Skip to content

Running out of memory when loading pre-trained embeddings #114

Open
@mortenterhart

Description

@mortenterhart

I'm running a laptop with an RTX 3070 GPU, an AMD Ryzen 9 5900HX processor and 16 GB RAM on Linux Mint. When trying to load the transe_wikidata5m.pkl model as following

with open('./transe_wikidata5m.pkl', 'rb') as embeddings_file:
    transe_embedding = pickle.load(embeddings_file)

I'm running out of memory and the process gets killed by the OS after some freezing time. I don't have access to another machine at the moment. Is there any alternative way of loading the model or are there smaller versions available?

I would be grateful for any help.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions