GPT4All does not show Q5_K_M model in model discovery about gpt4all HOT 5 CLOSED

I created this process to add ollama models to GPT4All:
https://github.com/ll3N1GmAll/AI_GPT4All_Ollama_Models

from gpt4all.

But when I search for the model ID, it is not recognized.

The main problem is that GPT4All currently ignores models on HF that are not in Q4_0, Q4_1, FP16, or FP32 format, as those are the only model types supported by our GPU backend that is used on Windows and Linux.

from gpt4all.

@cebtenzzre That seems very limited. So no k-quantization support or even bf16?

from gpt4all.

That seems very limited. So no k-quantization support or even bf16?

You can run a Q5_K_M with GPT4All, just not on GPU.

from gpt4all.

This situation is briefly referenced in the README here. If you'd like to improve the README, we would gladly accept a PR.

Duplicate of #1551

from gpt4all.