{
"input": [
{
"id": "732c90a4",
"content": [
{
"text": "Please recognize the speech and only output the recognized content:"
},
{
"audio": "[BASE64_AUDIO: wav, ~22.7MB]",
"format": "wav"
}
]
}
],
"target": "Eleven o'clock had struck it was a fine clear night they were the only persons on the road and they sauntered leisurely along to avoid paying the price of fatigue for the recreation provided for the toledans in their valley or on the banks of ... [TRUNCATED 7380 chars] ... less surprised than they and the better to assure himself of so wonderful a fact he begged leocadia to give him some token which should make perfectly clear to him that which indeed he did not doubt since it was authenticated by his parents.",
"id": 0,
"group_id": 0,
"metadata": {
"audio_id": "5639-40744",
"audio_duration": 496.6899719238281
}
}
注:部分内容因显示需要被截断。
提示模板
提示模板:
Please recognize the speech and only output the recognized content:
使用方法
使用 CLI
evalscope eval\
--model YOUR_MODEL \
--api-url OPENAI_API_COMPAT_URL \
--api-key EMPTY_TOKEN \
--datasets librispeech \
--limit 10 # Remove this line for formal evaluation
使用 Python
fromevalscopeimportrun_taskfromevalscope.configimportTaskConfigtask_cfg = TaskConfig(
model='YOUR_MODEL',
api_url='OPENAI_API_COMPAT_URL',
api_key='EMPTY_TOKEN',
datasets=['librispeech'],
dataset_args={
'librispeech': {
# subset_list: ['test_clean', 'test_other'] # optional, evaluate specific subsets
}
},
limit=10, # Remove this line for formal evaluation
)
run_task(task_cfg=task_cfg)