4.1.7

Changelog from version 4.1.6 to 4.1.7:

f0a0af9 new version 4.1.7
efd8c97 fixing license comments
8275fe4 ifndef guard names fixed
eff180d new llama logs
60187e4 huggingface-hub upgraded to 0.27.0
44bbb71 Falcon3 example added
9c2ac6d llama.cpp updated
beb4a22 llama.cpp updated
741b01e vendro C++ standard set to 17
39db082 llama.cpp updated + new penalize sampling
8bf89d9 Update README.md
2b9a0df workflow names fixed

4.1.6

Changelog from version 4.1.5 to 4.1.6:

379e971 new version 4.1.6
b2e0997 close inactive issues workflow
a1db9de updating workflows to use permissions
3679ba1 llama.cpp updated + new kv_cache types
f977a38 llama.cpp updated

4.1.5

Changelog from version 4.1.4 to 4.1.5:

701dda1 new version 4.1.5
460bfaf license comments fixed
f32182d llama_params ifndef fixed
868c70c llama.cpp updated
efad6bc Mistral model added
59957a0 Hermes model updated
edd93d3 llama.cpp updated
1106ec7 cron added to docker build workflow

4.1.4

Changelog from version 4.1.3 to 4.1.4:

893ae80 new version 4.1.4
7374334 fixing logs
be3ad64 llama.cpp updated
39f5283 llama.cpp updated + ggml_amx removed

4.1.3

Changelog from version 4.1.2 to 4.1.3:

5212102 new version 4.1.3
a729e44 new workflow to create releases
d0294ea removig wrong option for CURL
0fe12b0 llama.cpp updated
c86501b debug param removed
4ca5739 llama.cpp updated
3e770cf devices param added
879a1ab llama.cpp updated + new params sampling

4.1.2

• Improving includes order
• Remove bos from llama-3 prompt
• llama.cpp b4157

4.1.1

• minor fixes to langchain integration
• find_stop_word fixed by trimming string
• llama.cpp b4154

4.1.0

• Getting metadata from GGUF model through llama.cpp
• New metadata msgs (Metadata, GeneralInfo, ModelInfo, AttentionInfo, RoPEInfo, TokenizerInfo)
• New service to get the metadata of the LLM/VLM
• llama.cpp b4149

4.0.11

• defrag_thold param added
• target properties for ggml-base, ggml-cpu, ggml-amx, ggml-cuda, ggml-cuda
• minor fix in README
• llama.cpp b4112

4.0.10

• new github actions (docker push, doxygen)
• llama.cpp b4050