Prashanth velidandi commited on
Serving large numbers of models with spiky traffic creates cold starts, poor GPU utilization, and complex scheduling tradeoffs.
Browse filesInferX is an inference platform designed for environments where many models are served with irregular traffic. Rather than pinning models to GPUs, InferX treats models as swappable runtime state, allowing GPUs to stay busy while avoiding cold starts.
This Space outlines how InferX fits into long-tail inference scenarios, including platforms like Hugging Face Inference, multi-model agents, and on-demand enterprise workloads.
README.md
CHANGED
|
@@ -1,10 +1,12 @@
|
|
| 1 |
---
|
| 2 |
-
title:
|
| 3 |
-
emoji:
|
| 4 |
colorFrom: blue
|
| 5 |
-
colorTo:
|
| 6 |
sdk: static
|
| 7 |
pinned: false
|
|
|
|
|
|
|
| 8 |
---
|
| 9 |
|
| 10 |
-
Edit this `README.md` markdown file to author your organization card.
|
|
|
|
| 1 |
---
|
| 2 |
+
title: Serverless inference platform for multi-model workloads
|
| 3 |
+
emoji: ⚡
|
| 4 |
colorFrom: blue
|
| 5 |
+
colorTo: blue
|
| 6 |
sdk: static
|
| 7 |
pinned: false
|
| 8 |
+
license: apache-2.0
|
| 9 |
+
short_description: Inference platform focused on cold starts and GPU utilizatio
|
| 10 |
---
|
| 11 |
|
| 12 |
+
Edit this `README.md` markdown file to author your organization card.
|