Aggregate changes for v1

[avik-pal]

Main Changes

• fixes Roadmap to v1 #735
• fixes Remove Auto-Flattening of Chains #682
• fixes Drop Tracker.jl support for SimpleChains #753
• fixes Missing statistics tracking in normalization layers #780
• migrate to MLDataDevices.jl
• reexport NNlib.jl (partially)
• bias in parameters are now flattened
• use feat: NilArray for fast size propagation #811 to define the fallback outputsize
• removes inputsize since that was never used and is ambiguous for most models
• fixes Move uncommon layers to Boltz.jl #854
• layer map doesn't need the correction handling anymore
• fixes Change defaults for Layers to match Pytorch #808
• finish the migration guide
• outpad similar to the recent addition to Flux
• LPPool layers need tests

Final Leftovers

• Testing failures (minor tolerance issues)

Examples TODOs

• ConvMixer
• DDIM -- gradients are nothing. needs debugging

[codecov]

Codecov Report

Attention: Patch coverage is 98.03371% with 7 lines in your changes missing coverage. Please review.

Project coverage is 93.98%. Comparing base (59f83fc) to head (0bd7099).
Report is 95 commits behind head on main.

Files with missing lines	Patch %	Lines
src/utils.jl	60.00%	6 Missing ⚠️
src/layers/containers.jl	91.66%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #744      +/-   ##
==========================================
+ Coverage   93.71%   93.98%   +0.26%     
==========================================
  Files          59       57       -2     
  Lines        2960     2793     -167     
==========================================
- Hits         2774     2625     -149     
+ Misses        186      168      -18     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.