EngineX/xc-llm-ascend
3
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

[doc][main] Correct more doc mistakes (#4958)

Browse Source 
### What this PR does / why we need it?
Correct more doc mistakes

- vLLM version: v0.12.0
- vLLM main:
ad32e3e19c

Signed-off-by: lilinsiman <lilinsiman@gmail.com>
This commit is contained in:
lilinsiman
2025-12-13 18:36:58 +08:00
committed by GitHub
parent 4721e4f53f
commit 31c94b7e7b
7 changed files with 509 additions and 228 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
docs/source/locale/zh_CN/LC_MESSAGES/user_guide/support_matrix/supported_models.po
View File

@@ -169,10 +169,6 @@ msgstr "Qwen2-VL"
msgid "Qwen2.5-VL"
msgstr "Qwen2.5-VL"
#: ../../user_guide/support_matrix/supported_models.md
msgid "[#553](https://github.com/vllm-project/vllm-ascend/issues/553)"
msgstr "[#553](https://github.com/vllm-project/vllm-ascend/issues/553)"
#: ../../user_guide/support_matrix/supported_models.md
msgid "LLaVA-Next"
msgstr "LLaVA-Next"

220
docs/source/tutorials/Qwen2_audio.md
View File

@@ -1,220 +0,0 @@
# Qwen2-Audio-7B
## Run vllm-ascend on Single NPU
### Offline Inference on Single NPU
Run docker container:
```{code-block} bash
:substitutions:
# Update the vllm-ascend image
export IMAGE=quay.io/ascend/vllm-ascend:|vllm_ascend_version|
docker run --rm \
--name vllm-ascend \
--shm-size=1g \
--device /dev/davinci0 \
--device /dev/davinci_manager \
--device /dev/devmm_svm \
--device /dev/hisi_hdc \
-v /usr/local/dcmi:/usr/local/dcmi \
-v /usr/local/bin/npu-smi:/usr/local/bin/npu-smi \
-v /usr/local/Ascend/driver/lib64/:/usr/local/Ascend/driver/lib64/ \
-v /usr/local/Ascend/driver/version.info:/usr/local/Ascend/driver/version.info \
-v /etc/ascend_install.info:/etc/ascend_install.info \
-v /root/.cache:/root/.cache \
-p 8000:8000 \
-it $IMAGE bash
```
Set up environment variables:
```bash
# Load model from ModelScope to speed up download
export VLLM_USE_MODELSCOPE=True
# Set `max_split_size_mb` to reduce memory fragmentation and avoid out of memory
export PYTORCH_NPU_ALLOC_CONF=max_split_size_mb:256
```
:::{note}
`max_split_size_mb` prevents the native allocator from splitting blocks larger than this size (in MB). This can reduce fragmentation and may allow some borderline workloads to complete without running out of memory. You can find more details [<u>here</u>](https://www.hiascend.com/document/detail/zh/CANNCommunityEdition/800alpha003/apiref/envref/envref_07_0061.html).
:::
Install packages required for audio processing:
```bash
pip config set global.index-url https://mirrors.tuna.tsinghua.edu.cn/pypi/web/simple
pip install librosa soundfile
```
Run the following script to execute offline inference on a single NPU:
```python
from vllm import LLM, SamplingParams
from vllm.assets.audio import AudioAsset
from vllm.utils import FlexibleArgumentParser
# If network issues prevent AudioAsset from fetching remote audio files, retry or check your network.
audio_assets = [AudioAsset("mary_had_lamb"), AudioAsset("winning_call")]
question_per_audio_count = {
1: "What is recited in the audio?",
2: "What sport and what nursery rhyme are referenced?"
}
def prepare_inputs(audio_count: int):
audio_in_prompt = "".join([
f"Audio {idx+1}: <|audio_bos|><|AUDIO|><|audio_eos|>\n"
for idx in range(audio_count)
])
question = question_per_audio_count[audio_count]
prompt = ("<|im_start|>system\nYou are a helpful assistant.<|im_end|>\n"
"<|im_start|>user\n"
f"{audio_in_prompt}{question}<|im_end|>\n"
"<|im_start|>assistant\n")
mm_data = {
"audio":
[asset.audio_and_sample_rate for asset in audio_assets[:audio_count]]
}
# Merge text prompt and audio data into inputs
inputs = {"prompt": prompt, "multi_modal_data": mm_data}
return inputs
def main(audio_count: int):
# NOTE: The default `max_num_seqs` and `max_model_len` may result in OOM on
# lower-end GPUs.
# Unless specified, these settings have been tested to work on a single L4.
# `limit_mm_per_prompt`: the max num items for each modality per prompt.
llm = LLM(model="Qwen/Qwen2-Audio-7B-Instruct",
max_model_len=4096,
max_num_seqs=5,
limit_mm_per_prompt={"audio": audio_count})
inputs = prepare_inputs(audio_count)
sampling_params = SamplingParams(temperature=0.2,
max_tokens=64,
stop_token_ids=None)
outputs = llm.generate(inputs, sampling_params=sampling_params)
for o in outputs:
generated_text = o.outputs[0].text
print(generated_text)
if __name__ == "__main__":
audio_count = 2
main(audio_count)
```
If you run this script successfully, you can see the info shown below:
```bash
The sport referenced is baseball, and the nursery rhyme is 'Mary Had a Little Lamb'.
```
### Online Serving on Single NPU
Currently, the `chat_template` for `Qwen2-Audio` has some issues which caused audio placeholder failed to be inserted, find more details [<u>here</u>](https://github.com/vllm-project/vllm/issues/19977).
Nevertheless, we could use a custom template for online serving, which is shown below:
```jinja
{% set audio_count = namespace(value=0) %}
{% for message in messages %}
{% if loop.first and message['role'] != 'system' %}
<|im_start|>system\nYou are a helpful assistant.<|im_end|>\n
{% endif %}
<|im_start|>{{ message['role'] }}\n
{% if message['content'] is string %}
{{ message['content'] }}<|im_end|>\n
{% else %}
{% for content in message['content'] %}
{% if 'audio' in content or 'audio_url' in content or message['type'] == 'audio' or content['type'] == 'audio' %}
{% set audio_count.value = audio_count.value + 1 %}
Audio {{ audio_count.value }}: <|audio_bos|><|AUDIO|><|audio_eos|>\n
{% elif 'text' in content %}
{{ content['text'] }}
{% endif %}
{% endfor %}
<|im_end|>\n
{% endif %}
{% endfor %}
{% if add_generation_prompt %}
<|im_start|>assistant\n
{% endif %}
```
:::{note}
You can find this template at `vllm-ascend/examples/chat_templates/template_qwen2_audio.jinja`.
:::
Run docker container to start the vLLM server on a single NPU:
```{code-block} bash
# Update the vllm-ascend image
export IMAGE=quay.io/ascend/vllm-ascend:|vllm_ascend_version|
docker run --rm \
--name vllm-ascend \
--shm-size=1g \
--device /dev/davinci0 \
--device /dev/davinci_manager \
--device /dev/devmm_svm \
--device /dev/hisi_hdc \
-v /usr/local/dcmi:/usr/local/dcmi \
-v /usr/local/bin/npu-smi:/usr/local/bin/npu-smi \
-v /usr/local/Ascend/driver/lib64/:/usr/local/Ascend/driver/lib64/ \
-v /usr/local/Ascend/driver/version.info:/usr/local/Ascend/driver/version.info \
-v /etc/ascend_install.info:/etc/ascend_install.info \
-v /root/.cache:/root/.cache \
-p 8000:8000 \
-e VLLM_USE_MODELSCOPE=True \
-e PYTORCH_NPU_ALLOC_CONF=max_split_size_mb:256 \
-it $IMAGE \
vllm serve Qwen/Qwen2-Audio-7B-Instruct \
--max_model_len 16384 \
--max-num-batched-tokens 16384 \
--limit-mm-per-prompt '{"audio":2}' \
--chat-template /path/to/your/vllm-ascend/examples/chat_templates/template_qwen2_audio.jinja
```
:::{note}
Replace `/path/to/your/vllm-ascend` with your own path.
:::
If your service start successfully, you can see the info shown below:
```bash
INFO: Started server process [2736]
INFO: Waiting for application startup.
INFO: Application startup complete.
```
Once your server is started, you can query the model with input prompts:
```bash
curl -X POST http://localhost:8000/v1/chat/completions \
-H "Content-Type: application/json" \
-d '{
"model": "/root/.cache/modelscope/models/Qwen/Qwen2-Audio-7B-Instruct",
"messages": [
{"role": "system", "content": "You are a helpful assistant."},
{"role": "user", "content": [
{"type": "audio_url", "audio_url": {"url": "https://vllm-public-assets.s3.us-west-2.amazonaws.com/multimodal_asset/winning_call.ogg"}},
{"type": "text", "text": "What is in this audio? How does it sound?"}
]}
],
"max_tokens": 100
}'
```
If you query the server successfully, you can see the info shown below (client):
```bash
{"id":"chatcmpl-31f5f698f6734a4297f6492a830edb3f","object":"chat.completion","created":1761097383,"model":"/root/.cache/modelscope/models/Qwen/Qwen2-Audio-7B-Instruct","choices":[{"index":0,"message":{"role":"assistant","content":"The audio contains a background of a crowd cheering, a ball bouncing, and an object being hit. A man speaks in English saying 'and the o one pitch on the way to edgar martinez swung on and lined out.' The speech has a happy mood.","refusal":null,"annotations":null,"audio":null,"function_call":null,"tool_calls":[],"reasoning_content":null},"logprobs":null,"finish_reason":"stop","stop_reason":null,"token_ids":null}],"service_tier":null,"system_fingerprint":null,"usage":{"prompt_tokens":689,"total_tokens":743,"completion_tokens":54,"prompt_tokens_details":null},"prompt_logprobs":null,"prompt_token_ids":null,"kv_transfer_params":null}
```

1
docs/source/tutorials/index.md
View File

@@ -4,7 +4,6 @@
:caption: Deployment
:maxdepth: 1
Qwen2.5-Omni.md
Qwen2_audio
Qwen2.5-7B
Qwen3-Dense
Qwen-VL-Dense.md

4
docs/source/tutorials/ray.md
View File

@@ -131,6 +131,10 @@ ray start --address='{head_node_ip}:6379' --node-ip-address={local_ip}
Once the cluster is started on multiple nodes, execute `ray status` and `ray list nodes` to verify the Ray cluster's status. You should see the correct number of nodes and NPUs listed.
After Ray is successfully started, the following content will appear:\
A local Ray instance has started successfully.\
Dashboard URL: The access address for the Ray Dashboard (default: http://localhost:8265); Node status (CPU/memory resources, number of healthy nodes); Cluster connection address (used for adding multiple nodes).
## Start the Online Inference Service on Multi-node scenario
In the container, you can use vLLM as if all NPUs were on a single node. vLLM will utilize NPU resources across all nodes in the Ray cluster.

1
docs/source/user_guide/feature_guide/index.md
View File

@@ -16,4 +16,5 @@ netloader
dynamic_batch
kv_pool
external_dp
large_scale_ep
:::

504
docs/source/user_guide/feature_guide/large_scale_ep.md Normal file
View File

@@ -0,0 +1,504 @@
# Distributed DP Server With Large Scale Expert Parallelism
## Getting Start
vLLM-Ascend now supports prefill-decode (PD) disaggregation in the large scale **Expert Parallelism (EP)** scenario. To achieve better performance，the distributed DP server is applied in vLLM-Ascend. In the PD separation scenario, different optimization strategies can be implemented based on the distinct characteristics of PD nodes, thereby enabling more flexible model deployment. \
Take the deepseek model as an example, use 8 Atlas 800T A3 servers to deploy the model. Assume the ip of the servers start from 192.0.0.1, and end by 192.0.0.8. Use the first 4 servers as prefiller nodes and the last 4 servers as decoder nodes. And the prefiller nodes deployed as master node independently, the decoder nodes set 192.0.0.5 node to be the master node.
## Verify Multi-Node Communication Environment
### Physical Layer Requirements:
- The physical machines must be located on the same WLAN, with network connectivity.
- All NPUs must be interconnected. For the Atlas A2 generation, intra-node connectivity is via HCCS, and inter-node connectivity is via RDMA. For the Atlas A3 generation, both intra-node and inter-node connectivity are via HCCS.
### Verification Process:
:::::{tab-set}
::::{tab-item} A3
1. Single Node Verification:
Execute the following commands on each node in sequence. The results must all be `success` and the status must be `UP`:
```bash
# Check the remote switch ports
for i in {0..15}; do hccn_tool -i $i -lldp -g | grep Ifname; done
# Get the link status of the Ethernet ports (UP or DOWN)
for i in {0..15}; do hccn_tool -i $i -link -g ; done
# Check the network health status
for i in {0..15}; do hccn_tool -i $i -net_health -g ; done
# View the network detected IP configuration
for i in {0..15}; do hccn_tool -i $i -netdetect -g ; done
# View gateway configuration
for i in {0..15}; do hccn_tool -i $i -gateway -g ; done
# View NPU network configuration
cat /etc/hccn.conf
```
2. Get NPU IP Addresses
```bash
for i in {0..15}; do hccn_tool -i $i -vnic -g;done
```
3. Get superpodid and SDID
```bash
for i in {0..15}; do npu-smi info -t spod-info -i $i -c 0;npu-smi info -t spod-info -i $i -c 1;done
```
4. Cross-Node PING Test
```bash
# Execute on the target node (replace 'x.x.x.x' with actual npu ip address)
for i in {0..15}; do hccn_tool -i $i -hccs_ping -g address x.x.x.x;done
```
::::
::::{tab-item} A2
1. Single Node Verification:
Execute the following commands on each node in sequence. The results must all be `success` and the status must be `UP`:
```bash
# Check the remote switch ports
for i in {0..7}; do hccn_tool -i $i -lldp -g | grep Ifname; done
# Get the link status of the Ethernet ports (UP or DOWN)
for i in {0..7}; do hccn_tool -i $i -link -g ; done
# Check the network health status
for i in {0..7}; do hccn_tool -i $i -net_health -g ; done
# View the network detected IP configuration
for i in {0..7}; do hccn_tool -i $i -netdetect -g ; done
# View gateway configuration
for i in {0..7}; do hccn_tool -i $i -gateway -g ; done
# View NPU network configuration
cat /etc/hccn.conf
```
2. Get NPU IP Addresses
```bash
for i in {0..7}; do hccn_tool -i $i -ip -g;done
```
3. Cross-Node PING Test
```bash
# Execute on the target node (replace 'x.x.x.x' with actual npu ip address)
for i in {0..7}; do hccn_tool -i $i -ping -g address x.x.x.x;done
```
::::
:::::
## Large Scale EP model deployment
### Generate script with configurations
In the PD separation scenario, we provide a optimized configuration. You can use the following shell script for configuring the prefiller and decoder nodes respectively.
:::::{tab-set}
::::{tab-item} Prefiller node
```shell
# run_dp_template.sh
#!/bin/sh
# this obtained through ifconfig
# nic_name is the network interface name corresponding to local_ip
nic_name="xxxx"
local_ip="xxxx"
# basic configuration for HCCL and connection
export HCCL_IF_IP=$local_ip
export GLOO_SOCKET_IFNAME=$nic_name
export TP_SOCKET_IFNAME=$nic_name
export HCCL_SOCKET_IFNAME=$nic_name
export OMP_PROC_BIND=false
export OMP_NUM_THREADS=10
export HCCL_BUFFSIZE=256
# obtain parameters from distributed DP server
export VLLM_DP_SIZE=$1
export VLLM_DP_MASTER_IP=$2
export VLLM_DP_MASTER_PORT=$3
export VLLM_DP_RANK_LOCAL=$4
export VLLM_DP_RANK=$5
export VLLM_DP_SIZE_LOCAL=$7
#pytorch_npu settings and vllm settings
export PYTORCH_NPU_ALLOC_CONF=expandable_segments:True
export TASK_QUEUE_ENABLE=1
export VLLM_USE_MODELSCOPE="True"
# enable the distributed DP server
export VLLM_WORKER_MULTIPROC_METHOD="fork"
export VLLM_ASCEND_EXTERNAL_DP_LB_ENABLED=1
# The w8a8 weight can obtained from https://www.modelscope.cn/models/vllm-ascend/DeepSeek-R1-W8A8
# "--additional-config" is used to enable characteristics from vllm-ascend
vllm serve vllm-ascend/DeepSeek-R1-W8A8 \
--host 0.0.0.0 \
--port $6 \
--tensor-parallel-size 8 \
--enable-expert-parallel \
--seed 1024 \
--served-model-name deepseek_r1 \
--max-model-len 17000 \
--max-num-batched-tokens 16384 \
--trust-remote-code \
--max-num-seqs 4 \
--gpu-memory-utilization 0.9 \
--quantization ascend \
--speculative-config '{"num_speculative_tokens": 1, "method":"deepseek_mtp"}' \
--enforce-eager \
--kv-transfer-config \
'{"kv_connector": "MooncakeConnector",
"kv_buffer_device": "npu",
"kv_role": "kv_producer",
"kv_parallel_size": "1",
"kv_port": "20001",
"engine_id": "0",
"kv_connector_module_path": "vllm_ascend.distributed.mooncake_connector"
}'
--additional-config '{"enable_weight_nz_layout":true,"enable_prefill_optimizations":true}'
```
::::
::::{tab-item} Decoder node
```shell
# run_dp_template.sh
#!/bin/sh
# this obtained through ifconfig
# nic_name is the network interface name corresponding to local_ip
nic_name="xxxx"
local_ip="xxxx"
# basic configuration for HCCL and connection
export HCCL_IF_IP=$local_ip
export GLOO_SOCKET_IFNAME=$nic_name
export TP_SOCKET_IFNAME=$nic_name
export HCCL_SOCKET_IFNAME=$nic_name
export OMP_PROC_BIND=false
export OMP_NUM_THREADS=10
export HCCL_BUFFSIZE=1024
# obtain parameters from distributed DP server
export VLLM_DP_SIZE=$1
export VLLM_DP_MASTER_IP=$2
export VLLM_DP_MASTER_PORT=$3
export VLLM_DP_RANK_LOCAL=$4
export VLLM_DP_RANK=$5
export VLLM_DP_SIZE_LOCAL=$7
#pytorch_npu settings and vllm settings
export PYTORCH_NPU_ALLOC_CONF=expandable_segments:True
export TASK_QUEUE_ENABLE=1
export VLLM_USE_MODELSCOPE="True"
# enable the distributed DP server
export VLLM_WORKER_MULTIPROC_METHOD="fork"
export VLLM_ASCEND_EXTERNAL_DP_LB_ENABLED=1
# The w8a8 weight can obtained from https://www.modelscope.cn/models/vllm-ascend/DeepSeek-R1-W8A8
# "--additional-config" is used to enable characteristics from vllm-ascend
vllm serve vllm-ascend/DeepSeek-R1-W8A8 \
--host 0.0.0.0 \
--port $6 \
--tensor-parallel-size 1 \
--enable-expert-parallel \
--seed 1024 \
--served-model-name deepseek_r1 \
--max-model-len 17000 \
--max-num-batched-tokens 256 \
--trust-remote-code \
--max-num-seqs 28 \
--gpu-memory-utilization 0.9 \
--quantization ascend \
--speculative-config '{"num_speculative_tokens": 1, "method":"deepseek_mtp"}' \
--kv-transfer-config \
'{"kv_connector": "MooncakeConnector",
"kv_buffer_device": "npu",
"kv_role": "kv_consumer",
"kv_parallel_size": "1",
"kv_port": "20001",
"engine_id": "0",
"kv_connector_module_path": "vllm_ascend.distributed.mooncake_connector"
}' \
--additional-config '{"enable_weight_nz_layout":true}'
```
::::
:::::
### Start Distributed DP Server for prefill-decode disaggregation
Execute the following Python file on all nodes to use the distributed DP server. (We recommend using this feature on the v0.9.1 official release)
:::::{tab-set}
::::{tab-item} Prefiller node
```python
import multiprocessing
import os
import sys
dp_size = 2 # total number of DP engines for decode/prefill
dp_size_local = 2 # number of DP engines on the current node
dp_rank_start = 0 # starting DP rank for the current node
# dp_ip is different on prefiller nodes in this example
dp_ip = "192.0.0.1" # master node ip for DP communication
dp_port = 13395 # port used for DP communication
engine_port = 9000 # starting port for all DP groups on the current node
template_path = "./run_dp_template.sh"
if not os.path.exists(template_path):
print(f"Template file {template_path} does not exist.")
sys.exit(1)
def run_command(dp_rank_local, dp_rank, engine_port_):
command = f"bash ./run_dp_template.sh {dp_size} {dp_ip} {dp_port} {dp_rank_local} {dp_rank} {engine_port_} {dp_size_local}"
os.system(command)
processes = []
for i in range(dp_size_local):
dp_rank = dp_rank_start + i
dp_rank_local = i
engine_port_ = engine_port + i
process = multiprocessing.Process(target=run_command, args=(dp_rank_local, dp_rank, engine_port_))
processes.append(process)
process.start()
for process in processes:
process.join()
```
::::
::::{tab-item} Decoder node
```python
import multiprocessing
import os
import sys
dp_size = 64 # total number of DP engines for decode/prefill
dp_size_local = 16 # number of DP engines on the current node
dp_rank_start = 0 # starting DP rank for the current node. e.g. 0/16/32/48
# dp_ip is the same on decoder nodes in this example
dp_ip = "192.0.0.5" # master node ip for DP communication.
dp_port = 13395 # port used for DP communication
engine_port = 9000 # starting port for all DP groups on the current node
template_path = "./run_dp_template.sh"
if not os.path.exists(template_path):
print(f"Template file {template_path} does not exist.")
sys.exit(1)
def run_command(dp_rank_local, dp_rank, engine_port_):
command = f"bash ./run_dp_template.sh {dp_size} {dp_ip} {dp_port} {dp_rank_local} {dp_rank} {engine_port_} {dp_size_local}"
os.system(command)
processes = []
for i in range(dp_size_local):
dp_rank = dp_rank_start + i
dp_rank_local = i
engine_port_ = engine_port + i
process = multiprocessing.Process(target=run_command, args=(dp_rank_local, dp_rank, engine_port_))
processes.append(process)
process.start()
for process in processes:
process.join()
```
::::
:::::
Note that the prefiller nodes and the decoder nodes may have different configurations. In this example, each prefiller node deployed as master node independently, but all decoder nodes take the first node as the master node. So it leads to difference in 'dp_size_local' and 'dp_rank_start'
## Example proxy for Distributed DP Server
In the PD separation scenario, we need a proxy to distribute requests. Execute the following commands to enable the example proxy:
```shell
python load_balance_proxy_server_example.py \
--port 8000 \
--host 0.0.0.0 \
--prefiller-hosts \
192.0.0.1 \
192.0.0.2 \
192.0.0.3 \
192.0.0.4 \
--prefiller-hosts-num \
2 2 2 2 \
--prefiller-ports \
9000 9000 9000 9000 \
--prefiller-ports-inc \
2 2 2 2\
--decoder-hosts \
192.0.0.5 \
192.0.0.6 \
192.0.0.7 \
192.0.0.8 \
--decoder-hosts-num \
16 16 16 16 \
--decoder-ports \
9000 9000 9000 9000 \
--decoder-ports-inc \
16 16 16 16 \
```
|Parameter | meaning |
| --- | --- |
| --port | Proxy service Port |
| --host | Proxy service Host IP|
| --prefiller-hosts | Hosts of prefiller nodes |
| --prefiller-hosts-num | Number of repetitions for prefiller node hosts |
| --prefiller-ports | Ports of prefiller nodes |
| --prefiller-ports-inc | Number of increments for prefiller node ports |
| --decoder-hosts | Hosts of decoder nodes |
| --decoder-hosts-num | Number of repetitions for decoder node hosts |
| --decoder-ports | Ports of decoder nodes |
| --decoder-ports-inc | Number of increments for decoder node ports |
You can get the proxy program in the repository's examples, [load\_balance\_proxy\_server\_example.py](https://github.com/vllm-project/vllm-ascend/blob/v0.9.1-dev/examples/disaggregate_prefill_v1/load_balance_proxy_server_example.py)
## Benchmark
We recommend use aisbench tool to assess performance. [aisbench](https://gitee.com/aisbench/benchmark) Execute the following commands to install aisbench
```shell
git clone https://gitee.com/aisbench/benchmark.git
cd benchmark/
pip3 install -e ./
```
You need to canncel the http proxy before assessing performance, as following
```shell
# unset proxy
unset http_proxy
unset https_proxy
```
- You can place your datasets in the dir: `benchmark/ais_bench/datasets`
- You can change the configurationin the dir :`benchmark/ais_bench/benchmark/configs/models/vllm_api` Take the ``vllm_api_stream_chat.py`` for examples
```python
models = [
dict(
attr="service",
type=VLLMCustomAPIChatStream,
abbr='vllm-api-stream-chat',
path="vllm-ascend/DeepSeek-R1-W8A8",
model="dsr1",
request_rate = 28,
retry = 2,
host_ip = "192.0.0.1", # Proxy service host IP
host_port = 8000, # Proxy service Port
max_out_len = 10,
batch_size=1536,
trust_remote_code=True,
generation_kwargs = dict(
temperature = 0,
seed = 1024,
ignore_eos=False,
)
)
]
```
- Take gsm8k dataset for example, execute the following commands to assess performance.
```shell
ais_bench --models vllm_api_stream_chat --datasets gsm8k_gen_0_shot_cot_str_perf --debug --mode perf
```
- For more details for commands and parameters for aisbench, refer to [aisbench](https://gitee.com/aisbench/benchmark)
## Prefill & Decode Configuration Details
In the PD separation scenario, we provide a optimized configuration.
- **prefiller node**
1. set HCCL_BUFFSIZE=256
2. add '--enforce-eager' command to 'vllm serve'
3. Take '--kv-transfer-config' as follow
```shell
--kv-transfer-config \
'{"kv_connector": "MooncakeConnector",
"kv_buffer_device": "npu",
"kv_role": "kv_producer",
"kv_parallel_size": "1",
"kv_port": "20001",
"engine_id": "0",
"kv_connector_module_path": "vllm_ascend.distributed.mooncake_connector"
}'
```
4. Take '--additional-config' as follow
```shell
--additional-config '{"enable_weight_nz_layout":true,"enable_prefill_optimizations":true}'
```
- **decoder node**
1. set HCCL_BUFFSIZE=1024
2. Take '--kv-transfer-config' as follow
```shell
--kv-transfer-config
'{"kv_connector": "MooncakeConnector",
"kv_buffer_device": "npu",
"kv_role": "kv_consumer",
"kv_parallel_size": "1",
"kv_port": "20001",
"engine_id": "0",
"kv_connector_module_path": "vllm_ascend.distributed.mooncake_connector"
}'
```
3. Take '--additional-config' as follow
```shell
--additional-config '{"enable_weight_nz_layout":true}'
```
### Parameters Description
1.'--additional-config' Parameter Introduction:
- **"enable_weight_nz_layout"：** Whether to convert quantized weights to NZ format to accelerate matrix multiplication.
- **"enable_prefill_optimizations"：** Whether to enable DeepSeek models' prefill optimizations.
<br>
3.enable MTP
Add the following command to your configurations.
```shell
--speculative-config '{"num_speculative_tokens": 1, "method":"deepseek_mtp"}'
```
### Recommended Configuration Example
For example，if the average input length is 3.5k, and the output length is 1.1k, the context length is 16k, the max length of the input dataset is 7K. In this scenario, we give a recommended configuration for distributed DP server with high EP. Here we use 4 nodes for prefill and 4 nodes for decode.
| node | DP | TP | EP | max-model-len | max-num-batched-tokens | max-num-seqs | gpu-memory-utilization |
|----------|----|----|----|---------------|------------------------|--------------|-----------|
| prefill | 2 | 8 | 16 | 17000 | 16384 | 4 | 0.9 |
| decode | 64 | 1 | 64 | 17000 | 256 | 28 | 0.9 |
:::{note}
Note that these configurations are not related to optimization. You need to adjust these parameters based on actual scenarios.
:::
## FAQ
### 1. Prefiller nodes need to warmup
Since the computation of some NPU operators requires several rounds of warm-up to achieve best performance, we recommend preheating the service with some requests before conducting performance tests to achieve the best end-to-end throughput.

3
docs/source/user_guide/support_matrix/supported_models.md
View File

@@ -63,9 +63,6 @@ Get the latest info here: https://github.com/vllm-project/vllm-ascend/issues/160
| Qwen3-VL-MOE | ✅ | |||||||||||||||||||
| Qwen2.5-Omni | ✅ ||||||||||||||||||| [Qwen2.5-Omni](../../tutorials/Qwen2.5-Omni.md) |
| QVQ | ✅ | |||||||||||||||||||
| LLaVA 1.5/1.6 | ✅ | [1962](https://github.com/vllm-project/vllm-ascend/issues/1962) |||||||||||||||||||
| InternVL2 | ✅ | |||||||||||||||||||
| InternVL2.5 | ✅ | |||||||||||||||||||
| Qwen2-Audio | ✅ | |||||||||||||||||||
| Aria | ✅ | |||||||||||||||||||
| LLaVA-Next | ✅ | |||||||||||||||||||