README.md

metadata

library_name: transformers
tags:
  - deeplearning
  - indobert
  - SMOTE
license: apache-2.0
language:
  - id
base_model:
  - indobenchmark/indobert-base-p1

Hasil Train Loop

Epoch	Train Loss	Train ACC	Train F1	Train REC	Train PRE	Valid Loss	Valid ACC	Valid F1	Valid REC	Valid PRE	Catatan
1	0.8074	0.6357	0.6347	0.6357	0.6343	0.6218	0.7230	0.6995	0.7050	0.7110	Model terbaik disimpan
2	0.5472	0.7728	0.7732	0.7728	0.7739	0.5335	0.7824	0.7515	0.7531	0.7563	Model terbaik disimpan
3	0.4165	0.8307	0.8313	0.8307	0.8321	0.4123	0.8380	0.8113	0.8127	0.8104	Model terbaik disimpan
4	0.3166	0.8751	0.8755	0.8751	0.8762	0.4554	0.8248	0.7951	0.7973	0.7971	VALID LOSS tidak membaik (1/2)
5	0.2593	0.8970	0.8973	0.8970	0.8979	0.4023	0.8441	0.8230	0.8300	0.8219	Model terbaik disimpan
6	0.2175	0.9160	0.9161	0.9160	0.9163	0.3470	0.8850	0.8633	0.8609	0.8665	Model terbaik disimpan
7	0.1940	0.9268	0.9269	0.9268	0.9271	0.3848	0.8704	0.8480	0.8484	0.8478	VALID LOSS tidak membaik (1/2)
8	0.1616	0.9411	0.9411	0.9411	0.9411	0.4156	0.8596	0.8377	0.8414	0.8354	VALID LOSS tidak membaik (2/2) — Early stopping

Accuracy per Epoch

Loss per Epoch

Classification Report pada Data Testing

Confusion Matrix pada Data Testing

Distribusi Sentimen Hasil Pred Pada Data Testing

WordCloud Hasil Prediksi Pada Data Testing

Frekuensi Kata Hasil Prediksi Pada Data Testing