ggml/llama.cpp releases: b11057

TL;DR

The ggml/llama.cpp project added a dedicated Ling 3.0 (Bailing V3) parser to improve chat functionality by pre-opening think blocks in generation prompts, allowing for more seamless tool calls without premature

Detailed Summary

The ggml/llama.cpp project has released version b11057, which includes a new dedicated parser for Ling 3.0 (Bailing V3). This update pre-opens the think block in generation prompts, ensuring that the model does not emit an opening closing tags. The change impacts how chat interactions are structured within the model's responses, potentially enhancing the flow and coherence of generated dialogue.

Key Points

• • chat: add dedicated Ling 3.0 (Bailing V3) parser
• • Ling 3.0 Flash templates pre-open the think block in generation prompt
• • Model never emits an opening