README.md

metadata

license: apache-2.0
tags:
  - generated_from_trainer
metrics:
  - accuracy
model-index:
  - name: TSE_ALBERT_5E
    results: []

TSE_ALBERT_5E

This model is a fine-tuned version of albert-base-v2 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.3667
• Accuracy: 0.9333

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: 1e-05
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.5712	0.06	50	0.4047	0.82
0.3198	0.12	100	0.2883	0.9
0.3254	0.17	150	0.4352	0.84
0.2898	0.23	200	0.2892	0.9133
0.2826	0.29	250	0.3565	0.8867
0.2696	0.35	300	0.2263	0.9333
0.274	0.4	350	0.2068	0.94
0.2393	0.46	400	0.2270	0.9333
0.2067	0.52	450	0.2118	0.9333
0.2332	0.58	500	0.4454	0.88
0.3099	0.63	550	0.2777	0.9067
0.2687	0.69	600	0.2077	0.9333
0.2053	0.75	650	0.1923	0.9533
0.2359	0.81	700	0.3891	0.9067
0.2492	0.87	750	0.2765	0.9333
0.2589	0.92	800	0.1879	0.9467
0.2161	0.98	850	0.2733	0.9267
0.1752	1.04	900	0.3108	0.92
0.2213	1.1	950	0.3318	0.92
0.1665	1.15	1000	0.4124	0.8933
0.1832	1.21	1050	0.3448	0.92
0.1671	1.27	1100	0.3343	0.9067
0.184	1.33	1150	0.3929	0.9067
0.2788	1.38	1200	0.3888	0.8933
0.1768	1.44	1250	0.4028	0.9
0.2368	1.5	1300	0.3154	0.9133
0.2055	1.56	1350	0.2603	0.9267
0.1693	1.61	1400	0.2994	0.9267
0.1447	1.67	1450	0.3247	0.9267
0.226	1.73	1500	0.3410	0.9267
0.1744	1.79	1550	0.3105	0.9267
0.1943	1.85	1600	0.2760	0.94
0.2093	1.9	1650	0.2087	0.9467
0.2027	1.96	1700	0.2773	0.9333
0.1806	2.02	1750	0.3386	0.9267
0.1161	2.08	1800	0.4301	0.9067
0.0916	2.13	1850	0.3693	0.92
0.1586	2.19	1900	0.2929	0.94
0.1336	2.25	1950	0.4015	0.9133
0.1746	2.31	2000	0.3027	0.92
0.1353	2.36	2050	0.3224	0.9267
0.116	2.42	2100	0.3609	0.9267
0.1807	2.48	2150	0.3044	0.9267
0.1016	2.54	2200	0.3706	0.9133
0.0634	2.6	2250	0.3391	0.92
0.167	2.65	2300	0.3463	0.92
0.1718	2.71	2350	0.3254	0.92
0.1269	2.77	2400	0.2640	0.9333
0.1848	2.83	2450	0.2660	0.9267
0.116	2.88	2500	0.2532	0.94
0.1804	2.94	2550	0.3538	0.92
0.1315	3.0	2600	0.4146	0.9067
0.1024	3.06	2650	0.2899	0.9333
0.0904	3.11	2700	0.3191	0.9333
0.0596	3.17	2750	0.3569	0.9333
0.1144	3.23	2800	0.3373	0.9267
0.0782	3.29	2850	0.3447	0.9267
0.064	3.34	2900	0.2932	0.94
0.118	3.4	2950	0.3099	0.94
0.1286	3.46	3000	0.3404	0.9267
0.0963	3.52	3050	0.4026	0.9067
0.1158	3.58	3100	0.3320	0.9267
0.0967	3.63	3150	0.2984	0.94
0.1122	3.69	3200	0.3149	0.9333
0.134	3.75	3250	0.3804	0.9133
0.0953	3.81	3300	0.3670	0.92
0.0776	3.86	3350	0.4140	0.92
0.0813	3.92	3400	0.3654	0.9333
0.0406	3.98	3450	0.4364	0.92
0.0538	4.04	3500	0.3553	0.94
0.0734	4.09	3550	0.3814	0.9267
0.0396	4.15	3600	0.3978	0.9267
0.0427	4.21	3650	0.4333	0.92
0.1472	4.27	3700	0.3816	0.92
0.0587	4.33	3750	0.3624	0.92
0.0549	4.38	3800	0.3461	0.9333
0.0606	4.44	3850	0.3562	0.94
0.0483	4.5	3900	0.3655	0.9333
0.0351	4.56	3950	0.3613	0.9333
0.0763	4.61	4000	0.3641	0.94
0.0835	4.67	4050	0.3669	0.9333
0.0542	4.73	4100	0.3569	0.9333
0.0804	4.79	4150	0.3575	0.9333
0.0336	4.84	4200	0.3655	0.9333
0.0631	4.9	4250	0.3646	0.9333
0.0183	4.96	4300	0.3667	0.9333

Framework versions

• Transformers 4.24.0
• Pytorch 1.13.0
• Datasets 2.3.2
• Tokenizers 0.13.1