size_categories:
- 10M<n<100M
Welcome!
This is the dataset for OptoGPT: A Foundation Model for Multilayer Thin Film Inverse Design (https://www.oejournal.org/article/doi/10.29026/oea.2024.240062)
Code for OptoGPT: https://github.com/taigaoma1997/optogpt
Model for OptoGPT: https://huggingface.co/mataigao/optogpt
File description:
/train
/data_train_0.csv
/data_train_1.csv
/data_train_2.csv
/data_train_3.csv
/data_train_4.csv
/data_train_5.csv
/data_train_6.csv
/data_train_7.csv
/data_train_8.csv
/data_train_9.csv
/test /data_test.csv
Each .csv file contains 1M thin film structures as well as their simulated reflection and transmission spectrum. The spectrum is ranging from 400nm to 1100nm with 10nm gap.
Column description:
structure: List of the thin film structure sequence. The order of the sequence is from top to bottom, namely, the first layer is touching the air, and the bottom layer is touching the glass substrate.
num_layers: Total number of layers of this structure.
R_400nm: Reflection at 400 nm.
T_400nm: Transmission at 400 nm.
R_...
T_...