FEP
The live control plane for Friday infrastructure. Monitor clusters, deploy models, schedule GPU work, and govern access from one surface — dashboard to fleet, hub to compute.
Dashboard
1/2 clusters online · 4 workers · 8 GPUs
RefreshCluster GPU utilization
LIVENode GPU utilization
LIVECompute idle · VRAM 0%
Node power draw
LIVEMemory utilization
LIVEMemory allocation
LIVEGPU temperature
LIVETasks running
LIVEApps serving
LIVEClusters & Nodes
Real-time overview of all your GPU clusters and their resources.
Overview
Live cluster health in one board
Workers, GPUs, queues, endpoints, utilization, power, and temperature — the same Dashboard operators use in the Friday console.
Cluster GPU utilization
Nodes, VRAM, and temperature
Queues and active jobs
Apps serving in real time
Overview
Fleet Map puts every site on the board
Track Veronica and edge assets by status and location. Healthy, attention, and offline nodes stay visible — even when the site is remote.
Live asset positions
Healthy vs attention
Group by type and region
Status without a delayed report
Fleet Map
Catalog
One-click deploy through FEP pipeline
RefreshGemma 4 12B Instruct
gemma-4-12b-it
Capabilities
ChatMin GPU
gemma — Google Gemma Terms of Use apply
Replaces Gemma 2 9B Instruct on the 2×T4 envelope.
GLM-4 9B Chat
glm-4-9b-chat
Capabilities
ChatMin GPU
glm-4 — THUDM/Zhipu GLM license terms apply
T4-fit pin for the GLM-4 family; single-GPU co-host candidate.
PrismLM Bonsai 8B
prism-bonsai-8b
Capabilities
ChatMin GPU
apache-2.0
Tiny VRAM envelope for co-hosting demos on cluster GPUs.
Hub
Catalog one-click deploy through FEP
Browse approved models, set a deployment name, and push to cluster GPUs. Min GPU, license, and capability stay on the card.
Chat and embedding models
Min GPU called out
Deploy from the card
Approved, versioned catalog
Hub
Deployments from job to serving
ClearML tasks via the FEP API — name, model, job status, serving, queue, and actions in one live table.
Completed and serving states
Metrics and reroute
Stop and delete controls
Auto-refresh from FEP
Deployments
ClearML tasks via FEP API · live (auto-refresh)
| Name | Model | Job status | Serving | Queue | Created | Actions |
|---|---|---|---|---|---|---|
Xx Test 1 | GLM-4 9B Chat glm-4-9b-chat | Completed | Serving | — | 09/09/2026, 04:37:05 | |
PrismLM Bonsai 8B(reg) | PrismLM Bonsai 8B prism-bonsai-8b | Completed | Serving | — | 09/09/2026, 03:51:22 | |
CoVec1 4B | CoVec1 4B covec-4b | Completed | Serving | — | 09/09/2026, 02:18:44 | |
Gemma 4 12B Instruct | Gemma 4 12B Instruct gemma-4-12b-it | Completed | Serving | — | 08/09/2026, 21:04:11 | |
GLM-4 9B Chat | GLM-4 9B Chat glm-4-9b-chat | Completed | Serving | — | 08/09/2026, 18:42:37 | |
PrismLM Bonsai 8B | PrismLM Bonsai 8B prism-bonsai-8b | Completed | Serving | — | 08/09/2026, 14:09:53 | |
TextGenerator | Gemma 4 12B Instruct gemma-4-12b-it | Completed | Serving | — | 07/09/2026, 11:26:18 |