MAINT: remove test dependance on linting

See thread in #208 (review)

Since we use GHA as a FOSS project, we really don't pay for usage. This will increase usage because tests will run even if linting fails, but will help PR feedback TAT because failed linting won't preclude getting test data back.

.github/workflows/tests.yaml

@@ -22,7 +22,6 @@ jobs:
       - uses: pre-commit/[email protected]
 
   TestStable:
-    needs: Linting
     name: Ubuntu / Python ${{ matrix.python-version }} / TensorFlow Stable / Scikit-Learn Stable
     runs-on: ubuntu-latest
     strategy:
@@ -55,7 +54,6 @@ jobs:
       - uses: codecov/codecov-action@v1
 
   TestDev:
-    needs: Linting
     name: Ubuntu / Python ${{ matrix.python-version }} / TensorFlow Nightly / Scikit-Learn Nightly
     runs-on: ubuntu-latest
     strategy:
@@ -98,7 +96,6 @@ jobs:
       - uses: codecov/codecov-action@v1
 
   TestOldest:
-    needs: Linting
     name: Ubuntu / Python ${{ matrix.python-version }} / TF ${{ matrix.tf-version }} / Scikit-Learn ${{ matrix.sklearn-version }}
     runs-on: ubuntu-latest
     strategy:
@@ -135,7 +132,6 @@ jobs:
       - uses: codecov/codecov-action@v1
 
   TestOSs:
-    needs: Linting
     name: ${{ matrix.os }} / Python ${{ matrix.python-version }} / TF Stable / Scikit-Learn Stable
     runs-on: ${{ matrix.os }}-latest
     strategy: