fix: check dtype when loading models

[ZiniuYu]

The dtype is automatically set to torch.float16 when running on GPU.
This pr fixes dtype checking logics to unblock torch.float16

[codecov]

Codecov Report

Merging #872 (e38c854) into main (1eebdd7) will increase coverage by 10.34%.
The diff coverage is 100.00%.

@@             Coverage Diff             @@
##             main     #872       +/-   ##
===========================================
+ Coverage   70.69%   81.04%   +10.34%     
===========================================
  Files          22       22               
  Lines        1498     1498               
===========================================
+ Hits         1059     1214      +155     
+ Misses        439      284      -155     

Flag	Coverage Δ
cas	81.04% <100.00%> (+10.34%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
server/clip_server/model/model.py	75.09% <100.00%> (ø)
server/clip_server/model/trt_utils.py	56.04% <0.00%> (+56.04%)	⬆️
server/clip_server/model/clip_trt.py	69.38% <0.00%> (+69.38%)	⬆️
server/clip_server/executors/clip_tensorrt.py	94.59% <0.00%> (+94.59%)	⬆️

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

[jemmyshin]

LGTM