iNNterpol: High-precision interpolation of stellar atmospheres with a deep neural network using a 1D convolutional auto encoder for feature extraction

Open Access

Table 3

Range of parameters tested for the fully connected NN and optimal values for MARCS and PHOENIX data.

n_l Layers	n_n Neurons/layer	Activation	Batch size

16–28	64–128	lkrelu, elu	128, 64, 32
Optimal values
(MARCS CAE71)
22	71	elu	64
(PHOENIX CAE96)
16	96	elu	64

Notes. For both MARCS and PHOENIX, the best configurations have the resulting dimensionality reduction (n_o) equal to the neurons per layer (n_n) which are n_o = n_l = 71 and 96, respectively.

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.