xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2025-12-09 18:46:25 +08:00
Code Issues Packages Projects Releases Wiki Activity

[OpenVINO] Updated documentation (#7687)

Browse Source
This commit is contained in:
Ilya Lavrenov 2024-08-20 17:33:56 +04:00 committed by GitHub
parent 5288c06aa0
commit 398521ad19
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 1 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
docs/source/getting_started/openvino-installation.rst
View File

@ -70,7 +70,7 @@ vLLM OpenVINO backend uses the following environment variables to control behavi
- ``VLLM_OPENVINO_CPU_KV_CACHE_PRECISION=u8`` to control KV cache precision. By default, FP16 / BF16 is used depending on platform.
- ``VLLM_OPENVINO_ENABLE_QUANTIZED_WEIGHTS=ON`` to enable U8 weights compression during model loading stage. By default, compression is turned off.
- ``VLLM_OPENVINO_ENABLE_QUANTIZED_WEIGHTS=ON`` to enable U8 weights compression during model loading stage. By default, compression is turned off. You can also export model with different compression techniques using `optimum-cli` and pass exported folder as `<model_id>`
To enable better TPOT / TTFT latency, you can use vLLM's chunked prefill feature (``--enable-chunked-prefill``). Based on the experiments, the recommended batch size is ``256`` (``--max-num-batched-tokens``)
@ -91,5 +91,3 @@ Limitations
- Only LLM models are currently supported. LLaVa and encoder-decoder models are not currently enabled in vLLM OpenVINO integration.
- Tensor and pipeline parallelism are not currently enabled in vLLM integration.
- Speculative sampling is not tested within vLLM integration.