关于“Build a DNN using the Keras Functional API”的评价

Hiroki I. · 评论over 4 years之前

GUTTENBERG F. · 评论over 4 years之前

Iyasu E. · 评论over 4 years之前

Good lab but the results are strange given the nature of TF random seed. First time I got a loss of around, and it wouldn't go down with more epochs. Second time I got a loss of 10 on the second epoch and then stayed there, but the fare prediction was negative. Third time was just like the first one.

Jesús R. · 评论over 4 years之前

Adrian B. · 评论over 4 years之前

Rochak L. · 评论over 4 years之前

Aldrich L. · 评论over 4 years之前

Seemed weird tho - I put in some wild variations of lat/lon and all fares were 18.something; didn't seem right.

Maurice E. · 评论over 4 years之前

Larry L. · 评论over 4 years之前

Jeff P. · 评论over 4 years之前

Daniel W. · 评论over 4 years之前

to big of a disconnect between training material and exercise. not once is keras.layers.input mentioned but we are expected to use it. A hint like go read up on this tf example page would already be enough of a hint what is expected in this particular exersize...

Marco v. · 评论over 4 years之前

Mohnish S. · 评论over 4 years之前

ABDELMONEM K. · 评论over 4 years之前

Ganesh P. · 评论over 4 years之前

Kornesh K. · 评论over 4 years之前

NEY C. · 评论over 4 years之前

marwan s. · 评论over 4 years之前

Nasiba Z. · 评论over 4 years之前

Brandon C. · 评论over 4 years之前

Basab K. · 评论over 4 years之前

Corina C. · 评论over 4 years之前

Dmitry O. · 评论over 4 years之前

Hasan M. · 评论over 4 years之前

Leonard P. · 评论over 4 years之前