|
|
@ -24,8 +24,11 @@ Most people do not have such a powerful computer or access to GPU hardware. By r
|
|
|
|
GPT4All models can run on your laptop using only 4-8GB of RAM enabling their wide-spread utility.
|
|
|
|
GPT4All models can run on your laptop using only 4-8GB of RAM enabling their wide-spread utility.
|
|
|
|
|
|
|
|
|
|
|
|
The GPT4All software ecosystem is currently compatible with three variants of the Transformer neural network architecture:
|
|
|
|
The GPT4All software ecosystem is currently compatible with three variants of the Transformer neural network architecture:
|
|
|
|
|
|
|
|
|
|
|
|
- LLaMa
|
|
|
|
- LLaMa
|
|
|
|
|
|
|
|
|
|
|
|
- GPT-J
|
|
|
|
- GPT-J
|
|
|
|
|
|
|
|
|
|
|
|
- MPT
|
|
|
|
- MPT
|
|
|
|
|
|
|
|
|
|
|
|
Any model trained with one of these architectures can be quantized and run locally with all GPT4All bindings and in the
|
|
|
|
Any model trained with one of these architectures can be quantized and run locally with all GPT4All bindings and in the
|
|
|
|