Skip tensorflow test test_forward_ssd (apache#10231)

Since this test is launched in a thread, errors aren't propagated up to the main thread and thusly pytest doesn't detect / report them, so this test will always succeed. Given that this single test can take [upwards of 30 minutes](https://ci.tlcpack.ai/job/tvm/job/main/2499/testReport/cython.tests.python.frontend.tensorflow/test_forward/), this PR disables it until someone lands a proper fix (given that this test takes so long some consideration should be given to reducing its runtime before enabling it again).

Co-authored-by: driazati <[email protected]>

tests/python/frontend/tensorflow/test_forward.py

@@ -3934,6 +3934,9 @@ def _test_ssd_impl():
                     tvm.testing.assert_allclose(tvm_output[i], tf_output[i], rtol=1e-3, atol=1e-3)
 
 
+@pytest.mark.skip(
+    reason="Use of threading module here hides errors, see https://github.com/apache/tvm/pull/10231"
+)
 def test_forward_ssd():
     run_thread = threading.Thread(target=_test_ssd_impl, args=())
     old_stack_size = threading.stack_size(100 * 1024 * 1024)