Yury

Just to clarify a bit: 1. The current logic always aligns the resolution by 32px, which is not ideal and most of the time produces wrong resolution (i.e., 3664x1920 becomes...

What do you mean exactly? In `start_http.bat`, you can change `--sample_mode "production"` to `--sample_mode ""`, and this will disable samples download and make it work offline. For fresh installation, you'd...

@epietrocola you have to get and upload your models into blank slots. If you want to use the models that are downloaded with default sample mode, you can copy `model_dir`...

> [WARN] args.model.use_harmonic_emb is None; use default False > [WARN] > call by:torchfcpe.tools.spawn_cf_naive_mel_pe > [WARN] no cuda_visible device found; fallback to cpu Looks like fcpe does not work correctly and...

`Extra` may cause very high CPU usage. You can try lower values and it should consume much less CPU

Wait, I just noticed that you use `Index`. Don't use it for real time, set it to 0. It also requires a lot of CPU since it's computed only on...

That still doesn't sound good enough, I was expecting a bigger impact... Could you consider trying the fork version I linked? It could be that I've already addressed this issue...

I'd still recommend trying a forked version I linked. You can check out the videos that demonstrate the performance on AMD and Nvidia (and no high CPU utilization on either...

In general, the current implementation must not be exposed to the internet (only localhost or VPN/local network) since it does not implement any access checks. And yes, model loading is...

> What do you mean by "make the requests directly"? You can simply visit the pages served by the backend without crafting a web page (i.e., by just navigating to...