TypeError: data type not understood about sequence_tagging

[guillaumegenthial]

Hi @Jonida88!
Check your numpy version but it might be the case that for some reasons it does not understand the shape given as a list, change it to a tuple like (len(vocab), dim) instead of [len(vocab), dim].
Or check that len(vocab) and dim are integers.

[Jonida88]

Hi @guillaumegenthial glad that you was answering me i was able to solve it but now i have an other issue i am runing the build_data without error so this mean:

> Building vocab...
-done.33970tokens
Building vocab...
-done.400000 tokens
Writing vocab..
-done.7398 tokens
Writing vocab..
-done.28227 tokens
Writing vocab..
-done.85 tokens

but when i open the file I have this issue:

> Error! C:\Users\admin\ML\data\glove.6B.300d.trimmed.npz is not UTF-8 encoded
> Saving disabled.
> See Console for more details.

the code
`np.savez_compressed(filename_trimmed, embeddings=embeddings)

def get_trimmed_glove_vectors(filename_trimmed):
try:
with np.load(filename_trimmed) as data:
return data["embeddings"]
except IOError:
raise MyIOError(filename_trimmed)`

any idea... thanks for your help...

[DimejiOladepo]

I also had this issue. I upgraded my Tensorflow to 2.0.0 and numpy to 1.16.4.

For your new problem, consider saving your glove embedding as a UTF-8 txt file. Opening it in this format doesn't give issues.

I hope this helps.