docs: clarify remaining v0 references #26311
Conversation
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
This pull request provides a number of documentation updates to remove references to the legacy v0 engine and clarify concepts for the current v1 engine. The changes are well-executed across multiple files, improving the clarity and relevance of the documentation for users. The updates are consistent with the stated goals of the PR, and I have no further suggestions.
|
|
||
| We have started the process of deprecating V0. Please read [RFC #18571](gh-issue:18571) for more details. | ||
|
|
||
| V1 is now enabled by default for all supported use cases, and we will gradually enable it for every use case we plan to support. Please share any feedback on [GitHub](https://github.com/vllm-project/vllm) or in the [vLLM Slack](https://inviter.co/vllm-slack). |
There was a problem hiding this comment.
Also update this paragraph?
| | **Mamba Models** | <nobr>🟢 (Mamba-2), 🟢 (Mamba-1)</nobr> | | ||
| | **Multimodal Models** | <nobr>🟢 Functional</nobr> | | ||
|
|
||
| vLLM V1 currently excludes model architectures with the `SupportsV0Only` protocol. |
There was a problem hiding this comment.
We should remove the V1 column from the Supported Models page and delete all models that don't support V1
| Chunked prefill allows vLLM to process large prefills in smaller chunks and batch them together with decode requests. This feature helps improve both throughput and latency by better balancing compute-bound (prefill) and memory-bound (decode) operations. | ||
|
|
||
| In vLLM V1, **chunked prefill is always enabled by default**. This is different from vLLM V0, where it was conditionally enabled based on model characteristics. | ||
| In vLLM V1, **chunked prefill is always enabled by default** so that behavior is consistent across supported models. |
There was a problem hiding this comment.
| In vLLM V1, **chunked prefill is always enabled by default** so that behavior is consistent across supported models. | |
| In vLLM V1, **chunked prefill is always enabled by default**. |
There was a problem hiding this comment.
There are probably some mistakes here. @markmc PTAL
There was a problem hiding this comment.
@njhill I guess this page can use a full clean up
| Speculative decoding with a draft model requires the V1 engine. | ||
| Older releases that predate V1 (such as the 0.10.x series) raise a `NotImplementedError`. |
There was a problem hiding this comment.
| Speculative decoding with a draft model requires the V1 engine. | |
| Older releases that predate V1 (such as the 0.10.x series) raise a `NotImplementedError`. | |
| Speculative decoding with a draft model is not supported in vLLM V1 version. | |
| You can use older version before the 0.10x series to continue to leverage it. |
There was a problem hiding this comment.
@DarkLight1337 PTAL
There was a problem hiding this comment.
We should remove the V1 column from the Supported Models page and delete all models that don't support V1
LGTM after doing this
There was a problem hiding this comment.
We can probably gradually remove this docs
Summary
Testing
https://chatgpt.com/codex/tasks/task_e_68e3f11c47408329bf2324ac7b1ad7bf