Pin MLJFlux at version 0.4

[trontrytel]

The NN aerosol activation emulator training was broken in the longruns because the API of MLJFlux package have changed:

ArgumentError: Flux.jl optimiser detected. Only optimisers from Optimisers.jl are supported.

I tried updating to using Adam optimiser from the Optimisers package as they want us to do. However I then get an error:

ERROR: Optimisers.jl cannot be used with Zygote.jl's implicit gradients, Params & Grads

At least for now I'm just pinning the MLJFlux package at the older version.

[codecov[bot]]

Codecov Report

All modified and coverable lines are covered by tests :white_check_mark:

Project coverage is 97.01%. Comparing base (f73555a) to head (1973e64). Report is 1 commits behind head on main.

Additional details and impacted files

```diff @@ Coverage Diff @@ ## main #417 +/- ## ======================================= Coverage 97.01% 97.01% ======================================= Files 37 37 Lines 1440 1440 ======================================= Hits 1397 1397 Misses 43 43 ``` | [Components](https://app.codecov.io/gh/CliMA/CloudMicrophysics.jl/pull/417/components?src=pr&el=components&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=CliMA) | Coverage Δ | | |---|---|---| | [src](https://app.codecov.io/gh/CliMA/CloudMicrophysics.jl/pull/417/components?src=pr&el=component&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=CliMA) | `98.95% <ø> (ø)` | | | [ext](https://app.codecov.io/gh/CliMA/CloudMicrophysics.jl/pull/417/components?src=pr&el=component&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=CliMA) | `69.79% <ø> (ø)` | |