Improved-MARBERT-twitter-sentiment-nodroput-Twitter

This model is a fine-tuned version of UBC-NLP/MARBERTv2 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.9314
• Accuracy: 0.85

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: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 15

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.5838	0.55	50	0.6058	0.71
0.3547	1.1	100	0.3875	0.84
0.2735	1.65	150	0.3429	0.87
0.1925	2.2	200	0.3658	0.87
0.1676	2.75	250	0.5335	0.83
0.1366	3.3	300	0.5710	0.81
0.1202	3.85	350	0.5037	0.84
0.0843	4.4	400	0.5798	0.83
0.086	4.95	450	0.9457	0.79
0.0589	5.49	500	1.1547	0.76
0.0599	6.04	550	0.8437	0.83
0.036	6.59	600	0.9878	0.78
0.0557	7.14	650	0.7223	0.86
0.041	7.69	700	0.7275	0.85
0.0256	8.24	750	0.7327	0.85
0.0273	8.79	800	0.7270	0.84
0.0165	9.34	850	0.8266	0.85
0.0154	9.89	900	0.7583	0.87
0.0075	10.44	950	0.8894	0.85
0.0073	10.99	1000	0.8858	0.85
0.0126	11.54	1050	0.9245	0.84
0.0047	12.09	1100	0.9335	0.84
0.0073	12.64	1150	0.9405	0.85
0.0112	13.19	1200	0.9801	0.84
0.0068	13.74	1250	0.9314	0.85

Framework versions

• Transformers 4.34.1
• Pytorch 2.1.0+cu118
• Datasets 2.14.7
• Tokenizers 0.14.1