64 lines
1.6 KiB
YAML
64 lines
1.6 KiB
YAML
|
# ==========================================
|
||
|
|
# ACTUAL TEST CASES
|
||
|
|
# ==========================================
|
||
|
|
|
||
|
|
test_cases:
|
||
|
|
- name: "Qwen2.5-VL-32B-Instruct-a3"
|
||
|
|
model: "Qwen/Qwen2.5-VL-32B-Instruct"
|
||
|
|
envs:
|
||
|
|
TASK_QUEUE_ENABLE: "1"
|
||
|
|
VLLM_ASCEND_ENABLE_NZ: "0"
|
||
|
|
HCCL_OP_EXPANSION_MODE: "AIV"
|
||
|
|
SERVER_PORT: "DEFAULT_PORT"
|
||
|
|
server_cmd:
|
||
|
|
- "--no-enable-prefix-caching"
|
||
|
|
- "--mm-processor-cache-gb"
|
||
|
|
- "0"
|
||
|
|
- "--tensor-parallel-size"
|
||
|
|
- "4"
|
||
|
|
- "--port"
|
||
|
|
- "$SERVER_PORT"
|
||
|
|
- "--max-model-len"
|
||
|
|
- "30000"
|
||
|
|
- "--max-num-batched-tokens"
|
||
|
|
- "40000"
|
||
|
|
- "--max-num-seqs"
|
||
|
|
- "400"
|
||
|
|
- "--trust-remote-code"
|
||
|
|
- "--gpu-memory-utilization"
|
||
|
|
- "0.8"
|
||
|
|
- "--compilation_config"
|
||
|
|
- '{"cudagraph_mode": "FULL_DECODE_ONLY"}'
|
||
|
|
test_content:
|
||
|
|
- "completion"
|
||
|
|
- "image"
|
||
|
|
benchmarks:
|
||
|
|
acc:
|
||
|
|
case_type: accuracy
|
||
|
|
dataset_path: vllm-ascend/textvqa-lite
|
||
|
|
request_conf: vllm_api_stream_chat
|
||
|
|
dataset_conf: textvqa/textvqa_gen_base64
|
||
|
|
max_out_len: 2048
|
||
|
|
batch_size: 128
|
||
|
|
baseline: 76.22
|
||
|
|
temperature: 0
|
||
|
|
top_k: -1
|
||
|
|
top_p: 1
|
||
|
|
repetition_penalty: 1
|
||
|
|
threshold: 5
|
||
|
|
perf:
|
||
|
|
case_type: performance
|
||
|
|
dataset_path: vllm-ascend/textvqa-perf-1080p
|
||
|
|
request_conf: vllm_api_stream_chat
|
||
|
|
dataset_conf: textvqa/textvqa_gen_base64
|
||
|
|
num_prompts: 512
|
||
|
|
max_out_len: 256
|
||
|
|
batch_size: 128
|
||
|
|
temperature: 0
|
||
|
|
top_k: -1
|
||
|
|
top_p: 1
|
||
|
|
repetition_penalty: 1
|
||
|
|
request_rate: 0
|
||
|
|
baseline: 1
|
||
|
|
threshold: 0.97
|