Selected Model is Unsupported

[xavierallem]

Using Alpaca and gpt4all

Have the following

• ggml-alpaca-7b-q4.bin
• gpt4all-lora-quantized.bin

After selecting the path it shows Selected model is of an unsupported version

[alexrozanski]

Hey @xavierallem, sorry - the docs and UI aren't super clear on this yet. Looks like you have one of the slightly older model versions. I'm going to add native flows for these cases, but for now you can use:

• convert-unversioned-ggml-to-ggml.py for the Alpaca model
• convert-gpt4all-to-ggml.py for the GPT4All model

from the llama.cpp repo to upgrade the models to the new format required by llama.cpp (which LlamaChat uses internally). Lmk if this works!

[alexrozanski]

added some details to the README for now

[danieldunderfelt]

The instructions are quite obtuse. It took me a while to figure out that the tokenizer.model is found among the Llama files.

[alexrozanski]

@danieldunderfelt good feedback. I added some more notes in https://github.com/alexrozanski/LlamaChat/blob/main/README.md#-models, but this should really be better supported/documented in the conversion flow

[danieldunderfelt]

@danieldunderfelt good feedback. I added some more notes in https://github.com/alexrozanski/LlamaChat/blob/main/README.md#-models, but this should really be better supported/documented in the conversion flow

Epic, thanks!

[alexrozanski]

This is fixed in v1.2.0 -- older models are now supported directly in LlamaChat.