164 return "Draft model has " + std::to_string(
n_layer)
165 +
" layer(s) — this looks like an MTP head GGUF. "
166 "Set speculative.mtp: true in your config to route through "
167 "the target-owned shared-KV MTP path instead of crashing.";
Activate model on GPU (WARM → ACTIVE).
@ optional
optional: every escape asks the host's path approver
std::string mtp_head_classical_path_error(int n_layer)
Actionable error message for gh#107: MTP head on classical path.
bool looks_like_mtp_head(int n_layer)
True when the model's layer count is consistent with an MTP head GGUF.
std::string mtp_unsupported_reason(float temperature, bool has_grammar, bool streaming)
Reason MTP cannot run for a request, or "" when the envelope is safe.