rubert-tiny2_finetuned_emotion_experiment_modified_CE_LOSS_resampling

This model is a fine-tuned version of cointegrated/rubert-tiny2 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.4520
• Accuracy: 0.8621
• F1: 0.8616

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: 64
• eval_batch_size: 64
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 40

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
1.1134	1.0	53	0.9494	0.7716	0.7555
0.8421	2.0	106	0.7092	0.8204	0.8172
0.6488	3.0	159	0.6000	0.8319	0.8313
0.5392	4.0	212	0.5368	0.8376	0.8392
0.4616	5.0	265	0.4951	0.8549	0.8544
0.4138	6.0	318	0.4743	0.8621	0.8615
0.3694	7.0	371	0.4607	0.8563	0.8581
0.3375	8.0	424	0.4469	0.8693	0.8697
0.3049	9.0	477	0.4412	0.8649	0.8670
0.2804	10.0	530	0.4469	0.8635	0.8637
0.2787	11.0	583	0.4471	0.8693	0.8683
0.2284	12.0	636	0.4474	0.8693	0.8694
0.2188	13.0	689	0.4530	0.8649	0.8643
0.1998	14.0	742	0.4520	0.8621	0.8616

Framework versions

• Transformers 4.19.2
• Pytorch 1.11.0+cu113
• Datasets 2.2.2
• Tokenizers 0.12.1