EngineX-Hygon/sglang
5
0
Fork 0
Code Issues Pull Requests Actions 7 Projects Releases Wiki Activity

Enable MLA by default (#1447)

Browse Source
This commit is contained in:
Ke Bao
2024-09-17 19:42:48 +08:00
committed by GitHub
parent 90a26be31c
commit c6b6d2e71b
8 changed files with 16 additions and 18 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
docs/en/backend.md
View File

@@ -81,7 +81,6 @@ python -m sglang.launch_server --model-path meta-llama/Meta-Llama-3-8B-Instruct
- To enable torch.compile acceleration, add `--enable-torch-compile`. It accelerates small models on small batch sizes.
- To enable fp8 weight quantization, add `--quantization fp8` on a fp16 checkpoint or directly load a fp8 checkpoint without specifying any arguments.
- To enable fp8 kv cache quantization, add `--kv-cache-dtype fp8_e5m2`.
- To enable DeepSeek MLA acceleration, add `--enable-mla`.
- If the model does not have a chat template in the Hugging Face tokenizer, you can specify a [custom chat template](docs/en/custom_chat_template.md).
- To run tensor parallelism on multiple nodes, add `--nnodes 2`. If you have two nodes with two GPUs on each node and want to run TP=4, let `sgl-dev-0` be the hostname of the first node and `50000` be an available port.
```