feat: add fp16 inference support (torch/onnx)

[OrangeSodahub]

• clip_torch
• clip_onnx

[codecov]

Codecov Report

Merging #871 (bd1fe7c) into main (fd16e5a) will decrease coverage by 9.68%.
The diff coverage is 93.47%.

@@            Coverage Diff             @@
##             main     #871      +/-   ##
==========================================
- Coverage   81.04%   71.35%   -9.69%     
==========================================
  Files          22       22              
  Lines        1498     1529      +31     
==========================================
- Hits         1214     1091     -123     
- Misses        284      438     +154     

Flag	Coverage Δ
cas	71.35% <93.47%> (-9.69%)	⬇️

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	75.57% <81.81%> (+0.47%)	⬆️
server/clip_server/executors/clip_torch.py	87.77% <90.00%> (+0.87%)	⬆️
server/clip_server/executors/clip_onnx.py	87.91% <100.00%> (+2.68%)	⬆️
server/clip_server/executors/helper.py	97.18% <100.00%> (+0.12%)	⬆️
server/clip_server/helper.py	46.66% <100.00%> (+3.80%)	⬆️
server/clip_server/model/clip_onnx.py	76.19% <100.00%> (+3.46%)	⬆️
server/clip_server/executors/clip_tensorrt.py	0.00% <0.00%> (-94.60%)	⬇️
server/clip_server/model/clip_trt.py	0.00% <0.00%> (-69.39%)	⬇️
server/clip_server/model/trt_utils.py	0.00% <0.00%> (-56.05%)	⬇️
... and 1 more

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

[jemmyshin]

lgtm