Got stuck when processing code r1 data

[ck853178967]

stuck when running python examples/data_preprocess/coder1.py

        actualchildren[k] += 1
        if actualchildren[k] == children[k]:
                stack.append(k)
ans = []
for i in range(m):
        if actualused[i]:
                ans.append(str(i + 1))
print(len(ans))
print(' '.join(ans))

exec_succ = False
stdin = '1 0\n1\n' stdout = '-1\n'
output = 
 Failed to execute program: STDOUT:


STDERR:
Traceback (most recent call last):
  File "<string>", line 20, in <module>
ValueError: not enough values to unpack (expected 2, got 0)
Test code failed for codeforces
Map (num_proc=64):  98%|████████████████████████████████████████████████████████████████████████████████▋ | 12686/12898 [05:29<00:06, 32.51 examples/s]

Each time stuck at around the same position, if I clean the data cache at /root/.cache/huggingface/datasets/likaixin___taco-verified.

Otherwise it will stuck at around the 190th sample.

Is there any advice?

[FlowRays]

Why not use the processed dataset directly from https://huggingface.co/datasets/ganler/code-r1-12k?

If you want to test locally and the code_exec method here not work for your environment, you can try DeepCoder's code execution.

For the LeetCode dataset, you could try my locally verified version: https://huggingface.co/datasets/Raywithyou/LeetCodeDataset_verl