Fix automation errors

Author: mikeheddes

.github/workflows/publish.yml

@@ -50,6 +50,8 @@ jobs:
       env:
         ANACONDA_API_TOKEN: ${{ secrets.ANACONDA_TOKEN }}
       run: |
+        sleep 5m  # wait for all the PyPi servers to be updated with the latest version of torch-hd
+
         mkdir ./conda-build  # create the artifact dir
 
         # Get the version of the library

torchhd/tests/structures/test_sequence.py

@@ -120,12 +120,11 @@ def test_replace(self):
         assert functional.cosine_similarity(S[2], hv)[2] < 0.35
         assert functional.cosine_similarity(S[2], hv)[6] > 0.35
 
-        hv1 = functional.random_hv(10, 10000)
-        S2 = structures.BundleSequence.from_tensor(hv1)
-        assert functional.cosine_similarity(S2[2], hv1)[2] > 0.3
-        S2.replace(2, hv1[2], hv1[6])
-        assert functional.cosine_similarity(S2[2], hv1)[2] < 0.3
-        assert functional.cosine_similarity(S2[2], hv1)[6] > 0.3
+        S2 = structures.BundleSequence.from_tensor(hv[:7])
+        assert functional.cosine_similarity(S2[2], hv)[2] > 0.35
+        S2.replace(2, hv[2], hv[6])
+        assert functional.cosine_similarity(S2[2], hv)[2] < 0.35
+        assert functional.cosine_similarity(S2[2], hv)[6] > 0.35
 
     def test_concat(self):
         generator = torch.Generator()