Add from_pretrained with HFHub

[matteobeltrami]

Still missing:

• risoluzione di input dentro al checkpoint name;
• numero di canali deve andare dentro datasets_info;
• stringhe con " " ;
• fix filename checkpoint;

@fpaissan please check dependencies.

[matteobeltrami]

Added:

• risoluzione di input dentro al checkpoint name;
• numero di canali dentro datasets_info;
• stringhe con " " ;
• fix filename checkpoint;

@fpaissan please check dependencies.

[fpaissan]

Maybe add a comment explaining why and where this is used in the code.

[fpaissan]

If we implement the comment on line 46, we no longer need this.

[fpaissan]

For this, we can directly put a string; it should raise AssertionError by default -- I forgot about this.
Something like:

assert (
    num_classes == phinet.datasets_info[dataset]["Nclasses"]
),  "messaggio di errore"

-- see here alse

[fpaissan]

@matteobeltrami we should also handle the WARNING and debug strings via python logging library instead of prints. This will help check for bugs further down in the project.

[matteobeltrami]

Added description to the datasets_info dictionary, modified the assertion, changed print("WARNING ...") with logging.warning("...").

@fpaissan please check dependencies

[matteobeltrami]

Added description to the datasets_info dictionary, modified the assertion, changed print("WARNING ...") with logging.warning("...").
Performed black and flake8 checks

@fpaissan please check dependencies

[matteobeltrami]

added "classifier" (initialized as True) to the from_pretrained() method, rewritten the model initialization instruction as follows: model.load_state_dict(state_dict["state_dict"], strict=False). Notice that it differs from the previous version only in the "strict" parameter that determines whether to strictly enforce that the keys in state_dict match the keys returned by the state_dict() function, as explained in the pytorch documentation: https://pytorch.org/docs/stable/generated/torch.nn.Module.html#torch.nn.Module.load_state_dict.

@fpaissan please check dependencies

[fpaissan]

LGTM!