A python tool for evaluating the quality of sentence embeddings.
Other
2.09k
stars
309
forks
source link
ValueError: setting an array element with a sequence. The requested array has an inhomogeneous shape after 1 dimensions. The detected shape was (750,) + inhomogeneous part. #94
sent1 = np.array([s.split() for s in sent1])[not_empty_idx]
In this line, the element of [s.split() for s in sent1] is not the same len. It will report this error. Has anyone met the same one?
I change the code to
`sent1_lengths = [len(s.split()) for s in sent1]
sent2_lengths = [len(s.split()) for s in sent2]
max_length = max(max(sent1_lengths), max(sent2_lengths))
sent1 = np.array([s.split() + [''] (max_length - len(s.split())) for s in sent1])[not_empty_idx]
sent2 = np.array([s.split() + [''] (max_length - len(s.split())) for s in sent2])[not_empty_idx]
`
and then the code go right.
taofendenanhai
commented
1 year ago
sent1 = np.array([s.split() for s in sent1])[not_empty_idx]In this line, the element of [s.split() for s in sent1] is not the same len. It will report this error. Has anyone met the same one? I change the code to `sent1_lengths = [len(s.split()) for s in sent1] sent2_lengths = [len(s.split()) for s in sent2] max_length = max(max(sent1_lengths), max(sent2_lengths))sent1 = np.array([s.split() + [''] (max_length - len(s.split())) for s in sent1])[not_empty_idx] sent2 = np.array([s.split() + [''] (max_length - len(s.split())) for s in sent2])[not_empty_idx] ` and then the code go right.