upstream-archive/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2026-07-23 11:10:55 -05:00
Code Issues Packages Projects Releases Wiki Activity

app : add batched-bench, fit-params, quantize & perplexity (#23459)

Browse Source 
* app : add batched-bench, fit-params, quantize & perplexity

Signed-off-by: Adrien Gallouët <angt@huggingface.co>

* Add missing main.cpp

Signed-off-by: Adrien Gallouët <angt@huggingface.co>

* Add EOL

Signed-off-by: Adrien Gallouët <angt@huggingface.co>

---------

Signed-off-by: Adrien Gallouët <angt@huggingface.co>
This commit is contained in:
Adrien Gallouët
2026-05-21 09:29:44 +02:00
committed by GitHub
parent 12e5d99078
commit 1d7ab2b947
14 changed files with 118 additions and 20 deletions
Download Patch File Download Diff File Expand all files Collapse all files

11
app/CMakeLists.txt
View File

@@ -3,7 +3,16 @@ set(TARGET llama-app)
add_executable(${TARGET} llama.cpp)
set_target_properties(${TARGET} PROPERTIES OUTPUT_NAME llama)
target_link_libraries(${TARGET} PRIVATE llama-server-impl llama-cli-impl llama-completion-impl llama-bench-impl)
target_link_libraries(${TARGET} PRIVATE
llama-server-impl
llama-cli-impl
llama-completion-impl
llama-bench-impl
llama-batched-bench-impl
llama-fit-params-impl
llama-quantize-impl
llama-perplexity-impl
)
target_compile_features(${TARGET} PRIVATE cxx_std_17)
if(LLAMA_TOOLS_INSTALL)

22
app/llama.cpp
View File

@@ -4,12 +4,18 @@
#include <string>
#include <vector>
// visible
int llama_server(int argc, char ** argv);
int llama_cli(int argc, char ** argv);
// hidden
int llama_completion(int argc, char ** argv);
int llama_bench(int argc, char ** argv);
int llama_batched_bench(int argc, char ** argv);
int llama_fit_params(int argc, char ** argv);
int llama_quantize(int argc, char ** argv);
int llama_perplexity(int argc, char ** argv);
static int help(int argc, char ** argv);
static int version(int argc, char ** argv);
@@ -22,12 +28,16 @@ struct command {
};
static const command cmds[] = {
{"serve", "HTTP API server", {"server"}, false, llama_server },
{"cli", "Command-line interactive interface", {"client"}, false, llama_cli },
{"completion", "Text completion", {"complete"}, true, llama_completion },
{"bench", "Benchmarking tool", {}, true, llama_bench },
{"version", "Show version", {}, true, version },
{"help", "Show available commands", {}, true, help },
{"serve", "HTTP API server", {"server"}, false, llama_server },
{"cli", "Command-line interactive interface", {"client"}, false, llama_cli },
{"completion", "Text completion", {"complete"}, true, llama_completion },
{"bench", "Benchmark prompt processing and text generation", {}, true, llama_bench },
{"batched-bench", "Benchmark batched decoding performance", {}, true, llama_batched_bench},
{"fit-params", "Compute parameters to fit a model in device memory", {}, true, llama_fit_params },
{"quantize", "Quantize a model", {}, true, llama_quantize },
{"perplexity", "Compute model perplexity and KL divergence", {}, true, llama_perplexity },
{"version", "Show version", {}, true, version },
{"help", "Show available commands", {}, true, help },
};
static int version(int argc, char ** argv) {