ikawrakow/ik_llama.cpp
1
0
Fork 0
mirror of https://github.com/ikawrakow/ik_llama.cpp.git synced 2026-03-12 06:50:08 +00:00
Code Issues Packages Projects Releases Wiki Activity

Implement Adaptive-P Sampler (#1100)

Browse Source 
* initial implementation of adaptive-p sampler

* explicitly mark candidates unsorted + cleanup qualifiers

* cosmetic update

* reorg prototypes

* lockstep with mainline

* add _impl for _init + reorg

* add LLAMA_API to prototypes

* update sharpness to 10

* lockstep: rng seed

* delete llama_sampling member in llama_sampler_adaptive_p

* fix LLAMA_API return type

* lockstep: rng seed cont

* actually correct implementation

* lockstep: sorting behavior

* const -> constexpr for known constants

* add missing space

* fix softmax usage in adaptive p sampler

* cosmetic changes

* implement do-not-sort version of softmax

* simpify rng seed, add static to constexpr

* refactor: remove iface + use shared rng + use actually original probabilities

* adaptive-p: add dedicated rng back in

* fix initial max_logit + add float vector to adaptive p sampler context + stochastic sampling

* adaptive-p: fuse first softmax with transformation

* adaptive-p: implement binary search selection

* adaptive-p: update comment
This commit is contained in:
dungquixote42
2026-01-10 00:58:53 -05:00
committed by GitHub
parent c91cf84c8f
commit 6695c6c945
8 changed files with 226 additions and 10 deletions
Download Patch File Download Diff File Expand all files Collapse all files

14
common/common.cpp
View File

@@ -925,6 +925,16 @@ bool gpt_params_find_arg(int argc, char ** argv, const std::string & arg, gpt_pa
}
return true;
}
if (arg == "--adaptive-target") {
CHECK_ARG
sparams.adaptive_target = std::stof(argv[i]);
return true;
}
if (arg == "--adaptive-decay") {
CHECK_ARG
sparams.adaptive_decay = std::stof(argv[i]);
return true;
}
if (arg == "--spec-replace") {
CHECK_ARG
std::string target = argv[i];
@@ -2201,6 +2211,8 @@ void gpt_params_print_usage(int /*argc*/, char ** argv, const gpt_params & param
options.push_back({ "*", " --xtc-probability p", "xtc probability (default: %.1f, 0.0 = disabled)", (double)sparams.xtc_probability });
options.push_back({ "*", " --xtc-threshold t", "xtc threshold (default: %.1f, >0.5 = disabled)", (double)sparams.xtc_threshold});
options.push_back({ "*", " --top-n-sigma t", "top-n-sigma parmeter (default: %.1f, 0.0 = disabled)", (double)sparams.top_n_sigma});
options.push_back({ "*", " --adaptive-target", "adaptive-p sampling: (default: %.2f, <0.0 = disabled)", (double)sparams.adaptive_target});
options.push_back({ "*", " --adaptive-decay", "adaptive-p sampling: (default: %.2f)", (double)sparams.adaptive_decay});
options.push_back({ "*", " -l TOKEN_ID(+/-)BIAS", "modifies the likelihood of token appearing in the completion,\n"
"i.e. `--logit-bias 15043+1` to increase likelihood of token ' Hello',\n"
"or `--logit-bias 15043-1` to decrease likelihood of token ' Hello'" });
@@ -4174,6 +4186,8 @@ void yaml_dump_non_result_info(FILE * stream, const gpt_params & params, const l
fprintf(stream, "top_p: %f # default: 0.95\n", sparams.top_p);
fprintf(stream, "min_p: %f # default: 0.0\n", sparams.min_p);
fprintf(stream, "typical_p: %f # default: 1.0\n", sparams.typical_p);
fprintf(stream, "adaptive_target: %f # default: -1.0\n", sparams.adaptive_target);
fprintf(stream, "adaptive_decay: %f # default: 0.9\n", sparams.adaptive_decay);
fprintf(stream, "verbose_prompt: %s # default: false\n", params.verbose_prompt ? "true" : "false");
fprintf(stream, "display_prompt: %s # default: true\n", params.display_prompt ? "true" : "false");
}