common : add common_speculative_is_compat()

[ggerganov]

fix #19267

Memory modules that do not support removing the last tokens from the context (such as recurrent modules) cannot perform speculative decoding. Add new common_speculative_is_compat() to query this functionality and use it in llama-server to disable speculative decoding for those contexts.

[ngxson]

Just wondering if we can do this inside common_speculative_init instead.

For example, common_speculative_init can try to evaluate 2 tokens, then remove the first one. If llama_memory_seq_rm returns error, then we throw an error saying the model is not compatible.

Btw, I think it's better to throw an error and exit, rather than a warning.

[coder543]

I just ran into #19267, and it would be cool if there were a way to make this compatible rather than just disabling it, but disabling it is better than crashing. With Qwen3-Coder-Next, ngram-mod could provide large speedups during coding workflows.

[ggerganov]

@ngxson Implemented this idea in a new common_speculative_is_compat() helper function.

Btw, I think it's better to throw an error and exit, rather than a warning.

Do you have something specific in mind? In my server config, I want to set a default ngram-based spec decoding and have it applied for all routed models. When a routed model does not support it, it still continues to work. So I think a warning is better.