Skip to content
#

differentiable-programming

Here are 25 public repositories matching this topic...

fsaad
fsaad commented Mar 19, 2019
julia> @gen (static) function f()
           return @trace(normal(0,1), :foo)
           end
ERROR: LoadError: Syntax error when parsing static DSL function at return #= REPL[3]:2 =# @trace(normal(0, 1), :foo). 
Stacktrace:
 [1] error(::String) at ./error.jl:33
 [2] static_dsl_syntax_error(::Expr, ::String) at /scratch/fsaad/.julia/packages/Gen/n7UPH/src/dsl/static.jl:6 (repeats 2 
rxwei
rxwei commented Aug 24, 2019

For many raw APIs, we already have corresponding idiomatic APIs that implement all of the raw API's functionality, e.g. Tensor.init(_:) for Raw.cast(_:), Tensor.reshaped(to:) for Raw.reshape(_:shape:), etc. These raw APIs are not recommend also because they are not differentiable. However, raw APIs are more discoverable than the idiomatic APIs to users who are familiar with Python TensorFl

Loading…
You can’t perform that action at this time.