chore: fix devcontainer GPU memory bug

This commit ensures the model's folder is correctly mounted into
the dev container. By default, the local container is mounted under the
`/workspace` path when using a dev container. Because of this, the fp16
variant is not used, which causes the memory on smaller cards to be
insufficient.

runner/.devcontainer/devcontainer.json

@@ -3,7 +3,7 @@
 {
  "name": "ai-runner",
 
- // Or use a Dockerfile or Docker Compose file. More info: https://containers.dev/guide/dockerfile
+ // Image to use for the dev container. More info: https://containers.dev/guide/dockerfile.
  "build": {
  "dockerfile": "../Dockerfile"
  },
@@ -27,8 +27,13 @@
  // Use 'forwardPorts' to make a list of ports inside the container available locally.
  "forwardPorts": [
  8000
- ]
+ ],
 
+ // Use 'mounts' to make a list of local folders available inside the container.
+ "mounts": [
+ "source=${localWorkspaceFolder}/models,target=/models,type=bind"
+ ]
+
  // Use 'postCreateCommand' to run commands after the container is created.
  // "postCreateCommand": ""