[feature] Enhanced json workflow saving

### Feature Idea

In addition to the usual json data saved, I would like the json 1) to save the size (information) of the models used/involved in current workflow and 2) even check through internet and hf known repos to compare the name of such model and confirm or find the most likely model it is in case the name was changed by the user.

Reasoning:
- In rare cases: different models get the same names
- More frequent: users save the models with custom nodes or add prefixes or suffixes
- Sometimes models don't have names (xxx_pytorch. something), every user has a different name for them, but at least the parent folder of such model has a somewhat specific name
- Sometimes the only way to confirm a model CORRESPOND TO A MODEL AI SEARCHING FOR IS: to compare the size of model in hg and model in my local disk (or in the case of an enhanced json it would be to see what the json says and compare with repos and hf download links and confirm)
- Sometimes models can get confusing if we are in NATIVE vs WRAPPER (I am pretty sure we have a case where we had the same name for 2 different models native vs wrapper)

Solution 1 (without internet):
enhanced prompt would offer informations about the size of current models.
Combined with name of current model used (name possibly modified by user sharing his prompt or by ourself in previous machine while re using our own workflows) + its size -> we can search and find the correct model in HF (for example).
We avoid:
- not finding the right model
- doubting the model we found, since there are so many with the same name or iterations.

Solution 2 (with internet acess):
Solution 1 + the following:
- Perhaps though use of HF token or without it or similar, allow a portion of the process to go perform a research in hf places to find the model, and use an algo to find most proximate name models perhaps just prefix/suffix change.
- or Make a registery of known models and their paths and their sizes, that comfy can compare to it to find if the currrent custom named model correpond to any of them.

Thanks.
works well with another feature I am suggesting here (https://github.com/Comfy-Org/ComfyUI_frontend/issues/6142) where we can auto fill a section of used models in the current workflow.

### Existing Solutions

_No response_

### Other

_No response_