added github workflows, minor fixes on init/deps

.github/tests.json

@@ -0,0 +1,16 @@
+[
+  {
+    "hardwareConfig": {
+      "endpointConfig": {
+        "gpuIds": "AMPERE_16",
+        "name": "16GB GPU"
+      }
+    },
+    "input": {
+      "audio": "https://github.com/runpod-workers/sample-inputs/raw/main/audio/gettysburg.wav",
+      "model_name": "tiny",
+      "language": null,
+      "batch_size": 16
+    }
+  }
+]

.github/workflows/CD-docker_dev.yml

@@ -0,0 +1,35 @@
+name: Docker Image
+
+on:
+  push:
+    branches:
+      - "main"
+
+jobs:
+  docker:
+    runs-on: DO
+    steps:
+      - name: Set up QEMU
+        uses: docker/setup-qemu-action@v2
+
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v2
+
+      - name: Login to Docker Hub
+        uses: docker/login-action@v2
+        with:
+          username: ${{ secrets.DOCKERHUB_USERNAME }}
+          password: ${{ secrets.DOCKERHUB_TOKEN }}
+
+      - name: Build and push
+        uses: docker/build-push-action@v4
+        with:
+          push: true
+          tags: ${{ vars.DOCKERHUB_REPO }}/${{ vars.DOCKERHUB_IMG }}:dev
+      - uses: actions/checkout@v3
+      - name: Run Tests
+        uses: direlines/[email protected]
+        with:
+          image-tag: ${{ vars.DOCKERHUB_REPO }}/${{ vars.DOCKERHUB_IMG }}:dev
+          runpod-api-key: ${{ secrets.RUNPOD_API_KEY }}
+          request-timeout: 600

.github/workflows/CD-docker_release.yml

@@ -0,0 +1,27 @@
+name: Docker Image
+
+on:
+  release:
+    types: [published]
+
+jobs:
+  docker:
+    runs-on: DO
+    steps:
+      - name: Set up QEMU
+        uses: docker/setup-qemu-action@v2
+
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v2
+
+      - name: Login to Docker Hub
+        uses: docker/login-action@v2
+        with:
+          username: ${{ secrets.DOCKERHUB_USERNAME }}
+          password: ${{ secrets.DOCKERHUB_TOKEN }}
+
+      - name: Build and push
+        uses: docker/build-push-action@v4
+        with:
+          push: true
+          tags: ${{ secrets.DOCKERHUB_REPO }}/${{ secrets.DOCKERHUB_IMG }}:${{ github.event.release.tag_name }}

.github/workflows/CI-runpod_dep.yml

@@ -0,0 +1,52 @@
+name: CI | Update runpod package version
+
+on:
+  repository_dispatch:
+    types: [python-package-release]
+
+  push:
+    branches:
+      - "main"
+      - "master"
+
+  workflow_dispatch:
+
+jobs:
+  check_dep:
+    runs-on: ubuntu-latest
+    name: Check python requirements file and update
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v2
+
+      - name: Check for new package version and update
+        run: |
+          # Get current version
+          current_version=$(grep -oP 'runpod==\K[^"]+' ./builder/requirements.txt)
+
+          # Get new version
+          new_version=$(curl -s https://pypi.org/pypi/runpod/json | jq -r .info.version)
+          echo "NEW_VERSION_ENV=$new_version" >> $GITHUB_ENV
+
+          if [ -z "$new_version" ]; then
+              echo "Failed to fetch the new version."
+              exit 1
+          fi
+
+          # Check if the version is already up-to-date
+          if [ "$current_version" = "$new_version" ]; then
+              echo "The package version is already up-to-date."
+              exit 0
+          fi
+
+          # Update requirements.txt
+          sed -i "s/runpod==.*/runpod==$new_version/" ./builder/requirements.txt
+
+      - name: Create Pull Request
+        uses: peter-evans/create-pull-request@v3
+        with:
+          token: ${{ secrets.GITHUB_TOKEN }}
+          commit-message: Update package version
+          title: Update runpod package version
+          body: The package version has been updated to ${{ env.NEW_VERSION_ENV }}
+          branch: runpod-package-update

.github/workflows/CI-test_handler.yml

@@ -0,0 +1,85 @@
+name: CI | Test Worker
+
+on:
+  push:
+    branches:
+      - main
+
+  pull_request:
+    branches:
+      - main
+
+  workflow_dispatch:
+
+jobs:
+  initialize_worker:
+    runs-on: ubuntu-latest
+    outputs:
+      id: ${{ steps.extract_id.outputs.runpod_job_id }}
+
+    steps:
+      - name: Deploy Worker
+        id: deploy
+        uses: fjogeleit/http-request-action@v1
+        with:
+          url: "https://api.runpod.ai/v2/${{ secrets.RUNPOD_ENDPOINT }}/run"
+          method: "POST"
+          customHeaders: '{"Content-Type": "application/json"}'
+          bearerToken: ${{ secrets.RUNPOD_API_KEY }}
+          data: '{"input":{"github_pat": "${{ secrets.GH_PAT }}", "github_org":"${{ secrets.GH_ORG }}"}}'
+
+      - name: Extract Job ID
+        id: extract_id
+        run: |
+          ID=$(echo '${{ steps.deploy.outputs.response }}' | jq -r '.id')
+          echo "::set-output name=runpod_job_id::$ID"
+
+  run_tests:
+    needs: initialize_worker
+    runs-on: runpod
+
+    steps:
+      - uses: actions/checkout@v3
+
+      - name: Set up environment
+        run: |
+          rm -f /etc/apt/sources.list.d/*.list
+          apt-get update -y
+          apt-get upgrade -y
+          apt-get install --yes --no-install-recommends sudo ca-certificates git wget curl bash libgl1 libx11-6 software-properties-common ffmpeg build-essential -y
+          apt-get autoremove -y
+          apt-get clean -y
+          rm -rf /var/lib/apt/lists/*
+
+      - name: Set up Python 3.10 & install dependencies
+        uses: actions/setup-python@v4
+        with:
+          python-version: "3.10.12"
+
+      - name: Install Dependencies
+        run: |
+
+          python -m pip install --upgrade pip
+          pip install -r builder/requirements.txt
+
+      - name: Fetch and run models
+        run: |
+          python builder/fetch_models.py
+
+      - name: Execute Tests
+        run: |
+          python src/rp_handler.py --test_input='{"input": {"audio": "https://github.com/runpod-workers/sample-inputs/raw/main/audio/gettysburg.wav"}}'
+
+  terminate_worker:
+    if: ${{ always() && !success() }}
+    needs: initialize_worker
+    runs-on: ubuntu-latest
+
+    steps:
+      - name: Shutdown Worker
+        uses: fjogeleit/http-request-action@v1
+        with:
+          url: "https://api.runpod.ai/v2/${{ secrets.RUNPOD_ENDPOINT }}/cancel/${{ needs.initialize_worker.outputs.id }}"
+          method: "POST"
+          customHeaders: '{"Content-Type": "application/json"}'
+          bearerToken: ${{ secrets.RUNPOD_API_KEY }}

builder/fetch_models.py

@@ -1,7 +1,7 @@
 from concurrent.futures import ThreadPoolExecutor
-import whisperx
+from faster_whisper import WhisperModel
 
-model_names = ["tiny"] #  "base", "small", "medium", "large-v1", "large-v2", "large-v3"
+model_names = ["tiny", "large-v2"]   #  ["tiny", "base", "small", "medium", "large-v1", "large-v2", "large-v3"]
 
 
 def load_model(selected_model):
@@ -11,7 +11,7 @@ def load_model(selected_model):
     for _attempt in range(5):
         while True:
             try:
-                loaded_model = whisperx.load_model(
+                loaded_model = WhisperModel(
                     selected_model, device="cpu", compute_type="int8")
             except (AttributeError, OSError):
                 continue

builder/requirements.txt

@@ -1,4 +1,9 @@
 runpod==1.6.2
-torch==2.0.0
-torchaudio==2.0.1 --index-url https://download.pytorch.org/whl/cu118
-faster-whisper==3.1.1
+torch>=2
+torchaudio>=2
+faster-whisper==1.0.0
+transformers
+pandas
+setuptools>=65
+nltk
+whisperx

src/predict.py

@@ -23,6 +23,7 @@ class Predictor:
     """
 
     models: Dict[str, FasterWhisperPipeline] = {}
+    model_dir: str = "/tmp"
 
     @property
     def device(self):
@@ -40,7 +41,14 @@ def load_model(self, model_name) -> Tuple[str, FasterWhisperPipeline]:
         loaded_model = load_model(
             model_name,
             device=self.device,
-            compute_type=self.compute_type)
+            compute_type=self.compute_type,
+            download_root=self.model_dir,
+            asr_options = {
+                "max_new_tokens": None,
+                "clip_timestamps": None,
+                "hallucination_silence_threshold": None,
+            }
+        )
 
         return model_name, loaded_model
 
@@ -49,7 +57,7 @@ def setup(self):
         Load the model into memory to make running multiple predictions efficient
         """
 
-        model_names = ["tiny", "base", "small", "medium", "large-v1", "large-v2", "large-v3"]
+        model_names = ["tiny", "large-v2"]   #  ["tiny", "base", "small", "medium", "large-v1", "large-v2", "large-v3"]
         with ThreadPoolExecutor() as executor:
             for model_name, model in executor.map(self.load_model, model_names):
                 if model_name is not None: