[ci] Run all tests on (self-hosted) GPU (#3020)

* Create self-hosted.yml

* Update self-hosted.yml

* Update self-hosted.yml

* Update self-hosted.yml

* Update self-hosted.yml

* Update self-hosted.yml

* do not run slow tests, for now

* [ci] For comparison with circleci, let's also run CPU-tests

* [ci] reorganize

* clearer filenames

* [ci] Final tweaks before merging

* rm slow tests on circle ci

* Trigger CI

* On GPU this concurrency was way too high

.github/workflows/self-push.yml

@@ -0,0 +1,47 @@
+name: Self-hosted runner (push)
+
+on: 
+  push:
+    branches:
+      - master
+  pull_request:
+
+
+jobs:
+  run_tests_torch_and_tf_gpu:
+    runs-on: self-hosted
+    steps:
+    - uses: actions/checkout@v2
+    - name: Python version
+      run: |
+        which python
+        python --version
+        pip --version
+    - name: Current dir
+      run: pwd
+    - run: nvidia-smi
+    - name: Create new python env (on self-hosted runners we have to handle isolation ourselves)
+      run: |
+        python -m venv .env
+        source .env/bin/activate
+        which python
+        python --version
+        pip --version
+    - name: Install dependencies
+      run: |
+        source .env/bin/activate
+        pip install .[sklearn,tf,torch,testing]
+
+    - name: Are GPUs recognized by our DL frameworks
+      run: |
+        source .env/bin/activate
+        python -c "import torch; print(torch.cuda.is_available())"
+        python -c "import tensorflow as tf; print(tf.test.is_built_with_cuda(), tf.config.list_physical_devices('GPU'))"
+
+    - name: Run all non-slow tests on GPU
+      env:
+        OMP_NUM_THREADS: 1
+        USE_CUDA: yes
+      run: |
+        source .env/bin/activate
+        python -m pytest -n 2 --dist=loadfile -s -v ./tests/