dc1a6cb503
### What this PR does / why we need it?
Add accuracy test for multiple models：
- Meta_Llama_3.1_8B_Instruct
- Qwen2.5-Omni-7B
- Qwen3-VL-8B-Instruct
- vLLM version: v0.11.0
- vLLM main:
83f478bb19
---------
Signed-off-by: MrZ20 <2609716663@qq.com>
17 lines
384 B
YAML
17 lines
384 B
YAML
model_name: "deepseek-ai/DeepSeek-V2-Lite"
|
|
hardware: "Atlas A2 Series"
|
|
tasks:
|
|
- name: "gsm8k"
|
|
metrics:
|
|
- name: "exact_match,strict-match"
|
|
value: 0.385
|
|
- name: "exact_match,flexible-extract"
|
|
value: 0.385
|
|
tensor_parallel_size: 2
|
|
batch_size: 32
|
|
gpu_memory_utilization: 0.7
|
|
apply_chat_template: False
|
|
fewshot_as_multiturn: False
|
|
trust_remote_code: True
|
|
enforce_eager: False
|