Le Xu

Converting Azure inference trace to workload trace format

1

### 🚀 Feature Description and Motivation Adding a workload generator that generate workload file based on [Azure inference trace](https://github.com/Azure/AzurePublicDataset/tree/master/data) workload trace format (e.g., csv file). ### Use Case The generator...

### 🚀 Feature Description and Motivation Adding a workload generator that generate workload file based on BurstGPT workload trace format (e.g., csv file). ### Use Case The generator should support...

Adding workload generator that converts internal workload traces.

1

### 🚀 Feature Description and Motivation Adding a workload generator that generate workload file based on internal workload trace format (e.g., csv file). ### Use Case The generator should support...

This PR reconfigures workload generator to adopt predefined load types.

Remove model name from client and generator

1

## Pull Request Description This PR removes model name from the workload generator and client.

Autoscaling Benchmark Initial

5

## Pull Request Description Supporting autoscaling experiment scripts

Improve cache awareness for mock app

1

## Pull Request Description This branch enhance cache awareness for aibrix mock inference app.

## Pull Request Description This PR removes await from worker thread for async IO. ## Related Issues #903

[RFC]: xDiT Video Generation API

2

### Summary The goal of this API is to expose a scalable, distributed video generation service for xDiT using models such as CogVideoX, ConsisID, and Latte. The API allows clients...