lmms_eval/evaluator.py", line 83, in simple_evaluate assert tasks != [], "No tasks specified, or no tasks found. Please verify the task names." AssertionError: No tasks specified, or no tasks found. Please verify the task names.
lucasjinreal opened this issue · comments
lmms_eval/evaluator.py", line 83, in simple_evaluate
assert tasks != [], "No tasks specified, or no tasks found. Please verify the task names."
AssertionError: No tasks specified, or no tasks found. Please verify the task names.
I dont understand, why?
03-22 20:22:48 [lewisjin/mllm/llms_eval.py:240] INFO Available Tasks:
- ai2d
- chartqa
- cmmmu
- cmmmu_test
- cmmmu_val
- docvqa
- docvqa_test
- docvqa_val
- gqa
- hallusion_bench_image
- iconqa
- iconqa_test
- iconqa_val
- infovqa
- infovqa_test
- infovqa_val
- mmbench
- mmbench_cn
- mmbench_cn_cc
- mmbench_cn_dev
- mmbench_cn_test
- mmbench_en
- mmbench_en_dev
- mmbench_en_test
- mmmu
- mmmu_test
- mmmu_val
- mmvet
- multidocvqa
- multidocvqa_test
- multidocvqa_val
- ok_vqa
- ok_vqa_val2014
- pope
And how to specific mmbench_cn custom local dataset path?
Hi, looks like even set local envrioment variable, the script still try to read from hf remote, how to fix this issue?
lmms_eval/evaluator.py", line 83, in simple_evaluate assert tasks != [], "No tasks specified, or no tasks found. Please verify the task names." AssertionError: No tasks specified, or no tasks found. Please verify the task names.
have you solved this problem? i met the same when using a machine that doesn't have internet access...
I am move to VLMEvalKit.....
For me, I have a local huggingface dataset builder which can be loaded with load_dataset('/path/to/my/dataset/dataset.py') but throw this error when I put
dataset_path: /path/to/my/datasetin the yaml. Step through the debugger, I found the error lies in TaskConfig's post_init method in
api/task.py`.
Remove the line 99-103 solves the problem, which fail on import_module('/path/to/my/dataset/')
:
def __post_init__(self):
#if self.dataset_path and os.path.exists(os.path.dirname(self.dataset_path)):
# import inspect
# from importlib import import_module
# self.dataset_path = inspect.getfile(import_module(self.dataset_path))
Not sure what's the purpose of this code but thing runs fine if I remove it.