epic: Cortex supports preset configuration list

[Van-QA]

Problem: Users are unable to use imported GGUFs effectively as they all use default templates and parameters, requiring individual editing of each model's model.json file. https://discord.com/channels/1107178041848909847/1244886808979046452

Success Criteria:

• [x] 1. Enable automatic updating of the model.yaml file when users change model parameters during runtime. Model.yaml files store model configurations. If users modify these configurations during runtime, the files should be updated accordingly.

• [x] https://github.com/janhq/cortex/issues/616 2. Ensure each model has its own model.yaml file, located in the same folder as the GGUF binary. Similar to the current model's folder structure

• [x] janhq/jan#2960 3. Attempt to extract metadata from a GGUF file.

• [ ] 4. Implement a list of popular template configurations for any models that users import. janhq/jan#2958 Missing configurations will load with fallback presets. The configurations list refers to a pre-populated list and local cache of model + thread settings with memorable names, allowing users to choose from during the import process.

• [x] 5. Allow users to add their own custom templates. Users should have the ability to add their own custom templates, maybe r‌‌eusing from available model that they configur‌ated

[louis-jan]

For number 2, there is no need to put the YAML file under the model folder, since symlink support is available; all YAML files are located at /models.

[louis-jan]

Do you want to take the 4 @Van-QA?

[Van-QA]

Do you want to take the 4 @Van-QA?

I will work with @hahuyhoang411 on this, thank @louis-jan