⚡ Edgen: Local, private GenAI server alternative to OpenAI. No GPU required. Run AI models locally: LLMs (Llama2, Mistral, Mixtral...), Speech-to-text (whisper) and many others.
Client applications want or even need to know if the current default model (from the config) is present or not, i.e. if a download will be necessary when the model is used. Currently, client applications abuse the progress field from the status. But that is not reliable.
Solution
Provide a flag in the status "model_present". This flag is set when edgen starts, when the config changes and after download. To test if the model is present, the hf-hub client should be used.
Problem
Client applications want or even need to know if the current default model (from the config) is present or not, i.e. if a download will be necessary when the model is used. Currently, client applications abuse the progress field from the status. But that is not reliable.
Solution
Provide a flag in the status "model_present". This flag is set when edgen starts, when the config changes and after download. To test if the model is present, the hf-hub client should be used.