GitLab

Separate GGUF and PyTorch guides

bump submodules

pin to 9e70cc03229df19ca2d28ce23cc817198f897278 for now since
438c2ca83045a00ef244093d27e9ed41a8cb4ea9 is breaking

ggufv3

update default log target

ggufv3 adds support for big endianness, mainly for s390x architecture.
while that's not currently supported for ollama, the change is simple.

loosen version check to be more forward compatible. unless specified,
gguf versions other v1 will be decoded into v2.

added python rag news summary

fix: nil pointer dereference

fix: ollama host for hostname

Co-authored-by: Jeffrey Morgan <jmorganca@gmail.com>

image: show parameters

basic request validation

move the empty response so it's more visible

unused after removing EMBED

add error for falcon and starcoder vocab compatibility
---------
Co-authored-by: Bruce MacDonald <brucewmacdonald@gmail.com>

- only reload the running llm if the model has changed, or the options for loading the running model have changed
- rename loaded llm to runner to differentiate from loaded model image
- remove logic which keeps the first system prompt in the generation context

Fix Issue with Leading Whitespaces in Decoded Context

model: native gotemplate adapter template