So I have a problem that I can’t pass HUGGINGFACEHUB_API_TOKEN as an environmental variable to MLFlow logging. For OpenAI, MLFlow has “MLFLOW_OPENAI_SECRET_SCOPE” environmental variable which stores token value. But is there anything similar for Huggingface models since API token is mandatory? I can log models with using envinronmental token variables but when calling model and trying to use it for prediction, getting the next error message all the time: “Did not find huggingfacehub_api_token, please add an environment variable HUGGINGFACEHUB_API_TOKEN which contains it, or pass huggingfacehub_api_token as a named parameter. (type=value_error)'” → haven’t found a way how to pass that token value to prediction part at all…
1 Like