search

huawei-noah / vega

AutoML tools chain
http://www.noahlab.com.hk/opensource/vega/
Other
842 stars 175 forks source link

using vega for vector samples? #174

Open exx8 opened 2 years ago

exx8 commented 2 years ago

I see that vega is mostly used for matrices (actually for images). Does vega support 1 dimension tensor (vector). I do understand, 2 dimensional filter doesn't make any sense (probably?), but does vega offer anything with regards to dense architecture for example?

zhangjiajin commented 2 years ago

The networks in Vega dosn't support 1 dimension tensor. If you have some optimization requirements, you can try Vega's hpo(hyper-parameters' optimization) algorithms. Vega is responsible for hyper-parameters' optimization. You need to give the training script file that support 1 or 2 dimension tensor. The example is following: https://github.com/huawei-noah/vega/tree/master/examples/features/script_runner