chatllm: fix loading of chats after #2676 #2693
Merged
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
After #2676, attempting to load old chats crashes GPT4All. This is because the values of LLModelType, which are serialized to disk, have changed. The new builds of GPT4All misinterpret the old values of LLModelType and read bad data (Llama is loaded as ChatAPI).
This PR adds back the old enumeration values and tries to make it as clear as possible that they are not arbitrary and should not be changed.
Serialization for the GPT-J model type is also restored so GPT4All does not hit Q_UNREACHABLE on exit if there is an existing chat with one of these models.