Session Status

{#if compute.hasSession} {compute.statusInfo.statusText} {:else} No Session {/if}

{#if compute.hasSession && compute.sessionData} Current Session

Session ID: {compute.sessionId}

Status: {compute.statusInfo.emoji} {compute.statusInfo.statusText}

Policy: {compute.sessionConfig?.policyPath}

Cameras: {compute.sessionConfig?.cameraNames.join(", ")}

📡 Inference Server Connections

Workspace: {compute.sessionData.workspace_id}

{#each Object.entries(compute.sessionData.camera_room_ids) as [camera, roomId]}

📹 {camera}: {roomId}

{/each}

📥 Joint Input: {compute.sessionData.joint_input_room_id}

📤 Joint Output: {compute.sessionData.joint_output_room_id}

{#if compute.canStart} {/if} {#if compute.canStop} {/if}

{/if} {#if !compute.hasSession} Create Inference Session

Session ID

Policy Path

Camera Names

Comma-separated camera names

Transport Server URL

Configure in settings panel

Use current workspace ({workspaceId})

This will create a new ACT inference session with dedicated rooms for camera inputs, joint inputs, and joint outputs in the inference server communication system.

{/if}

Inference Sessions require a trained ACT model and create dedicated communication rooms for video inputs, robot joint states, and control outputs in the inference server system.