6a125d36df
Add Qwen3.8 cloud API support with updated model lifecycles, multimodal routing, vision filtering, structured output handling, and regression tests.
kantan-kanto2026-08-22 14:21:14 +09:00
f10b8d195f
Add local Qwen3.8 GGUF support with Qwen35 handler routing and regression tests.
kantan-kanto2026-08-22 07:59:28 +09:00
0b065870a8
Prepare the project metadata, changelog, and documentation for the 1.0.15 release.
v1.0.15
kantan-kanto2026-07-31 00:58:21 +09:00
a2d8cc69bd
Improve cloud API retries with transient-error classification, interruptible backoff, and DashScope-safe Retry-After handling.
kantan-kanto2026-07-31 00:37:49 +09:00
7c471ac0f8
Preserve Wan API compatibility with shared automatic language detection.
kantan-kanto2026-07-31 00:03:32 +09:00
b12813d0ff
Format Qwen-Image quality suffixes with language-appropriate punctuation.
kantan-kanto2026-07-30 23:39:57 +09:00
a64b0a6f42
Resolve API keys relative to the actual custom node installation directory.
kantan-kanto2026-07-30 23:24:42 +09:00
51e7a50b50
Optimize GGUF model caching and handler initialization to reduce peak memory usage.
kantan-kanto2026-07-30 23:14:16 +09:00
14beb00d71
Improve local GGUF prompt enhancement with reliable output-language enforcement and correction.
kantan-kanto2026-07-30 22:19:07 +09:00
5ed1d44a85
Bump version to 1.0.14 and update release documentation for llama-cpp-python MTMD compatibility.
v1.0.14
kantan-kanto2026-06-16 19:40:12 +09:00
4fe7f02a05
Add compatibility for llama-cpp-python MTMD handlers that use either clip_model_path or mmproj_path.
kantan-kanto2026-06-16 19:10:56 +09:00
2c8b38210f
Refine README installation guidance for local llama-cpp-python setup
kantan-kanto2026-06-14 00:06:06 +09:00
8a8cc599a2
Bug Fix v1.0.8 - Fixed issue where Qwen2.5-VL were always loaded in text-only mode even when a valid mmproj file was specified. - Improved mmproj auto-detection logic.
v1.0.8
kantan-kanto2026-02-09 15:13:04 +09:00