vrr/koboldcpp
2
0
Fork 0
mirror of https://github.com/LostRuins/koboldcpp.git synced 2025-09-11 09:34:37 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Merge branch 'upstream' into concedo_experimental

Browse source 
# Conflicts:
#	.devops/full-cuda.Dockerfile
#	.devops/full.Dockerfile
#	.devops/main-cuda.Dockerfile
#	.devops/main-rocm.Dockerfile
#	.devops/main-vulkan.Dockerfile
#	.devops/main.Dockerfile
#	.devops/server-cuda.Dockerfile
#	.devops/server.Dockerfile
#	README.md
#	common/CMakeLists.txt
#	grammars/README.md
#	tests/test-grammar-integration.cpp
#	tests/test-grammar-parser.cpp
#	tests/test-json-schema-to-grammar.cpp
This commit is contained in:
Concedo 2024-06-09 17:30:05 +08:00
commit 562d980140
25 changed files with 881 additions and 676 deletions
Download patch file Download diff file Expand all files Collapse all files

114
common/common.cpp
View file

@ -201,19 +201,13 @@ void gpt_params_handle_model_default(gpt_params & params) {
}
params.hf_file = params.model;
} else if (params.model.empty()) {
std::string cache_directory = fs_get_cache_directory();
const bool success = fs_create_directory_with_parents(cache_directory);
if (!success) {
throw std::runtime_error("failed to create cache directory: " + cache_directory);
}
params.model = cache_directory + string_split(params.hf_file, '/').back();
params.model = fs_get_cache_file(string_split(params.hf_file, '/').back());
}
} else if (!params.model_url.empty()) {
if (params.model.empty()) {
auto f = string_split(params.model_url, '#').front();
f = string_split(f, '?').front();
f = string_split(f, '/').back();
params.model = "models/" + f;
params.model = fs_get_cache_file(string_split(f, '/').back());
}
} else if (params.model.empty()) {
params.model = DEFAULT_MODEL_PATH;
@ -274,6 +268,7 @@ bool gpt_params_parse(int argc, char ** argv, gpt_params & params) {
}
} catch (const std::invalid_argument & ex) {
fprintf(stderr, "%s\n", ex.what());
params = params_org;
return false;
}
@ -409,6 +404,20 @@ bool gpt_params_find_arg(int argc, char ** argv, const std::string & arg, gpt_pa
}
return true;
}
if (arg == "--in-file") {
if (++i >= argc) {
invalid_param = true;
return true;
}
std::ifstream file(argv[i]);
if (!file) {
fprintf(stderr, "error: failed to open file '%s'\n", argv[i]);
invalid_param = true;
return true;
}
params.in_files.push_back(argv[i]);
return true;
}
if (arg == "-n" || arg == "--predict" || arg == "--n-predict") {
if (++i >= argc) {
invalid_param = true;
@ -1082,7 +1091,15 @@ bool gpt_params_find_arg(int argc, char ** argv, const std::string & arg, gpt_pa
return true;
}
if (arg == "-v" || arg == "--verbose") {
params.verbose = true;
params.verbosity = 1;
return true;
}
if (arg == "--verbosity") {
if (++i >= argc) {
invalid_param = true;
return true;
}
params.verbosity = std::stoi(argv[i]);
return true;
}
if (arg == "--verbose-prompt") {
@ -1392,6 +1409,14 @@ bool gpt_params_find_arg(int argc, char ** argv, const std::string & arg, gpt_pa
params.timeout_write = std::stoi(argv[i]);
return true;
}
if (arg == "--threads-http") {
if (++i >= argc) {
invalid_param = true;
return true;
}
params.n_threads_http = std::stoi(argv[i]);
return true;
}
if (arg == "-spf" || arg == "--system-prompt-file") {
if (++i >= argc) {
invalid_param = true;
@ -1461,6 +1486,14 @@ bool gpt_params_find_arg(int argc, char ** argv, const std::string & arg, gpt_pa
params.chat_template = argv[i];
return true;
}
if (arg == "--slot-prompt-similarity" || arg == "-sps") {
if (++i >= argc) {
invalid_param = true;
return true;
}
params.slot_prompt_similarity = std::stof(argv[i]);
return true;
}
if (arg == "-pps") {
params.is_pp_shared = true;
return true;
@ -1538,6 +1571,46 @@ bool gpt_params_find_arg(int argc, char ** argv, const std::string & arg, gpt_pa
params.i_pos = std::stoi(argv[i]);
return true;
}
if (arg == "-o" || arg == "--output" || arg == "--output-file") {
if (++i >= argc) {
invalid_param = true;
return true;
}
params.out_file = argv[i];
return true;
}
if (arg == "-ofreq" || arg == "--output-frequency") {
if (++i >= argc) {
invalid_param = true;
return true;
}
params.n_out_freq = std::stoi(argv[i]);
return true;
}
if (arg == "--save-frequency") {
if (++i >= argc) {
invalid_param = true;
return true;
}
params.n_save_freq = std::stoi(argv[i]);
return true;
}
if (arg == "--process-output") {
params.process_output = true;
return true;
}
if (arg == "--no-ppl") {
params.compute_ppl = false;
return true;
}
if (arg == "--chunk" || arg == "--from-chunk") {
if (++i >= argc) {
invalid_param = true;
return true;
}
params.i_chunk = std::stoi(argv[i]);
return true;
}
#ifndef LOG_DISABLE_LOGS
// Parse args for logging parameters
if (log_param_single_parse(argv[i])) {
@ -1613,6 +1686,7 @@ void gpt_params_print_usage(int /*argc*/, char ** argv, const gpt_params & param
options.push_back({ "*", "-h, --help, --usage", "print usage and exit" });
options.push_back({ "*", " --version", "show version and build info" });
options.push_back({ "*", "-v, --verbose", "print verbose information" });
options.push_back({ "*", " --verbosity N", "set specific verbosity level (default: %d)", params.verbosity });
options.push_back({ "*", " --verbose-prompt", "print a verbose prompt before generation (default: %s)", params.verbose_prompt ? "true" : "false" });
options.push_back({ "*", " --no-display-prompt", "don't print prompt at generation (default: %s)", !params.display_prompt ? "true" : "false" });
options.push_back({ "*", "-co, --color", "colorise output to distinguish prompt and user input from generations (default: %s)", params.use_color ? "true" : "false" });
@ -1638,6 +1712,7 @@ void gpt_params_print_usage(int /*argc*/, char ** argv, const gpt_params & param
options.push_back({ "*", "-fa, --flash-attn", "enable Flash Attention (default: %s)", params.flash_attn ? "enabled" : "disabled" });
options.push_back({ "*", "-p, --prompt PROMPT", "prompt to start generation with (default: '%s')", params.prompt.c_str() });
options.push_back({ "*", "-f, --file FNAME", "a file containing the prompt (default: none)" });
options.push_back({ "*", " --in-file FNAME", "an input file (repeat to specify multiple files)" });
options.push_back({ "*", "-bf, --binary-file FNAME", "binary file containing the prompt (default: none)" });
options.push_back({ "*", "-e, --escape", "process escapes sequences (\\n, \\r, \\t, \\', \\\", \\\\) (default: %s)", params.escape ? "true" : "false" });
options.push_back({ "*", " --no-escape", "do not process escape sequences" });
@ -1805,6 +1880,14 @@ void gpt_params_print_usage(int /*argc*/, char ** argv, const gpt_params & param
options.push_back({ "passkey", " --junk N", "number of times to repeat the junk text (default: %d)", params.n_junk });
options.push_back({ "passkey", " --pos N", "position of the passkey in the junk text (default: %d)", params.i_pos });
options.push_back({ "imatrix" });
options.push_back({ "imatrix", "-o, --output FNAME", "output file (default: '%s')", params.out_file.c_str() });
options.push_back({ "imatrix", " --output-frequency N", "output the imatrix every N iterations (default: %d)", params.n_out_freq });
options.push_back({ "imatrix", " --save-frequency N", "save an imatrix copy every N iterations (default: %d)", params.n_save_freq });
options.push_back({ "imatrix", " --process-output", "collect data for the output tensor (default: %s)", params.process_output ? "true" : "false" });
options.push_back({ "imatrix", " --no-ppl", "do not compute perplexity (default: %s)", params.compute_ppl ? "true" : "false" });
options.push_back({ "imatrix", " --chunk N", "start processing the input from chunk N (default: %d)", params.i_chunk });
options.push_back({ "bench" });
options.push_back({ "bench", "-pps", "is the prompt shared across parallel sequences (default: %s)", params.is_pp_shared ? "true" : "false" });
options.push_back({ "bench", "-npp n0,n1,...", "number of prompt tokens" });
@ -1821,6 +1904,7 @@ void gpt_params_print_usage(int /*argc*/, char ** argv, const gpt_params & param
options.push_back({ "server", " --ssl-key-file FNAME", "path to file a PEM-encoded SSL private key" });
options.push_back({ "server", " --ssl-cert-file FNAME", "path to file a PEM-encoded SSL certificate" });
options.push_back({ "server", " --timeout N", "server read/write timeout in seconds (default: %d)", params.timeout_read });
options.push_back({ "server", " --threads-http N", "number of threads used to process HTTP requests (default: %d)", params.n_threads_http });
options.push_back({ "server", " --system-prompt-file FNAME",
"set a file to load a system prompt (initial prompt of all slots), this is useful for chat applications" });
options.push_back({ "server", " --log-format {text,json}",
@ -1832,6 +1916,8 @@ void gpt_params_print_usage(int /*argc*/, char ** argv, const gpt_params & param
"set custom jinja chat template (default: template taken from model's metadata)\n"
"only commonly used templates are accepted:\n"
"https://github.com/ggerganov/llama.cpp/wiki/Templates-supported-by-llama_chat_apply_template" });
options.push_back({ "server", "-sps, --slot-prompt-similarity SIMILARITY",
"how much the prompt of a request must match the prompt of a slot in order to use that slot (default: %.2f, 0.0 = disabled)\n", params.slot_prompt_similarity });
#ifndef LOG_DISABLE_LOGS
options.push_back({ "logging" });
@ -2188,6 +2274,16 @@ std::string fs_get_cache_directory() {
return ensure_trailing_slash(cache_directory);
}
std::string fs_get_cache_file(const std::string & filename) {
GGML_ASSERT(filename.find(DIRECTORY_SEPARATOR) == std::string::npos);
std::string cache_directory = fs_get_cache_directory();
const bool success = fs_create_directory_with_parents(cache_directory);
if (!success) {
throw std::runtime_error("failed to create cache directory: " + cache_directory);
}
return cache_directory + filename;
}
//
// Model utils