Remove low scoring filter from LogitsDecoder

[zhiqwang]

This is a follow-up PR of #238 to make the exported ONNX model can be worked on TensorRT.

import torch
from torch import Tensor
from yolort.runtime import YOLOTRTModule
from yolort.v5 import attempt_download

model_path = "yolov5n.pt"
checkpoint_path = attempt_download(model_path)

model = YOLOTRTModule(checkpoint_path)
model.eval()
samples = torch.rand(1, 3, 320, 320)
outs = model(samples)

assert isinstance(outs[0], dict)
assert isinstance(outs[0]["boxes"], Tensor)
assert isinstance(outs[0]["scores"], Tensor)

model.to_onnx("yolov5n.onnx")

[codecov]

Codecov Report

Merging #242 (107a0cf) into main (e1f655a) will increase coverage by 0.12%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##             main     #242      +/-   ##
==========================================
+ Coverage   93.49%   93.62%   +0.12%     
==========================================
  Files          10       11       +1     
  Lines         661      674      +13     
==========================================
+ Hits          618      631      +13     
  Misses         43       43              

Flag	Coverage Δ
unittests	93.62% <100.00%> (+0.12%)	⬆️

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

Impacted Files	Coverage Δ
test/test_relaying.py	100.00% <100.00%> (ø)
test/test_runtime.py	100.00% <100.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update e1f655a...107a0cf. Read the comment docs.

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you all sign our Contributor License Agreement before we can accept your contribution.
1 out of 2 committers have signed the CLA.

✅ zhiqwang
❌ pre-commit-ci[bot]
_{You have signed the CLA already but the status is still pending? Let us recheck it.}