Fasttext: Exception: fastText: Cannot load asttext_wiki.bin due to C++ extension failed to allocate the memory

[omerarshad]

i have 8 gb free ram and when i try to load model of 3 GB, i get this error. Can anyone tell how to load pre trained vector in fasttext?

[EdouardGrave]

Hi @omerarshad,

The binary model is probably too large to fit in memory. We will maybe release smaller models in the future. Have you tried using the text model? Since words are sorted by frequency, it is possible to load the first k words only.

Best

[tarrekshaban]

I am having the same issue. I do not think it is memory as my computer has over 256 gb of memory. Not quiet sure what to do to solve the issue.

[gabrer]

I have got the same issue only using the Python package "fasttext".
Instead, the C++ function "load_model" works properly.

[stiebels]

Same issue with 32gb memory in Python package 'fasttext'.

[ArgentumBlack]

Same issue with the binary model， but the text model worked.

[kelly-tlz]

I have the same question, whatever model I load. @gabrer how to tackle this problem with c++

[Ellibeere]

I have the same issue with 8gb and trying to use wiki.ady.vec text model.

I also tried only loading the first few hundreds lines, didn't work either.

[lxw0109]

Same problem. Anybody help? Thank you ~

[omerarshad]

shift to ubuntu. it works there :/

[lxw0109]

Thank you @omerarshad , but actually I'm working on Ubuntu 16.04 64bits. And it did not work, which is really weird. So you have managed to make it work on Ubuntu ?