wmantly/llama.cpp-xrip
1
0 Fork 0
Code Issues Pull Requests Actions 23 Packages Projects Releases Wiki Activity

common, server : report model input/output modalities in GET /models (#29987)

Browse Source 
Signed-off-by: Adrien Gallouët <angt@huggingface.co>
This commit is contained in:
Adrien Gallouët authored and GitHub committed 2026-10-05 17:08:29 +02:00
1 parent c06f84160a
commit 4d60b4d087
10 files changed
+254 -60
Download Patch File Download Diff File Expand all files Collapse all files

No files matched your search

tools/server/server-context.cpp
+36 -30
View File
@@ -4537,40 +4537,41 @@ server_context_meta server_context::get_meta() const {
const char * ftype_name = llama_ftype_name(llama_model_ftype(impl->model_tgt));
return server_context_meta {
/* build_info */ std::string(llama_build_info()),
/* model_name */ impl->model_name,
/* model_aliases */ impl->model_aliases,
/* model_tags */ impl->model_tags,
/* model_path */ impl->params_base.model.path,
/* has_mtmd */ impl->mctx != nullptr,
/* has_inp_image */ impl->chat_params.allow_image,
/* has_inp_audio */ impl->chat_params.allow_audio,
/* has_inp_video */ impl->chat_params.allow_video,
/* json_ui_settings */ impl->json_ui_settings,
/* slot_n_ctx */ impl->n_ctx_slot(),
/* pooling_type */ llama_pooling_type(impl->ctx_tgt),
/* build_info */ std::string(llama_build_info()),
/* model_name */ impl->model_name,
/* model_aliases */ impl->model_aliases,
/* model_tags */ impl->model_tags,
/* model_path */ impl->params_base.model.path,
/* model_output_modalities */ server_model_output_modalities(common_get_decision_type(impl->model_tgt)),
/* has_mtmd */ impl->mctx != nullptr,
/* has_inp_image */ impl->chat_params.allow_image,
/* has_inp_audio */ impl->chat_params.allow_audio,
/* has_inp_video */ impl->chat_params.allow_video,
/* json_ui_settings */ impl->json_ui_settings,
/* slot_n_ctx */ impl->n_ctx_slot(),
/* pooling_type */ llama_pooling_type(impl->ctx_tgt),
/* chat_params */ impl->chat_params,
/* chat_template_caps */ common_chat_templates_get_caps(impl->chat_params.tmpls.get()),
/* chat_params */ impl->chat_params,
/* chat_template_caps */ common_chat_templates_get_caps(impl->chat_params.tmpls.get()),
/* bos_token_str */ bos_token_str,
/* eos_token_str */ eos_token_str,
/* fim_pre_token */ llama_vocab_fim_pre(impl->vocab),
/* fim_sub_token */ llama_vocab_fim_suf(impl->vocab),
/* fim_mid_token */ llama_vocab_fim_mid(impl->vocab),
/* fim_pad_token */ llama_vocab_fim_pad(impl->vocab),
/* fim_rep_token */ llama_vocab_fim_rep(impl->vocab),
/* fim_sep_token */ llama_vocab_fim_sep(impl->vocab),
/* bos_token_str */ bos_token_str,
/* eos_token_str */ eos_token_str,
/* fim_pre_token */ llama_vocab_fim_pre(impl->vocab),
/* fim_sub_token */ llama_vocab_fim_suf(impl->vocab),
/* fim_mid_token */ llama_vocab_fim_mid(impl->vocab),
/* fim_pad_token */ llama_vocab_fim_pad(impl->vocab),
/* fim_rep_token */ llama_vocab_fim_rep(impl->vocab),
/* fim_sep_token */ llama_vocab_fim_sep(impl->vocab),
/* logit_bias_eog */ impl->params_base.sampling.logit_bias_eog,
/* logit_bias_eog */ impl->params_base.sampling.logit_bias_eog,
/* model_vocab_type */ llama_vocab_type(impl->vocab),
/* model_vocab_n_tokens */ llama_vocab_n_tokens(impl->vocab),
/* model_n_ctx_train */ llama_model_n_ctx_train(impl->model_tgt),
/* model_n_embd_inp */ llama_model_n_embd(impl->model_tgt),
/* model_n_params */ llama_model_n_params(impl->model_tgt),
/* model_size */ llama_model_size(impl->model_tgt),
/* model_ftype */ ftype_name,
/* model_vocab_type */ llama_vocab_type(impl->vocab),
/* model_vocab_n_tokens */ llama_vocab_n_tokens(impl->vocab),
/* model_n_ctx_train */ llama_model_n_ctx_train(impl->model_tgt),
/* model_n_embd_inp */ llama_model_n_embd(impl->model_tgt),
/* model_n_params */ llama_model_n_params(impl->model_tgt),
/* model_size */ llama_model_size(impl->model_tgt),
/* model_ftype */ ftype_name,
};
}
@@ -4898,6 +4899,11 @@ static json get_res_model_info(const server_context_meta & meta) {
{"aliases", meta.model_aliases},
{"tags", meta.model_tags},
{"object", "model"},
{"architecture", server_model_architecture_json(
meta.has_inp_image,
meta.has_inp_audio,
meta.has_inp_video,
meta.model_output_modalities)},
{"created", std::time(0)},
{"owned_by", "llamacpp"},
{"meta", {