v2.1.0

• changed max_tokens=100 for runpod specific APIs.
• updated vllm to v0.7.3

v1.9.0

• vLLM's V1 engine is ready for testing! This is a rewritten engine designed for performance and architectural simplicity. You can turn it on by setting environment variable VLLM_USE_V1=1.

More updates:
Follow this link: https://github.com/vllm-project/vllm/releases/tag/v0.7.0

v1.8.0

• Update vllm version from v0.6.4 --> v0.6.6.
• Support for DeepSeek V3 Model.

v1.7.0

• Update vllm to 0.6.4.

v1.6.0

Merge pull request #125 from runpod-workers/up-0.6.3

update vllm

v1.5.0

• vllm version update 0.6.1 --> 0.6.2.
• Supports llama 3.2 Models.

v1.4.0: Merge pull request #109 from runpod-workers/0.5.5-update

vllm version update 0.5.5 --> 0.6.1

v1.3.1

vLLm version: 0.5.5

• OpenAI Completion Requests Bug fix.

v1.3.0

Version upgrade from vllm v0.5.4 -> v0.5.5

Various improvements and bug fixes.
[Known Issue]: OpenAI Completion Requests error.

v1.2.0

Version upgrade from vllm v0.5.3 -> v0.5.4

• Various improvements and bug fixes.
• [Known Issue]: OpenAI Completion Requests error.