更新 build-wheels-cu128-win.yml

更新 build-wheels-cu128-win.yml

更新 build-wheels-cu128-win.yml

Update build-wheels-cu128-win.yml

Update llama.cpp 20251115 and Move the ggml-related code to _ggml.py.

Use httplib to download model from an URL when the libcurl is disabled

Note: LLAMA_HTTPLIB is OFF, cannot build llama-server. Hint: to skip building server, set -DLLAMA_BUILD_SERVER=OFF

Use httplib to download model from an URL when the libcurl is disabled

Note: LLAMA_HTTPLIB is OFF, cannot build llama-server. Hint: to skip building server, set -DLLAMA_BUILD_SERVER=OFF

Use httplib to download model from an URL when the libcurl is disabled

Note: LLAMA_HTTPLIB is OFF, cannot build llama-server. Hint: to skip building server, set -DLLAMA_BUILD_SERVER=OFF

Use httplib to download model from an URL when the libcurl is disabled

Note: LLAMA_HTTPLIB is OFF, cannot build llama-server. Hint: to skip building server, set -DLLAMA_BUILD_SERVER=OFF

Use httplib to download model from an URL when the libcurl is disabled

Note: LLAMA_HTTPLIB is OFF, cannot build llama-server. Hint: to skip building server, set -DLLAMA_BUILD_SERVER=OFF