Olivier Chafik
Olivier Chafik
### Description of the bug: The only place w/ actual Python classes / methods docs seems to be in the repo under [docs/api/google/generativeai.md](https://github.com/google-gemini/generative-ai-python/blob/main/docs/api/google/generativeai.md) However, there's a prominent [ai.google.dev/gemini-api/docs](https://ai.google.dev/gemini-api/docs) link which...
This is already deployed in https://ochafik.com/openscad2/ but there's a fair bit of work left before merging. New features - Color support in preview & render, including green faces for result...
This moves Manifold (and its color support) out of experimental (https://github.com/openscad/openscad/issues/4825) Command line usage: ```bash echo '$fn=1000; cube(); color("red") sphere();" | \ openscad - -o out.off --backend=manifold ``` cc/ @kintel
While *casually* reading the Emscripten manual (esp. [emcc options](https://emscripten.org/docs/tools_reference/emcc.html) & [Emscripten Compiler Settings](https://emscripten.org/docs/tools_reference/settings_reference.html)), I stumbled across [-sNODERAWFS](https://emscripten.org/docs/tools_reference/settings_reference.html#noderawfs), which gives JS access to the filesystem. So... this PR adds a new...
color channels < 1.0f were zeroed out (first cast to int, then * 255).
Speeds up (re)builds with [ccache](https://ccache.dev/), if installed / detected (can disable w/ `-DASSIMP_BUILD_USE_CCACHE=0`) ```bash rm -fR build # First build takes a while: cmake -B build -DCMAKE_BUILD_TYPE=Debug && cmake --build...
This supersedes https://github.com/ggerganov/llama.cpp/pull/6389 (now using a fully C++ approach), https://github.com/ggerganov/llama.cpp/pull/5695 (first attempt at supporting Functionary) and https://github.com/ggerganov/llama.cpp/pull/9592 (more recent Python wrapper). ## Background It tackles two main problems related to...
WIP: server tests still not green Fixes https://github.com/ggerganov/llama.cpp/issues/9273 (and supersedes https://github.com/ggerganov/llama.cpp/pull/6941 ) This makes the non-streamed handling of completions to use a Sink-based API ([set_content_provider](https://github.com/yhirose/cpp-httplib?tab=readme-ov-file#send-content-with-the-content-provider)) similar to the streaming path,...
This is to be able to detect if the client connection was lost. Context: https://github.com/yhirose/cpp-httplib/issues/1952 / https://github.com/ggerganov/llama.cpp/pull/9679 As suggested by @ngxson, since is_writable / select_write may be a bit heavyweight,...
### Name and Version latest ### Operating systems _No response_ ### Which llama.cpp modules do you know to be affected? libllama (core library) ### Command line ```shell llama-cli -hf bartowski/DeepSeek-R1-Distill-Llama-8B-GGUF:Q4_K_M...