Modelpublic

Etelis/Sentiment140_ALBERT_5E

sourceHugging Faceapache-2.0updated 4y agoView on Hugging Face

0likes21downloads

Model Card

Sentiment140ALBERT5E

This model is a fine-tuned version of albert-base-v2 on the sentiment140 dataset. It achieves the following results on the evaluation set:

—Loss: 0.6103
—Accuracy: 0.8533

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
—trainbatchsize: 16
—evalbatchsize: 8
—seed: 42
—optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
—lrschedulertype: linear
—num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.6713	0.08	50	0.5704	0.7333
0.5742	0.16	100	0.4620	0.8
0.5104	0.24	150	0.5536	0.74
0.5313	0.32	200	0.5198	0.76
0.5023	0.4	250	0.4286	0.8
0.4871	0.48	300	0.4294	0.8267
0.4513	0.56	350	0.4349	0.8133
0.4647	0.64	400	0.4046	0.8333
0.4827	0.72	450	0.4218	0.8333
0.4517	0.8	500	0.4093	0.82
0.4417	0.88	550	0.3999	0.84
0.4701	0.96	600	0.3779	0.8867
0.397	1.04	650	0.3730	0.8667
0.3377	1.12	700	0.3833	0.8333
0.411	1.2	750	0.3704	0.84
0.3796	1.28	800	0.3472	0.86
0.3523	1.36	850	0.3512	0.8733
0.3992	1.44	900	0.3712	0.84
0.3641	1.52	950	0.3718	0.82
0.3973	1.6	1000	0.3508	0.84
0.3576	1.68	1050	0.3600	0.86
0.3701	1.76	1100	0.3287	0.8667
0.3721	1.84	1150	0.3794	0.82
0.3673	1.92	1200	0.3378	0.8733
0.4223	2.0	1250	0.3508	0.86
0.2745	2.08	1300	0.3835	0.86
0.283	2.16	1350	0.3500	0.8533
0.2769	2.24	1400	0.3334	0.8733
0.2491	2.32	1450	0.3519	0.8867
0.3237	2.4	1500	0.3438	0.86
0.2662	2.48	1550	0.3513	0.8667
0.2423	2.56	1600	0.3413	0.8867
0.2655	2.64	1650	0.3126	0.8933
0.2516	2.72	1700	0.3333	0.8733
0.252	2.8	1750	0.3316	0.88
0.2872	2.88	1800	0.3227	0.9
0.306	2.96	1850	0.3383	0.8733
0.248	3.04	1900	0.3474	0.8733
0.1507	3.12	1950	0.4140	0.8667
0.1994	3.2	2000	0.3729	0.8533
0.167	3.28	2050	0.3782	0.8867
0.1872	3.36	2100	0.4352	0.8867
0.1611	3.44	2150	0.4511	0.8667
0.2338	3.52	2200	0.4244	0.8533
0.1538	3.6	2250	0.4226	0.8733
0.1561	3.68	2300	0.4126	0.88
0.2156	3.76	2350	0.4382	0.86
0.1684	3.84	2400	0.4969	0.86
0.1917	3.92	2450	0.4439	0.8667
0.1584	4.0	2500	0.4759	0.86
0.1038	4.08	2550	0.5042	0.8667
0.0983	4.16	2600	0.5527	0.8533
0.1404	4.24	2650	0.5801	0.84
0.0844	4.32	2700	0.5884	0.86
0.1347	4.4	2750	0.5865	0.8467
0.1373	4.48	2800	0.5915	0.8533
0.1506	4.56	2850	0.5976	0.8467
0.1007	4.64	2900	0.6678	0.82
0.1311	4.72	2950	0.6082	0.8533
0.1402	4.8	3000	0.6180	0.8467
0.1363	4.88	3050	0.6107	0.8533
0.0995	4.96	3100	0.6103	0.8533

Framework versions

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