Release v0.7.2

[simon-mo]

Update (01/31/2025):

• This has been renamed to v0.7.2 as we are releasing v0.7.1 for Deepseek enhancements.

Blockers

• Support for Qwen-1M: Implements dual-chunk-flash-attn backend for dual chunk attention with sparse attention support #11844
• Support for Baichuan-M1: [Model] Enable Inference Support for the New Baichuan-M1 Model #12251
• Support for MiniCPM-o: [Model] Refactoring of MiniCPM-V and add MiniCPM-o-2.6 support for vLLM #12069
• Support for Qwen2.5VL [Draft] Qwen2.5-VL #12604

[simon-mo]

From v0.7.0 tracker (#12365)

• Accuracy error for Mistral model: Tests started failing between https://buildkite.com/vllm/ci/builds/11418#01942acd-a27b-477a-809b-1fd358291b41 and https://buildkite.com/vllm/ci/builds/11468#01942fb0-18f3-4deb-8bb4-8dbcaa2987e1
• Memory profiling error for Mamba model: Tests started failing between https://buildkite.com/vllm/ci/builds/11861#019458e3-1b40-45b0-9a45-d5cd2975ec1a and https://buildkite.com/vllm/ci/builds/11926#0194632f-b3a9-43e2-9d5f-a01bf7233bca
• Shape mismatch error in GGUF tests: Tests started failing between https://buildkite.com/vllm/ci/builds/11861#019458e3-1b40-45b0-9a45-d5cd2975ec1a and https://buildkite.com/vllm/ci/builds/11926#0194632f-b3a9-43e2-9d5f-a01bf7233bca
  • Fixed by [Bugfix] Fix gpt2 GGUF inference #12467

[Isotr0py]

Shape mismatch error in GGUF tests: Tests started failing between https://buildkite.com/vllm/ci/builds/11861#019458e3-1b40-45b0-9a45-d5cd2975ec1a and https://buildkite.com/vllm/ci/builds/11926#0194632f-b3a9-43e2-9d5f-a01bf7233bca

Seems that the CI links for this issue are wrong? It's same as Memory profiling error ones, and I didn't see the gguf failing in the logs of these two CI.

NVM, I missed the line. 😂

[DarkLight1337]

I posted two links for each error - the first link is the most recent nightly build without the error, and the second link is the earliest nightly build with the error

[hmellor]

Issue for Qwen2.5-VL: #12486