v0.3.14-cu128-AVX2-win-20250806

1. Sync llama.cpp API 20250806
2. Update vendor/llama.cpp to 0d8831543cdc368fb248bae6f1b4aa5516684edc
3. Add strftime_now function into Jinja2ChatFormatter Class for gpt-oss and gemma3 chat_template need

v0.3.14-cu128-AVX2-linux-20250806

1. Sync llama.cpp API 20250806
2. Update vendor/llama.cpp to 0d8831543cdc368fb248bae6f1b4aa5516684edc
3. Add strftime_now function into Jinja2ChatFormatter Class for gpt-oss and gemma3 chat_template need

v0.3.14-cu126-AVX2-win-20250806

1. Sync llama.cpp API 20250806
2. Update vendor/llama.cpp to 0d8831543cdc368fb248bae6f1b4aa5516684edc
3. Add strftime_now function into Jinja2ChatFormatter Class for gpt-oss and gemma3 chat_template need

v0.3.14-cu126-AVX2-linux-20250806

1. Sync llama.cpp API 20250806
2. Update vendor/llama.cpp to 0d8831543cdc368fb248bae6f1b4aa5516684edc
3. Add strftime_now function into Jinja2ChatFormatter Class for gpt-oss and gemma3 chat_template need

v0.3.14-cu124-AVX2-win-20250806

1. Sync llama.cpp API 20250806
2. Update vendor/llama.cpp to 0d8831543cdc368fb248bae6f1b4aa5516684edc
3. Add strftime_now function into Jinja2ChatFormatter Class for gpt-oss and gemma3 chat_template need

v0.3.14-cu124-AVX2-linux-20250806

1. Sync llama.cpp API 20250806
2. Update vendor/llama.cpp to 0d8831543cdc368fb248bae6f1b4aa5516684edc
3. Add strftime_now function into Jinja2ChatFormatter Class for gpt-oss and gemma3 chat_template need

v0.3.14-cu128-AVX2-win-20250801

1. Compile with cuda12.8.1 for Blackwell architecture (sm_100 and sm_120) computing cards
2. Sync llama.cpp API 20250801
3. Remove sm_70 in cuda12.8.1 action flow

v0.3.14-cu128-AVX2-linux-20250801

1. Compile with cuda12.8.1 for Blackwell architecture (sm_100 and sm_120) computing cards
2. Sync llama.cpp API 20250801
3. Remove sm_70 in cuda12.8.1 action flow

v0.3.14-cu126-AVX2-win-20250801

Sync llama.cpp API 20250801

v0.3.14-cu126-AVX2-linux-20250801

Sync llama.cpp API 20250801