This model is a fine-tuned version of avsolatorio/GIST-large-Embedding-v0 on the None dataset. It achieves the following results on the evaluation set:

Loss: 0.5549
F1: 0.6828
Roc Auc: 0.9255
Accuracy: 0.1053

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: 8
eval_batch_size: 8
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
num_epochs: 70
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	F1	Roc Auc	Accuracy
0.4089	1.0	50	0.3415	0.1336	0.7999	0.0702
0.3111	2.0	100	0.3084	0.2771	0.8501	0.0526
0.2544	3.0	150	0.2851	0.4233	0.8679	0.0526
0.209	4.0	200	0.2893	0.4545	0.8868	0.0526
0.1688	5.0	250	0.2560	0.5307	0.9137	0.1053
0.1335	6.0	300	0.2679	0.4982	0.9001	0.0702
0.1043	7.0	350	0.2689	0.5758	0.9070	0.1053
0.0813	8.0	400	0.2786	0.5994	0.9112	0.1228
0.0686	9.0	450	0.2742	0.6150	0.9119	0.1053
0.0553	10.0	500	0.2751	0.6498	0.9076	0.1404
0.0463	11.0	550	0.2905	0.5894	0.9156	0.1228
0.0401	12.0	600	0.2786	0.6313	0.9189	0.1579
0.0319	13.0	650	0.3090	0.6502	0.9127	0.1053
0.0277	14.0	700	0.2876	0.6024	0.9072	0.0877
0.0248	15.0	750	0.2991	0.6546	0.9275	0.0702
0.02	16.0	800	0.3128	0.6345	0.9217	0.0526
0.0176	17.0	850	0.3139	0.6782	0.9239	0.0877
0.0147	18.0	900	0.3128	0.6739	0.9232	0.1053
0.0128	19.0	950	0.3035	0.6718	0.9217	0.1228
0.0108	20.0	1000	0.3298	0.6531	0.9155	0.1053
0.0098	21.0	1050	0.3470	0.6596	0.9183	0.1053
0.0084	22.0	1100	0.3471	0.6674	0.9170	0.1404
0.0071	23.0	1150	0.3483	0.6756	0.9123	0.1228
0.0064	24.0	1200	0.3600	0.6734	0.9158	0.1053
0.0058	25.0	1250	0.3636	0.6734	0.9172	0.1228
0.0051	26.0	1300	0.3687	0.6826	0.9216	0.1053
0.0043	27.0	1350	0.3859	0.6627	0.9215	0.0877
0.0038	28.0	1400	0.3724	0.6759	0.9299	0.1053
0.0034	29.0	1450	0.4112	0.6869	0.9195	0.1228
0.0029	30.0	1500	0.3952	0.6985	0.9207	0.1404
0.0026	31.0	1550	0.4265	0.6762	0.9204	0.1228
0.0023	32.0	1600	0.4360	0.6861	0.9195	0.1053
0.002	33.0	1650	0.4182	0.6735	0.9271	0.0877
0.0018	34.0	1700	0.4394	0.6678	0.9211	0.0877
0.0016	35.0	1750	0.4406	0.6890	0.9288	0.0877
0.0014	36.0	1800	0.4398	0.6771	0.9240	0.1053
0.0013	37.0	1850	0.4394	0.6849	0.9226	0.0877
0.0012	38.0	1900	0.4642	0.6712	0.9147	0.0702
0.0011	39.0	1950	0.4667	0.6744	0.9223	0.0877
0.001	40.0	2000	0.4570	0.6662	0.9222	0.1053
0.0009	41.0	2050	0.4608	0.6871	0.9257	0.1053
0.0008	42.0	2100	0.4586	0.6771	0.9290	0.1053
0.0007	43.0	2150	0.4737	0.6903	0.9208	0.1228
0.0006	44.0	2200	0.4784	0.6812	0.9251	0.1053
0.0006	45.0	2250	0.4752	0.7063	0.9188	0.1404
0.0006	46.0	2300	0.4852	0.6938	0.9261	0.1053
0.0005	47.0	2350	0.4978	0.6881	0.9276	0.1053
0.0005	48.0	2400	0.5036	0.6664	0.9243	0.0877
0.0005	49.0	2450	0.5029	0.6782	0.9241	0.0877
0.0004	50.0	2500	0.5160	0.6713	0.9268	0.0877
0.0004	51.0	2550	0.5217	0.6789	0.9253	0.1053
0.0004	52.0	2600	0.5203	0.6842	0.9254	0.1228
0.0003	53.0	2650	0.5242	0.6773	0.9197	0.1228
0.0003	54.0	2700	0.5248	0.6887	0.9261	0.1053
0.0003	55.0	2750	0.5309	0.6796	0.9256	0.1053
0.0003	56.0	2800	0.5356	0.6827	0.9251	0.1228
0.0003	57.0	2850	0.5360	0.6693	0.9234	0.1053
0.0003	58.0	2900	0.5420	0.6866	0.9272	0.1053
0.0003	59.0	2950	0.5517	0.6793	0.9245	0.1053
0.0002	60.0	3000	0.5482	0.6855	0.9249	0.0877
0.0002	61.0	3050	0.5514	0.6798	0.9239	0.1053
0.0002	62.0	3100	0.5580	0.6824	0.9240	0.1053
0.0002	63.0	3150	0.5566	0.6821	0.9258	0.1053
0.0002	64.0	3200	0.5582	0.6776	0.9253	0.1053
0.0002	65.0	3250	0.5574	0.6816	0.9264	0.1053
0.0002	66.0	3300	0.5607	0.6767	0.9251	0.1053
0.0002	67.0	3350	0.5523	0.6851	0.9244	0.1053
0.0002	68.0	3400	0.5572	0.6804	0.9255	0.1053
0.0002	69.0	3450	0.5537	0.6828	0.9252	0.1053
0.0002	70.0	3500	0.5549	0.6828	0.9255	0.1053

Framework versions

Transformers 4.38.1
Pytorch 2.1.2
Datasets 2.1.0
Tokenizers 0.15.2

thanavut/output

output

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

thanavut/
output