[test] Fix flaky test for tensorflow 2.7.0 on GPU

[frankfliu]

Change-Id: If1a2a5cf58214f0cacee362e2806936ca6e271b3

Description

Brief description of what this PR is about

• If this change is a backward incompatible change, why must this change be made?
• Interesting edge cases to note here

[codecov-commenter]

Codecov Report

Merging #1475 (9f47aba) into master (bb5073f) will decrease coverage by 1.39%.
The diff coverage is 48.12%.

@@             Coverage Diff              @@
##             master    #1475      +/-   ##
============================================
- Coverage     72.08%   70.68%   -1.40%     
- Complexity     5126     5300     +174     
============================================
  Files           473      493      +20     
  Lines         21970    23227    +1257     
  Branches       2351     2535     +184     
============================================
+ Hits          15838    16419     +581     
- Misses         4925     5529     +604     
- Partials       1207     1279      +72     

Impacted Files	Coverage Δ
api/src/main/java/ai/djl/modality/cv/Image.java	69.23% <ø> (-4.11%)	⬇️
...i/djl/modality/cv/translator/BigGANTranslator.java	21.42% <ø> (-5.24%)	⬇️
...odality/cv/translator/BigGANTranslatorFactory.java	33.33% <0.00%> (+8.33%)	⬆️
...nslator/InstanceSegmentationTranslatorFactory.java	14.28% <0.00%> (-3.90%)	⬇️
.../modality/cv/translator/YoloTranslatorFactory.java	8.33% <0.00%> (-1.67%)	⬇️
...i/djl/modality/cv/translator/YoloV5Translator.java	5.69% <0.00%> (ø)
...odality/cv/translator/YoloV5TranslatorFactory.java	8.33% <0.00%> (-1.67%)	⬇️
...pi/src/main/java/ai/djl/ndarray/BytesSupplier.java	54.54% <0.00%> (-12.13%)	⬇️
...i/src/main/java/ai/djl/ndarray/NDArrayAdapter.java	15.23% <0.00%> (-0.83%)	⬇️
...l/training/loss/SigmoidBinaryCrossEntropyLoss.java	64.00% <0.00%> (ø)
... and 160 more

---

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 750c153...9f47aba. Read the comment docs.