scenario-no_kd_weight_copy-data-indolem_sentiment-model-xlmr_base_trained

This model is a fine-tuned version of haryoaw/scenario-normal-finetune-clf-data-indolem_sentiment-model-xlm-roberta-base on the indolem_sentiment dataset. It achieves the following results on the evaluation set:

• Loss: 0.9589
• Accuracy: 0.8596
• F1: 0.7686

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: 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: 6969

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
No log	0.88	100	0.5053	0.7995	0.7203
No log	1.75	200	0.5105	0.8145	0.7466
No log	2.63	300	0.3984	0.8321	0.7472
No log	3.51	400	0.5181	0.8496	0.7391
0.2291	4.39	500	0.5720	0.8496	0.7600
0.2291	5.26	600	0.8510	0.8446	0.7459
0.2291	6.14	700	0.6628	0.8672	0.7686
0.2291	7.02	800	0.9033	0.8346	0.7295
0.2291	7.89	900	0.8298	0.8571	0.7881
0.0628	8.77	1000	0.7277	0.8496	0.7297
0.0628	9.65	1100	0.8187	0.8471	0.7404
0.0628	10.53	1200	0.7711	0.8571	0.7729
0.0628	11.4	1300	1.1097	0.8496	0.7810
0.0628	12.28	1400	0.8916	0.8596	0.7846
0.0432	13.16	1500	1.0503	0.8571	0.7373
0.0432	14.04	1600	0.8906	0.8672	0.7871
0.0432	14.91	1700	1.0499	0.8471	0.7681
0.0432	15.79	1800	1.0312	0.8571	0.7654
0.0432	16.67	1900	1.0343	0.8471	0.7715
0.0283	17.54	2000	0.8457	0.8596	0.7879
0.0283	18.42	2100	0.9188	0.8697	0.7851
0.0283	19.3	2200	0.9986	0.8596	0.7846
0.0283	20.18	2300	1.0842	0.8471	0.7715
0.0283	21.05	2400	0.9589	0.8596	0.7686

Framework versions

• Transformers 4.33.3
• Pytorch 2.0.1
• Datasets 2.14.5
• Tokenizers 0.13.3