Is there a way to suppress bert_load_from_file:***** messages in the terminal?

[DaveParr]

When calling GPT4All embeddings:

from langchain.embeddings import GPT4AllEmbeddings

GPT4AllEmbeddings()

The function sends this to the terminal every time:

bert_load_from_file: gguf version     = 2
bert_load_from_file: gguf alignment   = 32
bert_load_from_file: gguf data offset = 695552
bert_load_from_file: model name           = BERT
bert_load_from_file: model architecture   = bert
bert_load_from_file: model file type      = 1
bert_load_from_file: bert tokenizer vocab = 30522

Having traced a little way through the source code and run some experiments setting verbose=False has no impact on this output. I've also not succeeded at finding a way to intercept and suppress these messages.

Does anyone have a known technique or suggestion on how to do this?

[jeremiah-k]

I'm looking for a way to do this too.

[sbeer1]

ditto

[khabouss]

The message you see in the terminal is printed when the GPT4AllEmbeddings object is first constructed, you can construct it just once and use it instead of reconstructing it every time
The following example will print the message just once for all documents:

...
from langchain.embeddings.gpt4all import GPT4AllEmbeddings

embedding = GPT4AllEmbeddings()

ElasticsearchStore.from_documents(
        documents,
        embedding=embedding,
        index_name=index,
        es_url=es_link,
        )

The following example will print the message for every documents:

...
from langchain.embeddings.gpt4all import GPT4AllEmbeddings

ElasticsearchStore.from_documents(
        documents,
        embedding= GPT4AllEmbeddings(),
        index_name=index,
        es_url=es_link,
        )

[DaveParr]

Thanks for the suggestion. In my case though I'm already doing something similar that only prints the message once. I want the message 0 times.