fix: resuming from a gpu model on cpu or vice versa

[SamuelLarkin]

For Marc to test.
fixes: #271

[github-actions]

CLI load time: 0:00.56
Pull Request HEAD: 06a6b29e4dbb19c011271b80d594aa7bdaf004de
Imports that take more than 0.1 s:
import time: self [us] | cumulative | imported package
import time:       890 |     115623 |       typer.rich_utils
import time:      1407 |     117796 |     typer.core
import time:      1142 |     141760 |   typer.main
import time:       342 |     162059 | typer
import time:       434 |     109720 |       everyvoice.model.feature_prediction.FastSpeech2_lightning.fs2.cli.synthesize
import time:       343 |     129353 |     everyvoice.model.feature_prediction.FastSpeech2_lightning.fs2.cli.cli
import time:       249 |     130143 |   everyvoice.model.feature_prediction.FastSpeech2_lightning.fs2.cli
import time:        25 |     130167 | everyvoice.model.feature_prediction.FastSpeech2_lightning.fs2.cli.preprocess

[codecov]

Codecov Report

Attention: 1 lines in your changes are missing coverage. Please review.

Comparison is base (2c0c8b0) 70.63% compared to head (06a6b29) 70.63%.

Files	Patch %	Lines
everyvoice/base_cli/helpers.py	0.00%	1 Missing ⚠️

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #272   +/-   ##
=======================================
  Coverage   70.63%   70.63%           
=======================================
  Files          39       39           
  Lines        2476     2476           
  Branches      385      385           
=======================================
  Hits         1749     1749           
  Misses        658      658           
  Partials       69       69           

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

[marctessier]

All look good Sam!

Ran various test with success :
GPU->CPU->GPU
CPU->GPU->CPU
GPU->GPU
CPU->CPU

[SamuelLarkin]

Thanks @marctessier for your help.