Throw informative errors when backends are not loaded #214
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #214 +/- ##
==========================================
- Coverage 96.40% 95.85% -0.55%
==========================================
Files 65 70 +5
Lines 3558 3791 +233
==========================================
+ Hits 3430 3634 +204
- Misses 128 157 +29 ☔ View full report in Codecov by Sentry. |
|
Can we check that these exceptions are indeed thrown? i.e. by defining a fake backend with no operators |
|
You can use 'check_available' before throwing the exception to confirm the "unloaded" diagnosis. Otherwise the user should probably open an issue, cause I can't think of any other situation where we hit that missing method on purpose |
|
Update with julia> gradient(f, AutoEnzyme(), x)
ERROR: failed to use Enzyme (forward/reverse) backend.
Backend package is not loaded. To fix, run
using Enzyme
Stacktrace:
[1] prepare_pullback_aux(f::Function, backend::AutoEnzyme{Nothing}, x::Vector{Float64}, dy::Float64, ::DifferentiationInterface.PullbackFast)
@ DifferentiationInterface ~/Developer/DifferentiationInterface.jl/DifferentiationInterface/src/first_order/pullback.jl:88
[2] prepare_pullback(f::Function, backend::AutoEnzyme{Nothing}, x::Vector{Float64}, dy::Float64)
@ DifferentiationInterface ~/Developer/DifferentiationInterface.jl/DifferentiationInterface/src/first_order/pullback.jl:68
[3] prepare_gradient(f::typeof(f), backend::AutoEnzyme{Nothing}, x::Vector{Float64})
@ DifferentiationInterface ~/Developer/DifferentiationInterface.jl/DifferentiationInterface/src/first_order/gradient.jl:48
[4] gradient(f::Function, backend::AutoEnzyme{Nothing}, x::Vector{Float64})
@ DifferentiationInterface ~/Developer/DifferentiationInterface.jl/DifferentiationInterface/src/first_order/gradient.jl:73
[5] top-level scope
@ REPL[13]:1
julia> gradient(f, AutoBrokenReverse(), x)
ERROR: failed to use AutoBrokenReverse (reverse) backend.
Please open an issue: https://github.com/gdalle/DifferentiationInterface.jl/issues/new
Stacktrace:
[1] prepare_pullback_aux(f::Function, backend::AutoBrokenReverse, x::Vector{Float64}, dy::Float64, ::DifferentiationInterface.PullbackFast)
@ DifferentiationInterface ~/Developer/DifferentiationInterface.jl/DifferentiationInterface/src/first_order/pullback.jl:88
[2] prepare_pullback(f::Function, backend::AutoBrokenReverse, x::Vector{Float64}, dy::Float64)
@ DifferentiationInterface ~/Developer/DifferentiationInterface.jl/DifferentiationInterface/src/first_order/pullback.jl:68
[3] prepare_gradient(f::typeof(f), backend::AutoBrokenReverse, x::Vector{Float64})
@ DifferentiationInterface ~/Developer/DifferentiationInterface.jl/DifferentiationInterface/src/first_order/gradient.jl:48
[4] gradient(f::Function, backend::AutoBrokenReverse, x::Vector{Float64})
@ DifferentiationInterface ~/Developer/DifferentiationInterface.jl/DifferentiationInterface/src/first_order/gradient.jl:73
[5] top-level scope
@ REPL[14]:1
julia> gradient(f, AutoForwardDiff(), x)
ERROR: failed to use ForwardDiff (forward) backend.
Backend package is not loaded. To fix, run
using ForwardDiff
Stacktrace:
[1] prepare_pushforward_aux(f::Function, backend::AutoForwardDiff{…}, x::Vector{…}, dy::FillArrays.OneElement{…}, ::DifferentiationInterface.PushforwardFast)
@ DifferentiationInterface ~/Developer/DifferentiationInterface.jl/DifferentiationInterface/src/first_order/pushforward.jl:77
[2] prepare_pushforward(f::Function, backend::AutoForwardDiff{…}, x::Vector{…}, dx::FillArrays.OneElement{…})
@ DifferentiationInterface ~/Developer/DifferentiationInterface.jl/DifferentiationInterface/src/first_order/pushforward.jl:56
[3] prepare_pullback_aux(f::Function, backend::AutoForwardDiff{…}, x::Vector{…}, dy::Float64, ::DifferentiationInterface.PullbackSlow)
@ DifferentiationInterface ~/Developer/DifferentiationInterface.jl/DifferentiationInterface/src/first_order/pullback.jl:77
[4] prepare_pullback(f::Function, backend::AutoForwardDiff{nothing, Nothing}, x::Vector{Float64}, dy::Float64)
@ DifferentiationInterface ~/Developer/DifferentiationInterface.jl/DifferentiationInterface/src/first_order/pullback.jl:68
[5] prepare_gradient(f::typeof(f), backend::AutoForwardDiff{nothing, Nothing}, x::Vector{Float64})
@ DifferentiationInterface ~/Developer/DifferentiationInterface.jl/DifferentiationInterface/src/first_order/gradient.jl:48
[6] gradient(f::Function, backend::AutoForwardDiff{nothing, Nothing}, x::Vector{Float64})
@ DifferentiationInterface ~/Developer/DifferentiationInterface.jl/DifferentiationInterface/src/first_order/gradient.jl:73
[7] top-level scope
@ REPL[15]:1
Some type information was truncated. Use `show(err)` to see complete types.
julia> gradient(f, AutoBrokenForward(), x)
ERROR: failed to use AutoBrokenForward (forward) backend.
Please open an issue: https://github.com/gdalle/DifferentiationInterface.jl/issues/new
Stacktrace:
[1] prepare_pushforward_aux(f::Function, backend::AutoBrokenForward, x::Vector{…}, dy::FillArrays.OneElement{…}, ::DifferentiationInterface.PushforwardFast)
@ DifferentiationInterface ~/Developer/DifferentiationInterface.jl/DifferentiationInterface/src/first_order/pushforward.jl:77
[2] prepare_pushforward(f::Function, backend::AutoBrokenForward, x::Vector{Float64}, dx::FillArrays.OneElement{Float64, 1, Tuple{…}, Tuple{…}})
@ DifferentiationInterface ~/Developer/DifferentiationInterface.jl/DifferentiationInterface/src/first_order/pushforward.jl:56
[3] prepare_pullback_aux(f::Function, backend::AutoBrokenForward, x::Vector{Float64}, dy::Float64, ::DifferentiationInterface.PullbackSlow)
@ DifferentiationInterface ~/Developer/DifferentiationInterface.jl/DifferentiationInterface/src/first_order/pullback.jl:77
[4] prepare_pullback(f::Function, backend::AutoBrokenForward, x::Vector{Float64}, dy::Float64)
@ DifferentiationInterface ~/Developer/DifferentiationInterface.jl/DifferentiationInterface/src/first_order/pullback.jl:68
[5] prepare_gradient(f::typeof(f), backend::AutoBrokenForward, x::Vector{Float64})
@ DifferentiationInterface ~/Developer/DifferentiationInterface.jl/DifferentiationInterface/src/first_order/gradient.jl:48
[6] gradient(f::Function, backend::AutoBrokenForward, x::Vector{Float64})
@ DifferentiationInterface ~/Developer/DifferentiationInterface.jl/DifferentiationInterface/src/first_order/gradient.jl:73
[7] top-level scope
@ REPL[16]:1
Some type information was truncated. Use `show(err)` to see complete types. |
gdalle
approved these changes
Apr 25, 2024
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
When forgetting to load a backend package, most operators error at either
prepare_pullback_auxorprepare_pushforward_aux, resulting in an uninformativeMethodError:This PR adds these missing methods, throwing an informative error that suggests loading the respective backend.
New errors
Previous errors