A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 15, 16:00 UTC
doc: fix some typos in doc (#16154)
Signed-off-by: yihong0618 <zouzou0208@gmail.com>
yihong committed
Apr 7, 2025 at 13:32 UTC
95d63f38c039e6fce57cf9cddb4c32bbc655a376
3 files changed
+3
-3
docs/source/design/mm_processing.md
+1
-1
@@ -8,7 +8,7 @@ Here are the main features of {class}`~vllm.multimodal.processing.BaseMultiModal
8
9
## Prompt Update Detection
10
11
-One of the main responsibilies of HF processor is to update the prompt with placeholder tokens. For example:
11
+One of the main responsibilities of HF processor is to update the prompt with placeholder tokens. For example:
12
13
- Insert feature placeholder tokens (e.g. `<image><image>...<image>`, the number of which equals to the feature size) at the start of the string.
14
- Replace existing input placeholder tokens (e.g. `<image>` for a single image) with feature placeholder tokens (e.g. `<image><image>...<image>`, the number of which equals to the feature size).
docs/source/design/v1/torch_compile.md
+1
-1
@@ -126,7 +126,7 @@ Unfortunately, because auto-tuning takes quite a long time (from seconds to minu
126
127
## Cudagraph Capture
128
129
-vLLM's V1 architecture uses piecewise cudagraph. The full computation graph is split as mentioned above, and we only capture the cudagraph for the piece of graph between attention operations (including the first graph before any attention operation, and the last graph after all the attention operation). This is based on a common observation: computation between attentions are usually token-wise and easy to deal with for cudagraph; while the attention operation is non-trival to be cudagraph compatible. Thus, by running the attention operation in eager mode while the rest operations in cudagraph, we keep the flexibility of the attention operation.
129
+vLLM's V1 architecture uses piecewise cudagraph. The full computation graph is split as mentioned above, and we only capture the cudagraph for the piece of graph between attention operations (including the first graph before any attention operation, and the last graph after all the attention operation). This is based on a common observation: computation between attentions are usually token-wise and easy to deal with for cudagraph; while the attention operation is non-trivial to be cudagraph compatible. Thus, by running the attention operation in eager mode while the rest operations in cudagraph, we keep the flexibility of the attention operation.
130
131
The piecewise cudagraph also has fine-grained memory management. The purpose is to only exclude the attention kernel from cudagraph, while keeping all the rest modules and the memory allocation operations in the cudagraph. This is why the attention operation in V1 has the output tensor as the input of the attention.
132
docs/source/features/quantization/gguf.md
+1
-1
@@ -29,7 +29,7 @@ vllm serve ./tinyllama-1.1b-chat-v1.0.Q4_K_M.gguf --tokenizer TinyLlama/TinyLlam
29
We recommend using the tokenizer from base model instead of GGUF model. Because the tokenizer conversion from GGUF is time-consuming and unstable, especially for some models with large vocab size.
30
:::
31
32
-GGUF assumes that huggingface can convert the metadata to a config file. In case huggingface doesn't support your model you can manually create a config and pass it as hf-confing-path
32
+GGUF assumes that huggingface can convert the metadata to a config file. In case huggingface doesn't support your model you can manually create a config and pass it as hf-config-path
33
34
```console
35
# If you model is not supported by huggingface you can manually provide a huggingface compatible config path