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>
20 lines
426 B
YAML
20 lines
426 B
YAML
model_name: "Qwen/Qwen3-30B-A3B"
|
|
hardware: "Atlas A2 Series"
|
|
tasks:
|
|
- name: "gsm8k"
|
|
metrics:
|
|
- name: "exact_match,strict-match"
|
|
value: 0.89
|
|
- name: "exact_match,flexible-extract"
|
|
value: 0.85
|
|
- name: "ceval-valid"
|
|
metrics:
|
|
- name: "acc,none"
|
|
value: 0.84
|
|
num_fewshot: 5
|
|
gpu_memory_utilization: 0.6
|
|
enable_expert_parallel: True
|
|
tensor_parallel_size: 2
|
|
apply_chat_template: False
|
|
fewshot_as_multiturn: False
|