Merge pull request #206 from roboflow/add-inference-ctas

Add Roboflow Inference sections
roboflow · Sep 28, 2023 · 8726cb2 · 8726cb2
2 parents 5d0a9bc + cb3b543
commit 8726cb2
Show file tree

Hide file tree

Showing 6 changed files with 596 additions and 59 deletions.
diff --git a/README.md b/README.md
diff --git a/automation/notebooks-table-data.csv b/automation/notebooks-table-data.csv
@@ -40,3 +40,4 @@ Track and Count Vehicles with YOLOv8 + ByteTRACK + Supervision, how-to-track-and
 Football Players Tracking with YOLOv5 + ByteTRACK, how-to-track-football-players.ipynb, https://blog.roboflow.com/track-football-players, https://youtu.be/QCG8QMhga9k, https://github.com/ifzhang/ByteTrack, 2110.06864, True, skills
 Create Segmentation Masks with Roboflow, how-to-generate-segmentation-mask-with-roboflow.ipynb, https://blog.roboflow.com/how-to-create-segmentation-masks-with-roboflow, , , , False, skills
 How to Use PolygonZone and Roboflow Supervision, how-to-use-polygonzone-annotate-and-supervision.ipynb, https://blog.roboflow.com/polygonzone/, , , , False, skills
+Image-to-Image Search with CLIP and faiss, image-to-image-search-clip-faiss.ipynb, https://blog.roboflow.com/clip-image-search-faiss/, , , , False, skills
diff --git a/notebooks/image-to-image-search-clip-faiss.ipynb b/notebooks/image-to-image-search-clip-faiss.ipynb
diff --git a/notebooks/train-yolov8-classification-on-custom-dataset.ipynb b/notebooks/train-yolov8-classification-on-custom-dataset.ipynb
@@ -1178,6 +1178,62 @@
         "      print(\"\\n\")"
       ]
     },
+    {
+      "cell_type": "markdown",
+      "metadata": {},
+      "source": [
+        "# Deploy Your Model to the Edge\n",
+        "\n",
+        "![Roboflow Inference banner](https://blog.roboflow.com/content/images/2023/08/banner.png)\n",
+        "\n",
+        "In addition to using the Roboflow hosted API for deployment, you can use [Roboflow Inference](https://inference.roboflow.com), an open source inference solution that has powered millions of API calls in production environments. Inference works with CPU and GPU, giving you immediate access to a range of devices, from the NVIDIA Jetson to TRT-compatible devices to ARM CPU devices.\n",
+        "\n",
+        "With Roboflow Inference, you can self-host and deploy your model on-device. You can deploy applications using the [Inference Docker containers](https://inference.roboflow.com/quickstart/docker/) or the pip package.\n",
+        "\n",
+        "For example, to install Inference on a device with an NVIDIA GPU, we can use:\n",
+        "\n",
+        "```\n",
+        "docker pull roboflow/roboflow-inference-server-gpu\n",
+        "```\n",
+        "\n",
+        "Then we can run inference via HTTP:\n",
+        "\n",
+        "```python\n",
+        "import requests\n",
+        "\n",
+        "workspace_id = \"\"\n",
+        "model_id = \"\"\n",
+        "image_url = \"\"\n",
+        "confidence = 0.75\n",
+        "api_key = \"\"\n",
+        "\n",
+        "infer_payload = {\n",
+        "    \"image\": {\n",
+        "        \"type\": \"url\",\n",
+        "        \"value\": image_url,\n",
+        "    },\n",
+        "    \"confidence\": confidence,\n",
+        "    \"iou_threshold\": iou_thresh,\n",
+        "    \"api_key\": api_key,\n",
+        "}\n",
+        "res = requests.post(\n",
+        "    f\"http://localhost:9001/{workspace_id}/{model_id}\",\n",
+        "    json=infer_object_detection_payload,\n",
+        ")\n",
+        "\n",
+        "predictions = res.json()\n",
+        "```\n",
+        "\n",
+        "Above, set your Roboflow workspace ID, model ID, and API key.\n",
+        "\n",
+        "- [Find your workspace and model ID](https://docs.roboflow.com/api-reference/workspace-and-project-ids?ref=blog.roboflow.com)\n",
+        "- [Find your API key](https://docs.roboflow.com/api-reference/authentication?ref=blog.roboflow.com#retrieve-an-api-key)\n",
+        "\n",
+        "Also, set the URL of an image on which you want to run inference. This can be a local file.\n",
+        "\n",
+        "_To use your YOLOv8 model commercially with Inference, you will need a Roboflow Enterprise license, through which you gain a pass-through license for using YOLOv8. An enterprise license also grants you access to features like advanced device management, multi-model containers, auto-batch inference, and more._"
+      ]
+    },
     {
       "attachments": {},
       "cell_type": "markdown",

diff --git a/notebooks/train-yolov8-instance-segmentation-on-custom-dataset.ipynb b/notebooks/train-yolov8-instance-segmentation-on-custom-dataset.ipynb
diff --git a/notebooks/train-yolov8-object-detection-on-custom-dataset.ipynb b/notebooks/train-yolov8-object-detection-on-custom-dataset.ipynb
@@ -1242,6 +1242,67 @@
         "pred"
       ]
     },
+    {
+      "cell_type": "markdown",
+      "metadata": {},
+      "source": [
+        "# Deploy Your Model to the Edge\n",
+        "\n",
+        "In addition to using the Roboflow hosted API for deployment, you can use [Roboflow Inference](https://inference.roboflow.com), an open source inference solution that has powered millions of API calls in production environments. Inference works with CPU and GPU, giving you immediate access to a range of devices, from the NVIDIA Jetson to TRT-compatible devices to ARM CPU devices.\n",
+        "\n",
+        "With Roboflow Inference, you can self-host and deploy your model on-device. You can deploy applications using the [Inference Docker containers](https://inference.roboflow.com/quickstart/docker/) or the pip package.\n",
+        "\n",
+        "For example, to install Inference on a device with an NVIDIA GPU, we can use:\n",
+        "\n",
+        "```\n",
+        "docker pull roboflow/roboflow-inference-server-gpu\n",
+        "```\n",
+        "\n",
+        "Then we can run inference via HTTP:\n",
+        "\n",
+        "```python\n",
+        "import requests\n",
+        "\n",
+        "workspace_id = \"\"\n",
+        "model_id = \"\"\n",
+        "image_url = \"\"\n",
+        "confidence = 0.75\n",
+        "api_key = \"\"\n",
+        "\n",
+        "infer_payload = {\n",
+        "    \"image\": {\n",
+        "        \"type\": \"url\",\n",
+        "        \"value\": image_url,\n",
+        "    },\n",
+        "    \"confidence\": confidence,\n",
+        "    \"iou_threshold\": iou_thresh,\n",
+        "    \"api_key\": api_key,\n",
+        "}\n",
+        "res = requests.post(\n",
+        "    f\"http://localhost:9001/{workspace_id}/{model_id}\",\n",
+        "    json=infer_object_detection_payload,\n",
+        ")\n",
+        "\n",
+        "predictions = res.json()\n",
+        "```\n",
+        "\n",
+        "Above, set your Roboflow workspace ID, model ID, and API key.\n",
+        "\n",
+        "- [Find your workspace and model ID](https://docs.roboflow.com/api-reference/workspace-and-project-ids?ref=blog.roboflow.com)\n",
+        "- [Find your API key](https://docs.roboflow.com/api-reference/authentication?ref=blog.roboflow.com#retrieve-an-api-key)\n",
+        "\n",
+        "Also, set the URL of an image on which you want to run inference. This can be a local file.\n",
+        "\n",
+        "_To use your YOLOv5 model commercially with Inference, you will need a Roboflow Enterprise license, through which you gain a pass-through license for using YOLOv5. An enterprise license also grants you access to features like advanced device management, multi-model containers, auto-batch inference, and more._"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": null,
+      "metadata": {},
+      "outputs": [],
+      "source": []
+    },
     {
       "attachments": {},
       "cell_type": "markdown",