Kyle Daruwalla
Kyle Daruwalla
This updates the GPU CI to test on LTS + latest.
This is a ~~non-breaking~~ (see below) step towards separating all the keywords out of `fmap`. It was a side-effect of working on #41, so I separated the changes out to...
Increase the LearnBase compat to 0.4 (last version before some more breaking changes which we will want to hold off on).
Great work! I've been working on a similar idea, and I was wondering if you would consider making this work a PR to MLDataPattern.jl? The key feature here is the...
The new save/load docs promote JLD2.jl which does not support saving/loading anonymous functions reliably. This most commonly occurs for activation functions. The solution is to use `Flux.state` + `Flux.loadmodel!` and...
**This issue is a WIP. I am starting from the template discussed on call, we need to break this into smaller steps.** # Flux documentation restructure It should surprise no...
Right now, the normalization scale and bias are required to be the same type. This is an unnecessary restriction that isn't there for the weight and biases of other layers....
When multiplying a vector by a matrix, the `rrule` triggers scalar indexing for GPU arrays. ```julia julia> using Zygote, CUDA julia> CUDA.allowscalar(false) julia> x, y = rand(Float32, 6), rand(Float32, 1,...
Functions like `tr(A * B)` will throw scalar indexing issues in the pullback for `*` when `A` and `B` are `CuArray`s. This is because the [pullback for `tr`](https://github.com/JuliaDiff/ChainRules.jl/blob/a9a84ba6cb8aa9ce079af9401600e7c96a8aff3a/src/rulesets/LinearAlgebra/dense.jl#L197) creates a...
The following markdown source: ```md Some text $$ y = ax + b $$ {#eq:my-equation} ``` will not get picked up by the filter. pandoc-crossref does seem to pick this...