[Meta-issue] Reported models with broken jinja templates

[ThiloteE]

Info

Starting with GPT4All 3.5.0, compatibility with Jinja templates that can be found in the large language models accompanying tokenizer_config.json has been introduced. Since GPT4All's jinja parser is c++ based and some model authors create their templates with a python based jinja tooling, there are some compatibility issues left to be addressed. In this issue we collect reports of model chat templates that do not (yet) work with GPT4All.

If you want your models chat template to be added to this list, please create a separate github isse and then link to this issue using Markdown syntax.

Chat template issues raised on Discord:

• MaziyarPanahi/calme-2.1-phi3.5-4b-GGUF (fixed) link
• QuantFactory/Hermes-3-Llama-3.2-3B-GGUF (already fixed by openorca substitution): link