Minor fixes (#285)

fix(rwkv): load tokenizer file from model path (#255)

:arrow_up: Update go-skynet/go-llama.cpp (#245) Signed-off-by: GitHub <noreply@github.com> Co-authored-by: mudler <mudler@users.noreply.github.com>

Update gpt4all to fix thread counts (#249)

feat: bert.cpp token embeddings (#241)

feat: add starcoder (#236)

refactor: drop code dups (#234)

Keep whisper models in memory (#233)

feat: add bert.cpp embeddings (#222)

doc: update README with embeddings docs

Mixed enhancements (#196)

fix: missing returning error and free callback stream (#187)

fix: hardcode default number of cores to '4' (#186)

Return usage in the API responses (#166)

Update docs (#163)

fix(deps): update github.com/go-skynet/go-llama.cpp digest to 377fd24 (#129) Co-authored-by: renovate[bot] <29139614+renovate[bot]@users.noreply.github.com>

update: bump llama.cpp to 7f15c5c (#122) Signed-off-by: mudler <mudler@mocaccino.org>

few typos and clarity changes (#91) (#92) Co-authored-by: antongisli <anton@huge.geek.nz>

feat: Return OpenAI errors and update docs (#80) Signed-off-by: mudler <mudler@mocaccino.org>

fix: make MacOS builds work (#61)