metadata

library_name: transformers
license: cc-by-4.0
base_model: hon9kon9ize/bert-large-cantonese
tags:
  - generated_from_trainer
metrics:
  - accuracy
model-index:
  - name: bert-suicide-detection-hk-large-new
    results: []

bert-suicide-detection-hk-large-new

This model is a fine-tuned version of hon9kon9ize/bert-large-cantonese on the None dataset. It achieves the following results on the evaluation set:

Loss: 0.2903
Accuracy: 0.9467

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: 4
eval_batch_size: 4
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: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.6203	0.0613	20	0.4148	0.8267
0.3246	0.1227	40	0.8805	0.8
0.5453	0.1840	60	0.3735	0.8667
0.4513	0.2454	80	0.4391	0.8867
0.7729	0.3067	100	0.4407	0.82
0.5867	0.3681	120	0.4013	0.8467
0.4073	0.4294	140	0.5397	0.86
0.1883	0.4908	160	0.7620	0.8667
0.4166	0.5521	180	0.6517	0.8933
0.4672	0.6135	200	0.6163	0.88
0.6858	0.6748	220	0.3484	0.8667
0.335	0.7362	240	0.6031	0.8533
0.4525	0.7975	260	0.6941	0.82
0.2385	0.8589	280	0.5618	0.88
0.4256	0.9202	300	0.5899	0.88
0.4934	0.9816	320	0.3289	0.9
0.277	1.0429	340	0.5671	0.88
0.5097	1.1043	360	0.5247	0.88
0.105	1.1656	380	0.4810	0.9
0.3976	1.2270	400	0.4562	0.8933
0.3506	1.2883	420	0.3943	0.8867
0.2057	1.3497	440	0.4944	0.8933
0.2788	1.4110	460	0.4718	0.9
0.4049	1.4724	480	0.5067	0.88
0.415	1.5337	500	0.4395	0.9
0.3565	1.5951	520	0.3682	0.9
0.3111	1.6564	540	0.3298	0.9
0.4191	1.7178	560	0.4493	0.8733
0.2731	1.7791	580	0.3832	0.9067
0.1803	1.8405	600	0.4403	0.8933
0.4462	1.9018	620	0.3844	0.9067
0.0025	1.9632	640	0.4563	0.9067
0.1574	2.0245	660	0.5508	0.8933
0.0927	2.0859	680	0.5529	0.9067
0.184	2.1472	700	0.5161	0.9
0.2446	2.2086	720	0.5064	0.8933
0.2498	2.2699	740	0.4034	0.92
0.2217	2.3313	760	0.5095	0.8733
0.2938	2.3926	780	0.3754	0.9067
0.109	2.4540	800	0.4771	0.8933
0.0282	2.5153	820	0.5535	0.8933
0.2455	2.5767	840	0.4206	0.9067
0.4728	2.6380	860	0.3018	0.9067
0.1145	2.6994	880	0.3053	0.9067
0.1045	2.7607	900	0.3431	0.9067
0.2207	2.8221	920	0.6482	0.86
0.427	2.8834	940	0.4396	0.9133
0.1898	2.9448	960	0.3327	0.92
0.0019	3.0061	980	0.3993	0.92
0.0842	3.0675	1000	0.4166	0.9267
0.1619	3.1288	1020	0.4181	0.9133
0.1849	3.1902	1040	0.4727	0.92
0.1949	3.2515	1060	0.3346	0.8933
0.1796	3.3129	1080	0.3471	0.9267
0.086	3.3742	1100	0.4089	0.8867
0.0187	3.4356	1120	0.3868	0.92
0.0768	3.4969	1140	0.4095	0.9267
0.0008	3.5583	1160	0.3780	0.9067
0.183	3.6196	1180	0.3827	0.9
0.204	3.6810	1200	0.5133	0.9
0.0758	3.7423	1220	0.4280	0.9133
0.0237	3.8037	1240	0.3942	0.92
0.2143	3.8650	1260	0.3680	0.9067
0.0106	3.9264	1280	0.5633	0.8867
0.2221	3.9877	1300	0.3815	0.92
0.0212	4.0491	1320	0.4599	0.9267
0.1678	4.1104	1340	0.3458	0.92
0.1153	4.1718	1360	0.3261	0.92
0.0006	4.2331	1380	0.3404	0.9133
0.0193	4.2945	1400	0.3602	0.92
0.0994	4.3558	1420	0.3303	0.94
0.0032	4.4172	1440	0.2885	0.94
0.0008	4.4785	1460	0.3112	0.92
0.0823	4.5399	1480	0.3145	0.9267
0.0086	4.6012	1500	0.2954	0.94
0.0009	4.6626	1520	0.3082	0.94
0.1619	4.7239	1540	0.2928	0.94
0.0004	4.7853	1560	0.2909	0.9333
0.0006	4.8466	1580	0.2879	0.9467
0.0005	4.9080	1600	0.2894	0.9467
0.0559	4.9693	1620	0.2903	0.9467

Framework versions

Transformers 4.47.1
Pytorch 2.5.1+cu121
Datasets 3.2.0
Tokenizers 0.21.0