Add 4GPU unit test (pytorch#82)

For now this literally just runs `NGPU=4 ./run_llama_train.sh` but I
verified at least it catches problems.

As a follow up, we should integrate mgpu test infra from pytorch and set
up actual unit tests to run in this job.

We should probably also keep testing the run_llama_train.sh script, and
add other combinations of 2D parallelism to ensure they all keep
working.

<img width="2120" alt="image"
src="https://github.com/pytorch/torchtrain/assets/4984825/2c235e9a-04ed-4f2d-9915-67de39d78e1c">

.github/workflows/unit_test_4gpu.yaml

@@ -0,0 +1,43 @@
+name: 4 GPU Unit Test
+
+on:
+  push:
+    branches: [ main ]
+  pull_request:
+
+concurrency:
+  group: unit-test${{ github.workflow }}-${{ github.ref == 'refs/heads/main' && github.run_number || github.ref }}
+  cancel-in-progress: true
+
+defaults:
+  run:
+    shell: bash -l -eo pipefail {0}
+
+jobs:
+  unit_tests_4gpu:
+    runs-on: linux.g5.12xlarge.nvidia.gpu
+    strategy:
+      matrix:
+        python-version: ['3.10']
+    steps:
+      - name: Check out repo
+        uses: actions/checkout@v3
+      - name: Setup conda env
+        uses: conda-incubator/setup-miniconda@v2
+        with:
+          auto-update-conda: true
+          miniconda-version: "latest"
+          activate-environment: test
+          python-version: ${{ matrix.python-version }}
+      - name: Update pip
+        run: python -m pip install --upgrade pip
+      - name: Install dependencies
+        run: |
+          pip3 install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/cu121
+          python -m pip install -r requirements.txt
+          python -m pip install -r dev-requirements.txt
+          python -m pip install -e .
+      - name: Run NGPU=4 ./run_llama_train.sh
+        run: NGPU=4 ./run_llama_train.sh
+      - name: Upload Coverage to Codecov
+        uses: codecov/codecov-action@v3

.github/workflows/unit_test.yaml → .github/workflows/unit_test_cpu.yaml

@@ -1,4 +1,4 @@
-name: Unit Test
+name: CPU Unit Test
 
 on:
   push:
@@ -14,7 +14,7 @@ defaults:
     shell: bash -l -eo pipefail {0}
 
 jobs:
-  unit_tests:
+  cpu_unit_tests:
     runs-on: ubuntu-latest
     strategy:
       matrix: