lookup: complement data from context with general text statistics (#5479)

* lookup: evaluation tools, use corpus/previous gens

* fixup! lookup: evaluation tools, use corpus/previous gens

* fixup! lookup: evaluation tools, use corpus/previous gens

* fixup! lookup: evaluation tools, use corpus/previous gens

* fixup! lookup: evaluation tools, use corpus/previous gens
This commit is contained in:
Johannes Gäßler 2024-03-23 01:24:36 +01:00 committed by GitHub
parent 56a00f0a2f
commit 50ccaf5eac
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
13 changed files with 774 additions and 63 deletions

View file

@ -963,6 +963,22 @@ static bool gpt_params_find_arg(int argc, char ** argv, const std::string & arg,
}
return true;
}
if (arg == "-lcs" || arg == "--lookup-cache-static") {
if (++i >= argc) {
invalid_param = true;
return true;
}
params.lookup_cache_static = argv[i];
return true;
}
if (arg == "-lcd" || arg == "--lookup-cache-dynamic") {
if (++i >= argc) {
invalid_param = true;
return true;
}
params.lookup_cache_dynamic = argv[i];
return true;
}
if (arg == "--save-all-logits" || arg == "--kl-divergence-base") {
if (++i >= argc) {
invalid_param = true;
@ -1436,6 +1452,10 @@ void gpt_print_usage(int /*argc*/, char ** argv, const gpt_params & params) {
printf(" Hugging Face model file (default: unused)\n");
printf(" -ld LOGDIR, --logdir LOGDIR\n");
printf(" path under which to save YAML logs (no logging if unset)\n");
printf(" -lcs FNAME, --lookup-cache-static FNAME\n");
printf(" path to static lookup cache to use for lookup decoding (not updated by generation)\n");
printf(" -lcd FNAME, --lookup-cache-dynamic FNAME\n");
printf(" path to dynamic lookup cache to use for lookup decoding (updated by generation)\n");
printf(" --override-kv KEY=TYPE:VALUE\n");
printf(" advanced option to override model metadata by key. may be specified multiple times.\n");
printf(" types: int, float, bool. example: --override-kv tokenizer.ggml.add_bos_token=bool:false\n");