Module 2 CI example (#7)

kyryl-opens-ml · Jul 3, 2024 · 25cc2f4 · 25cc2f4
1 parent a7b18c2
commit 25cc2f4
Show file tree

Hide file tree

Showing 8 changed files with 10 additions and 27 deletions.
diff --git a/.github/workflows/module-2.yaml b/.github/workflows/module-2.yaml
@@ -2,12 +2,13 @@ name: Module 2
 
 
 on:
-  workflow_dispatch:
-  # push:
+
   pull_request:
     branches:
       - main
-
+  push:
+    branches:
+      - main
 
 jobs:
 

diff --git a/module-2/.dvc/.gitignore b/module-2/.dvc/.gitignore
diff --git a/module-2/.dvc/config b/module-2/.dvc/config
diff --git a/module-2/.dvcignore b/module-2/.dvcignore
diff --git a/module-2/README.md b/module-2/README.md
@@ -102,10 +102,10 @@ Results.
 
 | Name of Inference    | Time (seconds)      |
 |----------------------|---------------------|
-| Inference 1 worker   | 12.90  |
-| Inference 16 workers (ThreadPoolExecutor) | 0.86  |
-| Inference 16 workers (ProcessPoolExecutor) | 3.88  |
-| Inference with Ray   | 2.15  |
+| Inference 1 worker   | 12.64  |
+| Inference 16 workers (ThreadPoolExecutor) | 0.85  |
+| Inference 16 workers (ProcessPoolExecutor) | 4.03  |
+| Inference with Ray   | 2.19  |
 
 
 # Streaming dataset
@@ -240,4 +240,4 @@ python ./labeling/create_dataset_synthetic.py
 
 ## Updated design doc
 
-[Google doc](https://docs.google.com/document/d/1dEzWd3pPozmU3AhMXjW3xcONUeNJee53djilN1A-wR8/edit)
+[Google doc](https://docs.google.com/document/d/1dEzWd3pPozmU3AhMXjW3xcONUeNJee53djilN1A-wR8/edit)
diff --git a/module-2/data/.gitignore b/module-2/data/.gitignore
diff --git a/module-2/data/big-data.csv.dvc b/module-2/data/big-data.csv.dvc
diff --git a/module-2/processing/inference_example.py b/module-2/processing/inference_example.py
@@ -36,7 +36,6 @@ def run_inference(
     model: DummyClassifier, x_test: np.ndarray, batch_size: int = 2048
 ) -> np.ndarray:
     y_pred = []
-    y_batch = predict(model, x_test)
 
     for i in tqdm(range(0, x_test.shape[0], batch_size)):
         x_batch = x_test[i : i + batch_size]
@@ -156,7 +155,7 @@ def run_pool(inference_size: int = 100_000_000, max_workers: int = 16):
 
 
 def run_ray(inference_size: int = 100_000_000, max_workers: int = 16):
-    ray.init()
+    ray.init(include_dashboard=True, dashboard_host='127.0.0.1', dashboard_port=5000)
 
     x_train, y_train, x_test = get_data(inference_size=inference_size)
     model = train_model(x_train, y_train)
@@ -165,7 +164,6 @@ def run_ray(inference_size: int = 100_000_000, max_workers: int = 16):
     res = run_inference_ray_main(model=model, x_test=x_test, max_workers=max_workers)
     print(f"Inference with Ray {time.monotonic() - s} result: {res.shape}")
 
-
 def run_dask(inference_size: int = 100_000_000, max_workers: int = 16):
     client = Client()