0.3389	0.3952	0.9177	0.9580
No log	23.0	46	0.5279	0.1905	0.3940	0.4395	0.5279	0.7265
No log	24.0	48	0.5443	0.1463	0.3361	0.3972	0.5443	0.7377
No log	25.0	50	0.6213	0.2684	0.4825	0.4858	0.6213	0.7882
No log	26.0	52	0.6193	0.1786	0.3848	0.4269	0.6193	0.7870
No log	27.0	54	0.6122	0.0921	0.2514	0.3470	0.6122	0.7825
No log	28.0	56	0.6808	0.0785	0.2263	0.3334	0.6808	0.8251
No log	29.0	58	0.6123	0.1147	0.2885	0.3678	0.6123	0.7825
No log	30.0	60	0.5252	0.2216	0.4340	0.4683	0.5252	0.7247
No log	31.0	62	0.7128	0.1836	0.4512	0.4701	0.7128	0.8443
No log	32.0	64	0.7461	0.1652	0.4316	0.4571	0.7461	0.8638
No log	33.0	66	0.5773	0.1935	0.4372	0.4619	0.5773	0.7598
No log	34.0	68	0.6495	0.1757	0.3733	0.4448	0.6495	0.8059
No log	35.0	70	0.7579	0.1544	0.3724	0.4569	0.7579	0.8706
No log	36.0	72	0.7860	0.1118	0.3041	0.4048	0.7860	0.8865
No log	37.0	74	0.7672	0.1347	0.3655	0.4537	0.7672	0.8759
No log	38.0	76	0.6904	0.1750	0.3771	0.4540	0.6904	0.8309
No log	39.0	78	0.6266	0.1754	0.4103	0.4700	0.6266	0.7916
No log	40.0	80	0.6285	0.2001	0.4478	0.5096	0.6285	0.7928
No log	41.0	82	0.5839	0.2061	0.4706	0.5007	0.5839	0.7641
No log	42.0	84	0.6171	0.1988	0.4569	0.4767	0.6171	0.7856
No log	43.0	86	0.6589	0.2099	0.4405	0.4607	0.6589	0.8117
No log	44.0	88	0.6384	0.2255	0.4378	0.4559	0.6384	0.7990
No log	45.0	90	0.5937	0.2001	0.4571	0.5036	0.5937	0.7705
No log	46.0	92	0.6509	0.2014	0.4448	0.5178	0.6509	0.8068
No log	47.0	94	0.7105	0.1674	0.4020	0.4304	0.7105	0.8429
No log	48.0	96	0.8152	0.1525	0.3370	0.3724	0.8152	0.9029
No log	49.0	98	0.7114	0.1453	0.3471	0.4335	0.7114	0.8434
No log	50.0	100	0.8984	0.1138	0.2238	0.3973	0.8984	0.9478
No log	51.0	102	0.8349	0.1447	0.2919	0.4351	0.8349	0.9138
No log	52.0	104	0.6250	0.2248	0.4395	0.5089	0.6250	0.7906
No log	53.0	106	0.6989	0.1902	0.4233	0.4535	0.6989	0.8360
No log	54.0	108	0.8627	0.1583	0.3213	0.3775	0.8627	0.9288
No log	55.0	110	0.7257	0.2005	0.4220	0.4537	0.7257	0.8519

Framework versions

• Transformers 4.51.1
• Pytorch 2.5.1+cu124
• Datasets 3.5.0
• Tokenizers 0.21.0