Skip to content

GitLab

Commit ff27a817 authored by wangkx1's avatar wangkx1
Browse files

init

parents
Pipeline #1521 canceled with stages
Hide whitespace changes
Inline Side-by-side
Showing with 3795 additions and 0 deletions
ollama/llm/llama.cpp/common/json-schema-to-grammar.h 0 → 100755
View file @ ff27a817
#pragma once
#include "ggml.h"
// Change JSON_ASSERT from assert() to GGML_ASSERT:
#define JSON_ASSERT GGML_ASSERT
#include "json.hpp"
std::string json_schema_to_grammar(const nlohmann::ordered_json& schema);
ollama/llm/llama.cpp/common/json.hpp 0 → 100755
View file @ ff27a817
This source diff could not be displayed because it is too large. You can view the blob instead.
ollama/llm/llama.cpp/common/log.h 0 → 100755
View file @ ff27a817
#pragma once
#include <chrono>
#include <cstring>
#include <sstream>
#include <iostream>
#include <thread>
#include <vector>
#include <algorithm>
#include <cinttypes>
// --------------------------------
//
// Basic usage:
//
// --------
//
// The LOG() and LOG_TEE() macros are ready to go by default
// they do not require any initialization.
//
// LOGLN() and LOG_TEELN() are variants which automatically
// include \n character at the end of the log string.
//
// LOG() behaves exactly like printf, by default writing to a logfile.
// LOG_TEE() additionally, prints to the screen too ( mimics Unix tee command ).
//
// Default logfile is named
// "llama.<threadID>.log"
// Default LOG_TEE() secondary output target is
// stderr
//
// Logs can be dynamically disabled or enabled using functions:
// log_disable()
// and
// log_enable()
//
// A log target can be changed with:
// log_set_target( string )
// creating and opening, or re-opening a file by string filename
// or
// log_set_target( FILE* )
// allowing to point at stderr, stdout, or any valid FILE* file handler.
//
// --------
//
// End of Basic usage.
//
// --------------------------------
// Specifies a log target.
// default uses log_handler() with "llama.log" log file
// this can be changed, by defining LOG_TARGET
// like so:
//
// #define LOG_TARGET (a valid FILE*)
// #include "log.h"
//
// or it can be simply redirected to stdout or stderr
// like so:
//
// #define LOG_TARGET stderr
// #include "log.h"
//
// The log target can also be redirected to a different function
// like so:
//
// #define LOG_TARGET log_handler_different()
// #include "log.h"
//
// FILE* log_handler_different()
// {
// return stderr;
// }
//
// or:
//
// #define LOG_TARGET log_handler_another_one("somelog.log")
// #include "log.h"
//
// FILE* log_handler_another_one(char*filename)
// {
// static FILE* logfile = nullptr;
// (...)
// if( !logfile )
// {
// fopen(...)
// }
// (...)
// return logfile
// }
//
#ifndef LOG_TARGET
#define LOG_TARGET log_handler()
#endif
#ifndef LOG_TEE_TARGET
#define LOG_TEE_TARGET stderr
#endif
// Utility for synchronizing log configuration state
// since std::optional was introduced only in c++17
enum LogTriState
{
LogTriStateSame,
LogTriStateFalse,
LogTriStateTrue
};
// Utility to obtain "pid" like unique process id and use it when creating log files.
inline std::string log_get_pid()
{
static std::string pid;
if (pid.empty())
{
// std::this_thread::get_id() is the most portable way of obtaining a "process id"
// it's not the same as "pid" but is unique enough to solve multiple instances
// trying to write to the same log.
std::stringstream ss;
ss << std::this_thread::get_id();
pid = ss.str();
}
return pid;
}
// Utility function for generating log file names with unique id based on thread id.
// invocation with log_filename_generator( "llama", "log" ) creates a string "llama.<number>.log"
// where the number is a runtime id of the current thread.
#define log_filename_generator(log_file_basename, log_file_extension) log_filename_generator_impl(LogTriStateSame, log_file_basename, log_file_extension)
// INTERNAL, DO NOT USE
inline std::string log_filename_generator_impl(LogTriState multilog, const std::string & log_file_basename, const std::string & log_file_extension)
{
static bool _multilog = false;
if (multilog != LogTriStateSame)
{
_multilog = multilog == LogTriStateTrue;
}
std::stringstream buf;
buf << log_file_basename;
if (_multilog)
{
buf << ".";
buf << log_get_pid();
}
buf << ".";
buf << log_file_extension;
return buf.str();
}
#ifndef LOG_DEFAULT_FILE_NAME
#define LOG_DEFAULT_FILE_NAME log_filename_generator("llama", "log")
#endif
// Utility for turning #define values into string literals
// so we can have a define for stderr and
// we can print "stderr" instead of literal stderr, etc.
#define LOG_STRINGIZE1(s) #s
#define LOG_STRINGIZE(s) LOG_STRINGIZE1(s)
#define LOG_TEE_TARGET_STRING LOG_STRINGIZE(LOG_TEE_TARGET)
// Allows disabling timestamps.
// in order to disable, define LOG_NO_TIMESTAMPS
// like so:
//
// #define LOG_NO_TIMESTAMPS
// #include "log.h"
//
#ifndef LOG_NO_TIMESTAMPS
#ifndef _MSC_VER
#define LOG_TIMESTAMP_FMT "[%" PRIu64 "] "
#define LOG_TIMESTAMP_VAL , (std::chrono::duration_cast<std::chrono::duration<std::uint64_t>>(std::chrono::system_clock::now().time_since_epoch())).count()
#else
#define LOG_TIMESTAMP_FMT "[%" PRIu64 "] "
#define LOG_TIMESTAMP_VAL , (std::chrono::duration_cast<std::chrono::duration<std::uint64_t>>(std::chrono::system_clock::now().time_since_epoch())).count()
#endif
#else
#define LOG_TIMESTAMP_FMT "%s"
#define LOG_TIMESTAMP_VAL ,""
#endif
#ifdef LOG_TEE_TIMESTAMPS
#ifndef _MSC_VER
#define LOG_TEE_TIMESTAMP_FMT "[%" PRIu64 "] "
#define LOG_TEE_TIMESTAMP_VAL , (std::chrono::duration_cast<std::chrono::duration<std::uint64_t>>(std::chrono::system_clock::now().time_since_epoch())).count()
#else
#define LOG_TEE_TIMESTAMP_FMT "[%" PRIu64 "] "
#define LOG_TEE_TIMESTAMP_VAL , (std::chrono::duration_cast<std::chrono::duration<std::uint64_t>>(std::chrono::system_clock::now().time_since_epoch())).count()
#endif
#else
#define LOG_TEE_TIMESTAMP_FMT "%s"
#define LOG_TEE_TIMESTAMP_VAL ,""
#endif
// Allows disabling file/line/function prefix
// in order to disable, define LOG_NO_FILE_LINE_FUNCTION
// like so:
//
// #define LOG_NO_FILE_LINE_FUNCTION
// #include "log.h"
//
#ifndef LOG_NO_FILE_LINE_FUNCTION
#ifndef _MSC_VER
#define LOG_FLF_FMT "[%24s:%5d][%24s] "
#define LOG_FLF_VAL , __FILE__, __LINE__, __FUNCTION__
#else
#define LOG_FLF_FMT "[%24s:%5ld][%24s] "
#define LOG_FLF_VAL , __FILE__, (long)__LINE__, __FUNCTION__
#endif
#else
#define LOG_FLF_FMT "%s"
#define LOG_FLF_VAL ,""
#endif
#ifdef LOG_TEE_FILE_LINE_FUNCTION
#ifndef _MSC_VER
#define LOG_TEE_FLF_FMT "[%24s:%5d][%24s] "
#define LOG_TEE_FLF_VAL , __FILE__, __LINE__, __FUNCTION__
#else
#define LOG_TEE_FLF_FMT "[%24s:%5ld][%24s] "
#define LOG_TEE_FLF_VAL , __FILE__, (long)__LINE__, __FUNCTION__
#endif
#else
#define LOG_TEE_FLF_FMT "%s"
#define LOG_TEE_FLF_VAL ,""
#endif
// INTERNAL, DO NOT USE
// USE LOG() INSTEAD
//
#if !defined(_MSC_VER) || defined(__INTEL_LLVM_COMPILER) || defined(__clang__)
#define LOG_IMPL(str, ...) \
do { \
if (LOG_TARGET != nullptr) \
{ \
fprintf(LOG_TARGET, LOG_TIMESTAMP_FMT LOG_FLF_FMT str "%s" LOG_TIMESTAMP_VAL LOG_FLF_VAL, __VA_ARGS__); \
fflush(LOG_TARGET); \
} \
} while (0)
#else
#define LOG_IMPL(str, ...) \
do { \
if (LOG_TARGET != nullptr) \
{ \
fprintf(LOG_TARGET, LOG_TIMESTAMP_FMT LOG_FLF_FMT str "%s" LOG_TIMESTAMP_VAL LOG_FLF_VAL "", ##__VA_ARGS__); \
fflush(LOG_TARGET); \
} \
} while (0)
#endif
// INTERNAL, DO NOT USE
// USE LOG_TEE() INSTEAD
//
#if !defined(_MSC_VER) || defined(__INTEL_LLVM_COMPILER) || defined(__clang__)
#define LOG_TEE_IMPL(str, ...) \
do { \
if (LOG_TARGET != nullptr) \
{ \
fprintf(LOG_TARGET, LOG_TIMESTAMP_FMT LOG_FLF_FMT str "%s" LOG_TIMESTAMP_VAL LOG_FLF_VAL, __VA_ARGS__); \
fflush(LOG_TARGET); \
} \
if (LOG_TARGET != nullptr && LOG_TARGET != stdout && LOG_TARGET != stderr && LOG_TEE_TARGET != nullptr) \
{ \
fprintf(LOG_TEE_TARGET, LOG_TEE_TIMESTAMP_FMT LOG_TEE_FLF_FMT str "%s" LOG_TEE_TIMESTAMP_VAL LOG_TEE_FLF_VAL, __VA_ARGS__); \
fflush(LOG_TEE_TARGET); \
} \
} while (0)
#else
#define LOG_TEE_IMPL(str, ...) \
do { \
if (LOG_TARGET != nullptr) \
{ \
fprintf(LOG_TARGET, LOG_TIMESTAMP_FMT LOG_FLF_FMT str "%s" LOG_TIMESTAMP_VAL LOG_FLF_VAL "", ##__VA_ARGS__); \
fflush(LOG_TARGET); \
} \
if (LOG_TARGET != nullptr && LOG_TARGET != stdout && LOG_TARGET != stderr && LOG_TEE_TARGET != nullptr) \
{ \
fprintf(LOG_TEE_TARGET, LOG_TEE_TIMESTAMP_FMT LOG_TEE_FLF_FMT str "%s" LOG_TEE_TIMESTAMP_VAL LOG_TEE_FLF_VAL "", ##__VA_ARGS__); \
fflush(LOG_TEE_TARGET); \
} \
} while (0)
#endif
// The '\0' as a last argument, is a trick to bypass the silly
// "warning: ISO C++11 requires at least one argument for the "..." in a variadic macro"
// so we can have a single macro which can be called just like printf.
// Main LOG macro.
// behaves like printf, and supports arguments the exact same way.
//
#if !defined(_MSC_VER) || defined(__clang__)
#define LOG(...) LOG_IMPL(__VA_ARGS__, "")
#else
#define LOG(str, ...) LOG_IMPL("%s" str, "", ##__VA_ARGS__, "")
#endif
// Main TEE macro.
// does the same as LOG
// and
// simultaneously writes stderr.
//
// Secondary target can be changed just like LOG_TARGET
// by defining LOG_TEE_TARGET
//
#if !defined(_MSC_VER) || defined(__clang__)
#define LOG_TEE(...) LOG_TEE_IMPL(__VA_ARGS__, "")
#else
#define LOG_TEE(str, ...) LOG_TEE_IMPL("%s" str, "", ##__VA_ARGS__, "")
#endif
// LOG macro variants with auto endline.
#if !defined(_MSC_VER) || defined(__clang__)
#define LOGLN(...) LOG_IMPL(__VA_ARGS__, "\n")
#define LOG_TEELN(...) LOG_TEE_IMPL(__VA_ARGS__, "\n")
#else
#define LOGLN(str, ...) LOG_IMPL("%s" str, "", ##__VA_ARGS__, "\n")
#define LOG_TEELN(str, ...) LOG_TEE_IMPL("%s" str, "", ##__VA_ARGS__, "\n")
#endif
// INTERNAL, DO NOT USE
inline FILE *log_handler1_impl(bool change = false, LogTriState append = LogTriStateSame, LogTriState disable = LogTriStateSame, const std::string & filename = LOG_DEFAULT_FILE_NAME, FILE *target = nullptr)
{
static bool _initialized = false;
static bool _append = false;
static bool _disabled = filename.empty() && target == nullptr;
static std::string log_current_filename{filename};
static FILE *log_current_target{target};
static FILE *logfile = nullptr;
if (change)
{
if (append != LogTriStateSame)
{
_append = append == LogTriStateTrue;
return logfile;
}
if (disable == LogTriStateTrue)
{
// Disable primary target
_disabled = true;
}
// If previously disabled, only enable, and keep previous target
else if (disable == LogTriStateFalse)
{
_disabled = false;
}
// Otherwise, process the arguments
else if (log_current_filename != filename || log_current_target != target)
{
_initialized = false;
}
}
if (_disabled)
{
// Log is disabled
return nullptr;
}
if (_initialized)
{
// with fallback in case something went wrong
return logfile ? logfile : stderr;
}
// do the (re)initialization
if (target != nullptr)
{
if (logfile != nullptr && logfile != stdout && logfile != stderr)
{
fclose(logfile);
}
log_current_filename = LOG_DEFAULT_FILE_NAME;
log_current_target = target;
logfile = target;
}
else
{
if (log_current_filename != filename)
{
if (logfile != nullptr && logfile != stdout && logfile != stderr)
{
fclose(logfile);
}
}
logfile = fopen(filename.c_str(), _append ? "a" : "w");
}
if (!logfile)
{
// Verify whether the file was opened, otherwise fallback to stderr
logfile = stderr;
fprintf(stderr, "Failed to open logfile '%s' with error '%s'\n", filename.c_str(), std::strerror(errno));
fflush(stderr);
// At this point we let the init flag be to true below, and let the target fallback to stderr
// otherwise we would repeatedly fopen() which was already unsuccessful
}
_initialized = true;
return logfile ? logfile : stderr;
}
// INTERNAL, DO NOT USE
inline FILE *log_handler2_impl(bool change = false, LogTriState append = LogTriStateSame, LogTriState disable = LogTriStateSame, FILE *target = nullptr, const std::string & filename = LOG_DEFAULT_FILE_NAME)
{
return log_handler1_impl(change, append, disable, filename, target);
}
// Disables logs entirely at runtime.
// Makes LOG() and LOG_TEE() produce no output,
// until enabled back.
#define log_disable() log_disable_impl()
// INTERNAL, DO NOT USE
inline FILE *log_disable_impl()
{
return log_handler1_impl(true, LogTriStateSame, LogTriStateTrue);
}
// Enables logs at runtime.
#define log_enable() log_enable_impl()
// INTERNAL, DO NOT USE
inline FILE *log_enable_impl()
{
return log_handler1_impl(true, LogTriStateSame, LogTriStateFalse);
}
// Sets target fir logs, either by a file name or FILE* pointer (stdout, stderr, or any valid FILE*)
#define log_set_target(target) log_set_target_impl(target)
// INTERNAL, DO NOT USE
inline FILE *log_set_target_impl(const std::string & filename) { return log_handler1_impl(true, LogTriStateSame, LogTriStateSame, filename); }
inline FILE *log_set_target_impl(FILE *target) { return log_handler2_impl(true, LogTriStateSame, LogTriStateSame, target); }
// INTERNAL, DO NOT USE
inline FILE *log_handler() { return log_handler1_impl(); }
// Enable or disable creating separate log files for each run.
// can ONLY be invoked BEFORE first log use.
#define log_multilog(enable) log_filename_generator_impl((enable) ? LogTriStateTrue : LogTriStateFalse, "", "")
// Enable or disable append mode for log file.
// can ONLY be invoked BEFORE first log use.
#define log_append(enable) log_append_impl(enable)
// INTERNAL, DO NOT USE
inline FILE *log_append_impl(bool enable)
{
return log_handler1_impl(true, enable ? LogTriStateTrue : LogTriStateFalse, LogTriStateSame);
}
inline void log_test()
{
log_disable();
LOG("01 Hello World to nobody, because logs are disabled!\n");
log_enable();
LOG("02 Hello World to default output, which is \"%s\" ( Yaaay, arguments! )!\n", LOG_STRINGIZE(LOG_TARGET));
LOG_TEE("03 Hello World to **both** default output and " LOG_TEE_TARGET_STRING "!\n");
log_set_target(stderr);
LOG("04 Hello World to stderr!\n");
LOG_TEE("05 Hello World TEE with double printing to stderr prevented!\n");
log_set_target(LOG_DEFAULT_FILE_NAME);
LOG("06 Hello World to default log file!\n");
log_set_target(stdout);
LOG("07 Hello World to stdout!\n");
log_set_target(LOG_DEFAULT_FILE_NAME);
LOG("08 Hello World to default log file again!\n");
log_disable();
LOG("09 Hello World _1_ into the void!\n");
log_enable();
LOG("10 Hello World back from the void ( you should not see _1_ in the log or the output )!\n");
log_disable();
log_set_target("llama.anotherlog.log");
LOG("11 Hello World _2_ to nobody, new target was selected but logs are still disabled!\n");
log_enable();
LOG("12 Hello World this time in a new file ( you should not see _2_ in the log or the output )?\n");
log_set_target("llama.yetanotherlog.log");
LOG("13 Hello World this time in yet new file?\n");
log_set_target(log_filename_generator("llama_autonamed", "log"));
LOG("14 Hello World in log with generated filename!\n");
#ifdef _MSC_VER
LOG_TEE("15 Hello msvc TEE without arguments\n");
LOG_TEE("16 Hello msvc TEE with (%d)(%s) arguments\n", 1, "test");
LOG_TEELN("17 Hello msvc TEELN without arguments\n");
LOG_TEELN("18 Hello msvc TEELN with (%d)(%s) arguments\n", 1, "test");
LOG("19 Hello msvc LOG without arguments\n");
LOG("20 Hello msvc LOG with (%d)(%s) arguments\n", 1, "test");
LOGLN("21 Hello msvc LOGLN without arguments\n");
LOGLN("22 Hello msvc LOGLN with (%d)(%s) arguments\n", 1, "test");
#endif
}
inline bool log_param_single_parse(const std::string & param)
{
if ( param == "--log-test")
{
log_test();
return true;
}
if ( param == "--log-disable")
{
log_disable();
return true;
}
if ( param == "--log-enable")
{
log_enable();
return true;
}
if (param == "--log-new")
{
log_multilog(true);
return true;
}
if (param == "--log-append")
{
log_append(true);
return true;
}
return false;
}
inline bool log_param_pair_parse(bool check_but_dont_parse, const std::string & param, const std::string & next = std::string())
{
if ( param == "--log-file")
{
if (!check_but_dont_parse)
{
log_set_target(log_filename_generator(next.empty() ? "unnamed" : next, "log"));
}
return true;
}
return false;
}
inline void log_print_usage()
{
printf("log options:\n");
/* format
printf(" -h, --help show this help message and exit\n");*/
/* spacing
printf("__-param----------------Description\n");*/
printf(" --log-test Run simple logging test\n");
printf(" --log-disable Disable trace logs\n");
printf(" --log-enable Enable trace logs\n");
printf(" --log-file Specify a log filename (without extension)\n");
printf(" --log-new Create a separate new log file on start. "
"Each log file will have unique name: \"<name>.<ID>.log\"\n");
printf(" --log-append Don't truncate the old log file.\n");
printf("\n");
}
#define log_dump_cmdline(argc, argv) log_dump_cmdline_impl(argc, argv)
// INTERNAL, DO NOT USE
inline void log_dump_cmdline_impl(int argc, char **argv)
{
std::stringstream buf;
for (int i = 0; i < argc; ++i)
{
if (std::string(argv[i]).find(' ') != std::string::npos)
{
buf << " \"" << argv[i] <<"\"";
}
else
{
buf << " " << argv[i];
}
}
LOGLN("Cmd:%s", buf.str().c_str());
}
#define log_tostr(var) log_var_to_string_impl(var).c_str()
inline std::string log_var_to_string_impl(bool var)
{
return var ? "true" : "false";
}
inline std::string log_var_to_string_impl(std::string var)
{
return var;
}
inline std::string log_var_to_string_impl(const std::vector<int> & var)
{
std::stringstream buf;
buf << "[ ";
bool first = true;
for (auto e : var)
{
if (first)
{
first = false;
}
else
{
buf << ", ";
}
buf << std::to_string(e);
}
buf << " ]";
return buf.str();
}
template <typename C, typename T>
inline std::string LOG_TOKENS_TOSTR_PRETTY(const C & ctx, const T & tokens)
{
std::stringstream buf;
buf << "[ ";
bool first = true;
for (const auto &token : tokens)
{
if (!first) {
buf << ", ";
} else {
first = false;
}
auto detokenized = llama_token_to_piece(ctx, token);
detokenized.erase(
std::remove_if(
detokenized.begin(),
detokenized.end(),
[](const unsigned char c) { return !std::isprint(c); }),
detokenized.end());
buf
<< "'" << detokenized << "'"
<< ":" << std::to_string(token);
}
buf << " ]";
return buf.str();
}
template <typename C, typename B>
inline std::string LOG_BATCH_TOSTR_PRETTY(const C & ctx, const B & batch)
{
std::stringstream buf;
buf << "[ ";
bool first = true;
for (int i = 0; i < batch.n_tokens; ++i)
{
if (!first) {
buf << ", ";
} else {
first = false;
}
auto detokenized = llama_token_to_piece(ctx, batch.token[i]);
detokenized.erase(
std::remove_if(
detokenized.begin(),
detokenized.end(),
[](const unsigned char c) { return !std::isprint(c); }),
detokenized.end());
buf
<< "\n" << std::to_string(i)
<< ":token '" << detokenized << "'"
<< ":pos " << std::to_string(batch.pos[i])
<< ":n_seq_id " << std::to_string(batch.n_seq_id[i])
<< ":seq_id " << std::to_string(batch.seq_id[i][0])
<< ":logits " << std::to_string(batch.logits[i]);
}
buf << " ]";
return buf.str();
}
#ifdef LOG_DISABLE_LOGS
#undef LOG
#define LOG(...) // dummy stub
#undef LOGLN
#define LOGLN(...) // dummy stub
#undef LOG_TEE
#define LOG_TEE(...) fprintf(stderr, __VA_ARGS__) // convert to normal fprintf
#undef LOG_TEELN
#define LOG_TEELN(...) fprintf(stderr, __VA_ARGS__) // convert to normal fprintf
#undef LOG_DISABLE
#define LOG_DISABLE() // dummy stub
#undef LOG_ENABLE
#define LOG_ENABLE() // dummy stub
#undef LOG_ENABLE
#define LOG_ENABLE() // dummy stub
#undef LOG_SET_TARGET
#define LOG_SET_TARGET(...) // dummy stub
#undef LOG_DUMP_CMDLINE
#define LOG_DUMP_CMDLINE(...) // dummy stub
#endif // LOG_DISABLE_LOGS
ollama/llm/llama.cpp/common/ngram-cache.cpp 0 → 100755
View file @ ff27a817
#include "ngram-cache.h"
#include "common.h"
#include "log.h"
#include <cstdint>
#include <fstream>
void llama_ngram_cache_update(llama_ngram_cache & ngram_cache, int ngram_min, int ngram_max,
std::vector<llama_token> & inp, int nnew, bool print_progress) {
const int64_t t_start_ms = ggml_time_ms();
const int64_t inp_size = inp.size();
const int64_t n_todo = inp_size * (ngram_max - ngram_min + 1);
int64_t n_done = 0;
for (int64_t ngram_size = ngram_min; ngram_size <= ngram_max; ++ngram_size) {
const int64_t i_start = std::max(inp_size - nnew, ngram_size);
for (int64_t i = i_start; i < inp_size; ++i) {
const int64_t ngram_start = i - ngram_size;
llama_ngram ngram(&inp[ngram_start], ngram_size);
const llama_token token = inp[i];
llama_ngram_cache::iterator part_it = ngram_cache.find(ngram);
if (part_it == ngram_cache.end()) {
llama_ngram_cache_part part;
part.emplace(token, 1);
ngram_cache.emplace(ngram, part);
} else {
llama_ngram_cache_part::iterator token_count_it = part_it->second.find(token);
if (token_count_it == part_it->second.end()) {
part_it->second.emplace(token, 1);
} else {
token_count_it->second++;
}
}
++n_done;
if (print_progress && n_done % 10000000 == 0) {
const int64_t t_now_ms = ggml_time_ms();
const int64_t eta_ms = (inp_size*(ngram_max-ngram_min+1) - n_done) * (t_now_ms - t_start_ms) / n_done;
const int64_t eta_min = eta_ms / (60*1000);
const int64_t eta_s = (eta_ms - 60*1000*eta_min) / 1000;
fprintf(stderr, "%s: %" PRId64 "/%" PRId64 " done, ETA: %02" PRId64 ":%02" PRId64 "\n", __func__, n_done, n_todo, eta_min, eta_s);
}
}
}
}
// Helper function to get a token from the combined, speculative sequence of inp and draft.
static llama_token get_token(const std::vector<llama_token> & inp, const std::vector<llama_token> & draft, const size_t i) {
return i < inp.size() ? inp[i] : draft[1 + i - inp.size()];
}
// If sample size or percentage are below these thresholds the draft is aborted early:
constexpr int draft_min_sample_size_lax[LLAMA_NGRAM_MAX] = { 2, 2, 1, 1};
constexpr int draft_min_percent_lax[LLAMA_NGRAM_MAX] = {66, 50, 50, 50};
constexpr int draft_min_sample_size_strict[LLAMA_NGRAM_MAX] = { 4, 3, 2, 2};
constexpr int draft_min_percent_strict[LLAMA_NGRAM_MAX] = {75, 66, 66, 66};
// Helper function that tries to draft a token from only the static ngram cache:
static llama_token try_draft(llama_ngram_cache & nc_static, const llama_ngram ngram_static) {
llama_ngram_cache::iterator part_static_it = nc_static.find(ngram_static);
if (part_static_it == nc_static.end()) {
return -1;
}
const llama_ngram_cache_part part_static = part_static_it->second;
int max_count_static = 0;
int sum_count_static = 0;
llama_token max_token = -1;
for (std::pair<llama_token, int> token_count_static : part_static) {
const llama_token token = token_count_static.first;
const int32_t count_static = token_count_static.second;
if (count_static > max_count_static) {
max_token = token;
max_count_static = count_static;
}
sum_count_static += count_static;
}
if (sum_count_static < draft_min_sample_size_lax[LLAMA_NGRAM_STATIC-1]) {
return -1;
}
if (100*max_count_static < draft_min_percent_lax[LLAMA_NGRAM_STATIC-1]*sum_count_static) {
return -1;
}
return max_token;
}
// Try to draft a token from primary cache (context/dynamic), validate with static cache:
static llama_token try_draft(
llama_ngram_cache & nc_primary, const std::vector<llama_ngram> & ngrams_primary, llama_ngram_cache_part & part_static,
const int * min_sample_size, const int * min_percent) {
llama_token drafted_token = -1;
for (int i = ngrams_primary.size()-1; i >= 0 && drafted_token == -1; --i) {
const llama_ngram ngram_primary = ngrams_primary[i];
llama_ngram_cache::iterator part_primary_it = nc_primary.find(ngram_primary);
if (part_primary_it == nc_primary.end()) {
continue;
}
const llama_ngram_cache_part part_primary = part_primary_it->second;
int max_count_primary = 0;
int max_count_static = 0;
int sum_count_primary = 0;
llama_token max_token = -1;
for (std::pair<llama_token, int> token_count_primary : part_primary) {
const llama_token token = token_count_primary.first;
llama_ngram_cache_part::iterator token_count_static_it = part_static.find(token);
const int32_t count_primary = token_count_primary.second;
const int32_t count_static = token_count_static_it != part_static.end() ? 100*token_count_static_it->second : 1;
if (count_primary*count_static > max_count_primary*max_count_static) {
max_token = token;
max_count_primary = count_primary;
max_count_static = count_static;
}
sum_count_primary += count_primary;
}
if (sum_count_primary < min_sample_size[i]) {
continue;
}
if (100*max_count_primary < min_percent[i]*sum_count_primary) {
continue;;
}
drafted_token = max_token;
}
return drafted_token;
}
void llama_ngram_cache_draft(
std::vector<llama_token> & inp, std::vector<llama_token> & draft, int n_draft, int ngram_min, int ngram_max,
llama_ngram_cache & nc_context, llama_ngram_cache & nc_dynamic, llama_ngram_cache & nc_static
) {
GGML_ASSERT(draft.size() == 1);
const int inp_size = inp.size();
if (inp_size < LLAMA_NGRAM_STATIC) {
return;
}
while ((int) draft.size()-1 < n_draft) {
llama_token drafted_token = -1;
const int ngram_start_static = inp_size-LLAMA_NGRAM_STATIC + draft.size()-1;
llama_ngram ngram_static;
for (int j = ngram_start_static; j < ngram_start_static + LLAMA_NGRAM_STATIC; ++j) {
ngram_static.tokens[j-ngram_start_static] = get_token(inp, draft, j);
}
llama_ngram_cache::iterator part_static_it = nc_static.find(ngram_static);
llama_ngram_cache_part part_static;
if (part_static_it != nc_static.end()) {
part_static = part_static_it->second;
}
// cd = context + dynamic
std::vector<llama_ngram> ngrams_cd;
for (int ngram_size_cd = ngram_min; ngram_size_cd <= ngram_max; ++ngram_size_cd) {
const int ngram_start_cd = inp_size-ngram_size_cd + draft.size()-1;
llama_ngram ngram_cd;
for (int j = ngram_start_cd; j < ngram_start_cd + ngram_size_cd; ++j) {
ngram_cd.tokens[j-ngram_start_cd] = get_token(inp, draft, j);
}
ngrams_cd.push_back(ngram_cd);
}
if (drafted_token == -1) {
drafted_token = try_draft(nc_context, ngrams_cd, part_static, draft_min_sample_size_lax, draft_min_percent_lax);
}
if (drafted_token == -1) {
drafted_token = try_draft(nc_dynamic, ngrams_cd, part_static, draft_min_sample_size_strict, draft_min_percent_strict);
}
if (drafted_token == -1) {
drafted_token = try_draft(nc_static, ngram_static);
}
if (drafted_token == -1) {
break;
}
LOG(" - draft candidate: token=%d\n", drafted_token);
draft.push_back(drafted_token);
}
}
void llama_ngram_cache_save(llama_ngram_cache & ngram_cache, std::string & filename) {
std::ofstream file_out(filename, std::ios::binary);
for (std::pair<llama_ngram, llama_ngram_cache_part> item : ngram_cache) {
const llama_ngram ngram = item.first;
llama_ngram_cache_part token_counts = item.second;
GGML_ASSERT(!token_counts.empty());
const int32_t ntokens = token_counts.size();
GGML_ASSERT(ntokens > 0);
file_out.write(reinterpret_cast<const char *>(&ngram), sizeof(llama_ngram));
file_out.write(reinterpret_cast<const char *>(&ntokens), sizeof(int32_t));
for (std::pair<llama_token, int32_t> item2 : token_counts) {
const llama_token token = item2.first;
const int32_t count = item2.second;
GGML_ASSERT(count > 0);
file_out.write(reinterpret_cast<const char *>(&token), sizeof(llama_token));
file_out.write(reinterpret_cast<const char *>(&count), sizeof(int32_t));
}
}
}
llama_ngram_cache llama_ngram_cache_load(std::string & filename) {
std::ifstream hashmap_file(filename, std::ios::binary);
if (!hashmap_file) {
throw std::ifstream::failure("Unable to open file " + filename);
}
llama_ngram_cache ngram_cache;
llama_ngram ngram;
int32_t ntokens;
llama_token token;
int32_t count;
char * ngramc = reinterpret_cast<char*>(&ngram);
char * ntokensc = reinterpret_cast<char*>(&ntokens);
char * tokenc = reinterpret_cast<char*>(&token);
char * countc = reinterpret_cast<char*>(&count);
while(hashmap_file.read(ngramc, sizeof(llama_ngram))) {
GGML_ASSERT(!hashmap_file.eof());
GGML_ASSERT(hashmap_file.read(ntokensc, sizeof(int32_t)));
GGML_ASSERT(ntokens > 0);
llama_ngram_cache_part token_counts;
for (int i = 0; i < ntokens; ++i) {
GGML_ASSERT(!hashmap_file.eof());
GGML_ASSERT(hashmap_file.read(tokenc, sizeof(llama_token)));
GGML_ASSERT(!hashmap_file.eof());
GGML_ASSERT(hashmap_file.read(countc, sizeof(int32_t)));
GGML_ASSERT(count > 0);
token_counts.emplace(token, count);
}
ngram_cache.emplace(ngram, token_counts);
}
GGML_ASSERT(hashmap_file.eof());
return ngram_cache;
}
void llama_ngram_cache_merge(llama_ngram_cache & ngram_cache_target, llama_ngram_cache & ngram_cache_add) {
for (std::pair<llama_ngram, llama_ngram_cache_part> ngram_part : ngram_cache_add) {
const llama_ngram ngram = ngram_part.first;
llama_ngram_cache_part part = ngram_part.second;
llama_ngram_cache::iterator part_merged_it = ngram_cache_target.find(ngram);
if (part_merged_it == ngram_cache_target.end()) {
ngram_cache_target.emplace(ngram, part);
continue;
}
for (std::pair<llama_token, int32_t> token_count : part) {
const llama_token token = token_count.first;
const int32_t count = token_count.second;
GGML_ASSERT(count > 0);
llama_ngram_cache_part::iterator token_count_merged_it = part_merged_it->second.find(token);
if (token_count_merged_it == part_merged_it->second.end()) {
part_merged_it->second.emplace(token, count);
continue;
}
token_count_merged_it->second += count;
}
}
}
ollama/llm/llama.cpp/common/ngram-cache.h 0 → 100755
View file @ ff27a817
#pragma once
#include "llama.h"
#include <unordered_map>
#include <string>
#include <vector>
#define LLAMA_NGRAM_MIN 1
#define LLAMA_NGRAM_MAX 4
#define LLAMA_NGRAM_STATIC 2
// Data structures to map n-grams to empirical token probabilities:
struct llama_ngram {
llama_token tokens[LLAMA_NGRAM_MAX];
llama_ngram() {
for (int i = 0; i < LLAMA_NGRAM_MAX; ++i) {
tokens[i] = -1;
}
}
llama_ngram(const llama_token * input, const int ngram_size) {
for (int i = 0; i < LLAMA_NGRAM_MAX; ++i) {
tokens[i] = i < ngram_size ? input[i] : -1;
}
}
bool operator==(const llama_ngram & other) const {
for (int i = 0; i < LLAMA_NGRAM_MAX; ++i) {
if (tokens[i] != other.tokens[i]) {
return false;
}
}
return true;
}
};
struct llama_ngram_hash_function {
size_t operator()(const llama_ngram & ngram) const {
size_t hash = 0;
for (int i = 0; i < LLAMA_NGRAM_MAX; ++i) {
hash ^= std::hash<llama_token>{}(ngram.tokens[i]);
}
return hash;
}
};
// token -> number of times token has been seen
typedef std::unordered_map<llama_token, int32_t> llama_ngram_cache_part;
// n-gram -> empirical distribution of following tokens
typedef std::unordered_map<llama_ngram, llama_ngram_cache_part, llama_ngram_hash_function> llama_ngram_cache;
// Update an ngram cache with tokens.
// ngram_cache: the cache to modify.
// ngram_min/ngram_max: the min/max size of the ngrams to extract from inp_data.
// inp_data: the token sequence with which to update ngram_cache.
// nnew: how many new tokens have been appended to inp_data since the last call to this function.
// print_progress: whether to print progress to stderr.
//
// In order to get correct results inp_data can ONLY BE APPENDED TO.
// Changes in the middle need a complete rebuild.
void llama_ngram_cache_update(
llama_ngram_cache & ngram_cache, int ngram_min, int ngram_max, std::vector<llama_token> & inp_data, int nnew, bool print_progress);
// Try to draft tokens from ngram caches.
// inp: the tokens generated so far.
// draft: the token sequence to draft. Expected to initially contain the previously sampled token.
// n_draft: maximum number of tokens to add to draft.
// ngram_min/gram_max: the min/max size of the ngrams in nc_context and nc_dynamic.
// nc_context: ngram cache based on current context.
// nc_dynamic: ngram cache based on previous user generations.
// nc_static: ngram cache generated from a large text corpus, used for validation.
void llama_ngram_cache_draft(
std::vector<llama_token> & inp, std::vector<llama_token> & draft, int n_draft, int ngram_min, int ngram_max,
llama_ngram_cache & nc_context, llama_ngram_cache & nc_dynamic, llama_ngram_cache & nc_static);
// Save an ngram cache to a file.
// ngram_cache: the ngram cache to save.
// filename: the path under which to save the ngram cache.
void llama_ngram_cache_save(llama_ngram_cache & ngram_cache, std::string & filename);
// Load an ngram cache saved with llama_ngram_cache_save.
// filename: the path from which to load the ngram cache.
// returns: an ngram cache containing the information saved to filename.
llama_ngram_cache llama_ngram_cache_load(std::string & filename);
// Merge two ngram caches.
// ngram_cache_target: the ngram cache to which to add the information from ngram_cache_add.
// ngram_cache_add: the ngram cache to add to ngram_cache_target.
void llama_ngram_cache_merge(llama_ngram_cache & ngram_cache_target, llama_ngram_cache & ngram_cache_add);
ollama/llm/llama.cpp/common/sampling.cpp 0 → 100755
View file @ ff27a817
#define LLAMA_API_INTERNAL
#include "sampling.h"
#include <random>
struct llama_sampling_context * llama_sampling_init(const struct llama_sampling_params & params) {
struct llama_sampling_context * result = new llama_sampling_context();
result->params = params;
result->grammar = nullptr;
// if there is a grammar, parse it
if (!params.grammar.empty()) {
result->parsed_grammar = grammar_parser::parse(params.grammar.c_str());
// will be empty (default) if there are parse errors
if (result->parsed_grammar.rules.empty()) {
fprintf(stderr, "%s: failed to parse grammar\n", __func__);
delete result;
return nullptr;
}
// Ensure that there is a "root" node.
if (result->parsed_grammar.symbol_ids.find("root") == result->parsed_grammar.symbol_ids.end()) {
fprintf(stderr, "%s: grammar does not contain a 'root' symbol\n", __func__);
delete result;
return nullptr;
}
std::vector<const llama_grammar_element *> grammar_rules(result->parsed_grammar.c_rules());
result->grammar = llama_grammar_init(
grammar_rules.data(),
grammar_rules.size(), result->parsed_grammar.symbol_ids.at("root"));
}
result->prev.resize(params.n_prev);
result->n_valid = 0;
llama_sampling_set_rng_seed(result, params.seed);
return result;
}
void llama_sampling_free(struct llama_sampling_context * ctx) {
if (ctx->grammar != NULL) {
llama_grammar_free(ctx->grammar);
}
delete ctx;
}
void llama_sampling_reset(llama_sampling_context * ctx) {
if (ctx->grammar != NULL) {
llama_grammar_free(ctx->grammar);
ctx->grammar = NULL;
}
if (!ctx->parsed_grammar.rules.empty()) {
std::vector<const llama_grammar_element *> grammar_rules(ctx->parsed_grammar.c_rules());
ctx->grammar = llama_grammar_init(
grammar_rules.data(),
grammar_rules.size(), ctx->parsed_grammar.symbol_ids.at("root"));
}
std::fill(ctx->prev.begin(), ctx->prev.end(), 0);
ctx->cur.clear();
ctx->n_valid = 0;
}
void llama_sampling_set_rng_seed(struct llama_sampling_context * ctx, uint32_t seed) {
if (seed == LLAMA_DEFAULT_SEED) {
seed = std::random_device{}();
}
ctx->rng.seed(seed);
}
void llama_sampling_cp(llama_sampling_context * src, llama_sampling_context * dst) {
if (dst->grammar) {
llama_grammar_free(dst->grammar);
dst->grammar = nullptr;
}
if (src->grammar) {
dst->grammar = llama_grammar_copy(src->grammar);
}
dst->prev = src->prev;
}
llama_token llama_sampling_last(llama_sampling_context * ctx) {
return ctx->prev.back();
}
std::string llama_sampling_prev_str(llama_sampling_context * ctx_sampling, llama_context * ctx_main, int n) {
const int size = ctx_sampling->prev.size();
n = std::min(n, size);
std::string result;
for (int i = size - n; i < size; i++) {
result += llama_token_to_piece(ctx_main, ctx_sampling->prev[i]);
}
return result;
}
std::string llama_sampling_print(const llama_sampling_params & params) {
char result[1024];
snprintf(result, sizeof(result),
"\trepeat_last_n = %d, repeat_penalty = %.3f, frequency_penalty = %.3f, presence_penalty = %.3f\n"
"\ttop_k = %d, tfs_z = %.3f, top_p = %.3f, min_p = %.3f, typical_p = %.3f, temp = %.3f\n"
"\tmirostat = %d, mirostat_lr = %.3f, mirostat_ent = %.3f",
params.penalty_last_n, params.penalty_repeat, params.penalty_freq, params.penalty_present,
params.top_k, params.tfs_z, params.top_p, params.min_p, params.typical_p, params.temp,
params.mirostat, params.mirostat_eta, params.mirostat_tau);
return std::string(result);
}
std::string llama_sampling_order_print(const llama_sampling_params & params) {
std::string result = "CFG -> Penalties ";
if (params.mirostat == 0) {
for (auto sampler_type : params.samplers_sequence) {
const auto sampler_type_name = llama_sampling_type_to_str(sampler_type);
if (!sampler_type_name.empty()) {
result += "-> " + sampler_type_name + " ";
}
}
} else {
result += "-> mirostat ";
}
return result;
}
std::string llama_sampling_type_to_str(llama_sampler_type sampler_type) {
switch (sampler_type) {
case llama_sampler_type::TOP_K: return "top_k";
case llama_sampler_type::TFS_Z: return "tfs_z";
case llama_sampler_type::TYPICAL_P: return "typical_p";
case llama_sampler_type::TOP_P: return "top_p";
case llama_sampler_type::MIN_P: return "min_p";
case llama_sampler_type::TEMPERATURE: return "temperature";
default : return "";
}
}
std::vector<llama_sampler_type> llama_sampling_types_from_names(const std::vector<std::string> & names, bool allow_alt_names) {
std::unordered_map<std::string, llama_sampler_type> sampler_canonical_name_map {
{"top_k", llama_sampler_type::TOP_K},
{"top_p", llama_sampler_type::TOP_P},
{"typical_p", llama_sampler_type::TYPICAL_P},
{"min_p", llama_sampler_type::MIN_P},
{"tfs_z", llama_sampler_type::TFS_Z},
{"temperature", llama_sampler_type::TEMPERATURE}
};
// since samplers names are written multiple ways
// make it ready for both system names and input names
std::unordered_map<std::string, llama_sampler_type> sampler_alt_name_map {
{"top-k", llama_sampler_type::TOP_K},
{"top-p", llama_sampler_type::TOP_P},
{"nucleus", llama_sampler_type::TOP_P},
{"typical-p", llama_sampler_type::TYPICAL_P},
{"typical", llama_sampler_type::TYPICAL_P},
{"min-p", llama_sampler_type::MIN_P},
{"tfs-z", llama_sampler_type::TFS_Z},
{"tfs", llama_sampler_type::TFS_Z},
{"temp", llama_sampler_type::TEMPERATURE}
};
std::vector<llama_sampler_type> sampler_types;
sampler_types.reserve(names.size());
for (const auto & name : names)
{
auto sampler_item = sampler_canonical_name_map.find(name);
if (sampler_item != sampler_canonical_name_map.end())
{
sampler_types.push_back(sampler_item->second);
}
else
{
if (allow_alt_names)
{
sampler_item = sampler_alt_name_map.find(name);
if (sampler_item != sampler_alt_name_map.end())
{
sampler_types.push_back(sampler_item->second);
}
}
}
}
return sampler_types;
}
std::vector<llama_sampler_type> llama_sampling_types_from_chars(const std::string & names_string) {
std::unordered_map<char, llama_sampler_type> sampler_name_map {
{'k', llama_sampler_type::TOP_K},
{'p', llama_sampler_type::TOP_P},
{'y', llama_sampler_type::TYPICAL_P},
{'m', llama_sampler_type::MIN_P},
{'f', llama_sampler_type::TFS_Z},
{'t', llama_sampler_type::TEMPERATURE}
};
std::vector<llama_sampler_type> sampler_types;
sampler_types.reserve(names_string.size());
for (const auto & c : names_string) {
const auto sampler_item = sampler_name_map.find(c);
if (sampler_item != sampler_name_map.end()) {
sampler_types.push_back(sampler_item->second);
}
}
return sampler_types;
}
// no reasons to expose this function in header
static void sampler_queue(
struct llama_context * ctx_main,
const llama_sampling_params & params,
llama_token_data_array & cur_p,
size_t min_keep) {
const float temp = params.temp;
const float dynatemp_range = params.dynatemp_range;
const float dynatemp_exponent = params.dynatemp_exponent;
const int32_t top_k = params.top_k;
const float top_p = params.top_p;
const float min_p = params.min_p;
const float tfs_z = params.tfs_z;
const float typical_p = params.typical_p;
const std::vector<llama_sampler_type> & samplers_sequence = params.samplers_sequence;
for (auto sampler_type : samplers_sequence) {
switch (sampler_type) {
case llama_sampler_type::TOP_K : llama_sample_top_k (ctx_main, &cur_p, top_k, min_keep); break;
case llama_sampler_type::TFS_Z : llama_sample_tail_free(ctx_main, &cur_p, tfs_z, min_keep); break;
case llama_sampler_type::TYPICAL_P: llama_sample_typical (ctx_main, &cur_p, typical_p, min_keep); break;
case llama_sampler_type::TOP_P : llama_sample_top_p (ctx_main, &cur_p, top_p, min_keep); break;
case llama_sampler_type::MIN_P : llama_sample_min_p (ctx_main, &cur_p, min_p, min_keep); break;
case llama_sampler_type::TEMPERATURE:
if (dynatemp_range > 0) {
float dynatemp_min = std::max(0.0f, temp - dynatemp_range);
float dynatemp_max = std::max(0.0f, temp + dynatemp_range);
llama_sample_entropy(ctx_main, &cur_p, dynatemp_min, dynatemp_max, dynatemp_exponent);
} else {
llama_sample_temp(ctx_main, &cur_p, temp);
}
break;
default : break;
}
}
}
static llama_token llama_sampling_sample_impl(
struct llama_sampling_context * ctx_sampling,
struct llama_context * ctx_main,
struct llama_context * ctx_cfg,
const int idx,
bool is_resampling) {
const llama_sampling_params & params = ctx_sampling->params;
const float temp = params.temp;
const int mirostat = params.mirostat;
const float mirostat_tau = params.mirostat_tau;
const float mirostat_eta = params.mirostat_eta;
std::vector<float> original_logits;
auto cur_p = llama_sampling_prepare(ctx_sampling, ctx_main, ctx_cfg, idx, /* apply_grammar= */ is_resampling, &original_logits);
if (ctx_sampling->grammar != NULL && !is_resampling) {
GGML_ASSERT(!original_logits.empty());
}
llama_token id = 0;
// Get a pointer to the logits
float * logits = llama_get_logits_ith(ctx_main, idx);
if (temp < 0.0) {
// greedy sampling, with probs
llama_sample_softmax(ctx_main, &cur_p);
id = cur_p.data[0].id;
} else if (temp == 0.0) {
// greedy sampling, no probs
id = llama_sample_token_greedy(ctx_main, &cur_p);
} else {
if (mirostat == 1) {
const int mirostat_m = 100;
llama_sample_temp(ctx_main, &cur_p, temp);
id = llama_sample_token_mirostat(ctx_main, &cur_p, mirostat_tau, mirostat_eta, mirostat_m, &ctx_sampling->mirostat_mu);
} else if (mirostat == 2) {
llama_sample_temp(ctx_main, &cur_p, temp);
id = llama_sample_token_mirostat_v2(ctx_main, &cur_p, mirostat_tau, mirostat_eta, &ctx_sampling->mirostat_mu);
} else {
// temperature sampling
size_t min_keep = std::max(1, params.min_keep);
sampler_queue(ctx_main, params, cur_p, min_keep);
id = llama_sample_token_with_rng(ctx_main, &cur_p, ctx_sampling->rng);
//{
// const int n_top = 10;
// LOG("top %d candidates:\n", n_top);
// for (int i = 0; i < n_top; i++) {
// const llama_token id = cur_p.data[i].id;
// (void)id; // To avoid a warning that id is unused when logging is disabled.
// LOG(" - %5d: '%12s' (%.3f)\n", id, llama_token_to_piece(ctx_main, id).c_str(), cur_p.data[i].p);
// }
//}
//LOG("sampled token: %5d: '%s'\n", id, llama_token_to_piece(ctx_main, id).c_str());
}
}
if (ctx_sampling->grammar != NULL && !is_resampling) {
// Create an array with a single token data element for the sampled id
llama_token_data single_token_data = {id, logits[id], 0.0f};
llama_token_data_array single_token_data_array = { &single_token_data, 1, false };
// Apply grammar constraints to the single token
llama_sample_grammar(ctx_main, &single_token_data_array, ctx_sampling->grammar);
// Check if the token is valid according to the grammar by seeing if its logit has been set to -INFINITY
bool is_valid = single_token_data_array.data[0].logit != -INFINITY;
// If the token is not valid according to the grammar, perform resampling
if (!is_valid) {
LOG("Resampling because token %d: '%s' does not meet grammar rules\n", id, llama_token_to_piece(ctx_main, id).c_str());
// Restore logits from the copy
std::copy(original_logits.begin(), original_logits.end(), logits);
return llama_sampling_sample_impl(ctx_sampling, ctx_main, ctx_cfg, idx, /* is_resampling= */ true);
}
}
ctx_sampling->n_valid = temp == 0.0f ? 0 : cur_p.size;
return id;
}
static llama_token_data_array llama_sampling_prepare_impl(
struct llama_sampling_context * ctx_sampling,
struct llama_context * ctx_main,
struct llama_context * ctx_cfg,
const int idx,
bool apply_grammar,
std::vector<float> * original_logits) {
const llama_sampling_params & params = ctx_sampling->params;
const int n_vocab = llama_n_vocab(llama_get_model(ctx_main));
const int32_t penalty_last_n = params.penalty_last_n < 0 ? params.n_prev : params.penalty_last_n;
const float penalty_repeat = params.penalty_repeat;
const float penalty_freq = params.penalty_freq;
const float penalty_present = params.penalty_present;
const bool penalize_nl = params.penalize_nl;
auto & prev = ctx_sampling->prev;
auto & cur = ctx_sampling->cur;
// Get a pointer to the logits
float * logits = llama_get_logits_ith(ctx_main, idx);
if (ctx_sampling->grammar != NULL && !apply_grammar) {
GGML_ASSERT(original_logits != NULL);
// Only make a copy of the original logits if we are not applying grammar checks, not sure if I actually have to do this.
*original_logits = {logits, logits + llama_n_vocab(llama_get_model(ctx_main))};
}
// apply params.logit_bias map
for (auto it = params.logit_bias.begin(); it != params.logit_bias.end(); it++) {
logits[it->first] += it->second;
}
if (ctx_cfg) {
float * logits_guidance = llama_get_logits_ith(ctx_cfg, idx);
llama_sample_apply_guidance(ctx_main, logits, logits_guidance, params.cfg_scale);
}
cur.clear();
for (llama_token token_id = 0; token_id < n_vocab; token_id++) {
cur.emplace_back(llama_token_data{token_id, logits[token_id], 0.0f});
}
llama_token_data_array cur_p = { cur.data(), cur.size(), false };
// apply penalties
const auto& penalty_tokens = params.use_penalty_prompt_tokens ? params.penalty_prompt_tokens : prev;
const int penalty_tokens_used_size = std::min((int)penalty_tokens.size(), penalty_last_n);
if (penalty_tokens_used_size) {
const float nl_logit = logits[llama_token_nl(llama_get_model(ctx_main))];
llama_sample_repetition_penalties(ctx_main, &cur_p,
penalty_tokens.data() + penalty_tokens.size() - penalty_tokens_used_size,
penalty_tokens_used_size, penalty_repeat, penalty_freq, penalty_present);
if (!penalize_nl) {
for (size_t idx = 0; idx < cur_p.size; idx++) {
if (cur_p.data[idx].id == llama_token_nl(llama_get_model(ctx_main))) {
cur_p.data[idx].logit = nl_logit;
break;
}
}
}
}
// apply grammar checks before sampling logic
if (apply_grammar && ctx_sampling->grammar != NULL) {
llama_sample_grammar(ctx_main, &cur_p, ctx_sampling->grammar);
}
return cur_p;
}
llama_token llama_sampling_sample(
struct llama_sampling_context * ctx_sampling,
struct llama_context * ctx_main,
struct llama_context * ctx_cfg,
const int idx) {
// Call the implementation function with is_resampling set to false by default
return llama_sampling_sample_impl(ctx_sampling, ctx_main, ctx_cfg, idx, /* is_resampling= */ false);
}
llama_token_data_array llama_sampling_prepare(
struct llama_sampling_context * ctx_sampling,
struct llama_context * ctx_main,
struct llama_context * ctx_cfg,
const int idx,
bool apply_grammar,
std::vector<float> * original_logits) {
return llama_sampling_prepare_impl(ctx_sampling,ctx_main, ctx_cfg, idx, apply_grammar, original_logits);
}
void llama_sampling_accept(
struct llama_sampling_context * ctx_sampling,
struct llama_context * ctx_main,
llama_token id,
bool apply_grammar) {
ctx_sampling->prev.erase(ctx_sampling->prev.begin());
ctx_sampling->prev.push_back(id);
if (ctx_sampling->grammar != NULL && apply_grammar) {
llama_grammar_accept_token(ctx_main, ctx_sampling->grammar, id);
}
}
ollama/llm/llama.cpp/common/sampling.h 0 → 100755
View file @ ff27a817
#pragma once
#include "llama.h"
#include "grammar-parser.h"
#include <random>
#include <string>
#include <unordered_map>
#include <vector>
// sampler types
enum class llama_sampler_type : char {
TOP_K = 'k',
TOP_P = 'p',
MIN_P = 'm',
TFS_Z = 'f',
TYPICAL_P = 'y',
TEMPERATURE = 't'
};
// sampling parameters
typedef struct llama_sampling_params {
int32_t n_prev = 64; // number of previous tokens to remember
int32_t n_probs = 0; // if greater than 0, output the probabilities of top n_probs tokens.
int32_t min_keep = 0; // 0 = disabled, otherwise samplers should return at least min_keep tokens
int32_t top_k = 40; // <= 0 to use vocab size
float top_p = 0.95f; // 1.0 = disabled
float min_p = 0.05f; // 0.0 = disabled
float tfs_z = 1.00f; // 1.0 = disabled
float typical_p = 1.00f; // 1.0 = disabled
float temp = 0.80f; // <= 0.0 to sample greedily, 0.0 to not output probabilities
float dynatemp_range = 0.00f; // 0.0 = disabled
float dynatemp_exponent = 1.00f; // controls how entropy maps to temperature in dynamic temperature sampler
int32_t penalty_last_n = 64; // last n tokens to penalize (0 = disable penalty, -1 = context size)
float penalty_repeat = 1.00f; // 1.0 = disabled
float penalty_freq = 0.00f; // 0.0 = disabled
float penalty_present = 0.00f; // 0.0 = disabled
int32_t mirostat = 0; // 0 = disabled, 1 = mirostat, 2 = mirostat 2.0
float mirostat_tau = 5.00f; // target entropy
float mirostat_eta = 0.10f; // learning rate
bool penalize_nl = false; // consider newlines as a repeatable token
uint32_t seed = LLAMA_DEFAULT_SEED; // the seed used to initialize llama_sampling_context
std::vector<llama_sampler_type> samplers_sequence = {
llama_sampler_type::TOP_K,
llama_sampler_type::TFS_Z,
llama_sampler_type::TYPICAL_P,
llama_sampler_type::TOP_P,
llama_sampler_type::MIN_P,
llama_sampler_type::TEMPERATURE
};
std::string grammar; // optional BNF-like grammar to constrain sampling
// Classifier-Free Guidance
// https://arxiv.org/abs/2306.17806
std::string cfg_negative_prompt; // string to help guidance
float cfg_scale = 1.f; // how strong is guidance
std::unordered_map<llama_token, float> logit_bias; // logit bias for specific tokens
std::vector<llama_token> penalty_prompt_tokens;
bool use_penalty_prompt_tokens = false;
} llama_sampling_params;
// general sampler context
// TODO: move to llama.h
struct llama_sampling_context {
// parameters that will be used for sampling
llama_sampling_params params;
// mirostat sampler state
float mirostat_mu;
llama_grammar * grammar;
// internal
grammar_parser::parse_state parsed_grammar;
// TODO: replace with ring-buffer
std::vector<llama_token> prev;
std::vector<llama_token_data> cur;
size_t n_valid; // Number of correct top tokens with correct probabilities.
std::mt19937 rng;
};
#include "common.h"
// Create a new sampling context instance.
struct llama_sampling_context * llama_sampling_init(const struct llama_sampling_params & params);
void llama_sampling_free(struct llama_sampling_context * ctx);
// Reset the sampler context
// - clear prev tokens
// - reset grammar
void llama_sampling_reset(llama_sampling_context * ctx);
// Set the sampler seed
void llama_sampling_set_rng_seed(struct llama_sampling_context * ctx, uint32_t seed);
// Copy the sampler context
void llama_sampling_cp(llama_sampling_context * src, llama_sampling_context * dst);
// Get the last sampled token
llama_token llama_sampling_last(llama_sampling_context * ctx);
// Get a string representation of the last sampled tokens
std::string llama_sampling_prev_str(llama_sampling_context * ctx_sampling, llama_context * ctx_main, int n);
// Print sampling parameters into a string
std::string llama_sampling_print(const llama_sampling_params & params);
// Print sampling order into a string
std::string llama_sampling_order_print(const llama_sampling_params & params);
std::string llama_sampling_type_to_str(llama_sampler_type sampler_type);
std::vector<llama_sampler_type> llama_sampling_types_from_names(const std::vector<std::string> & names, bool allow_alt_names);
std::vector<llama_sampler_type> llama_sampling_types_from_chars(const std::string & names_string);
// this is a common sampling function used across the examples for convenience
// it can serve as a starting point for implementing your own sampling function
// Note: When using multiple sequences, it is the caller's responsibility to call
// llama_sampling_reset when a sequence ends
//
// required:
// - ctx_main: context to use for sampling
// - ctx_sampling: sampling-specific context
//
// optional:
// - ctx_cfg: context to use for classifier-free guidance
// - idx: sample from llama_get_logits_ith(ctx, idx)
//
// returns:
// - token: sampled token
// - candidates: vector of candidate tokens
//
llama_token llama_sampling_sample(
struct llama_sampling_context * ctx_sampling,
struct llama_context * ctx_main,
struct llama_context * ctx_cfg,
int idx = -1);
// Prepares and adjusts the set of token candidates for sampling based on penalties, biases, and sampling parameters.
llama_token_data_array llama_sampling_prepare(
struct llama_sampling_context * ctx_sampling,
struct llama_context * ctx_main,
struct llama_context * ctx_cfg,
int idx = 0,
bool apply_grammar = true,
std::vector<float> * original_logits = nullptr);
void llama_sampling_accept(
struct llama_sampling_context * ctx_sampling,
struct llama_context * ctx_main,
llama_token id,
bool apply_grammar);
ollama/llm/llama.cpp/common/stb_image.h 0 → 100755
View file @ ff27a817
This source diff could not be displayed because it is too large. You can view the blob instead.
ollama/llm/llama.cpp/common/train.cpp 0 → 100755
View file @ ff27a817
#include "train.h"
#include "common.h"
#include <random>
#include <sstream>
#include <functional>
struct random_normal_distribution {
std::mt19937 gen;
std::normal_distribution<float> rd;
float min;
float max;
};
struct random_uniform_distribution {
std::mt19937 gen;
std::uniform_real_distribution<float> rd;
};
struct train_state * init_train_state() {
struct train_state * state = new struct train_state;
state->train_its = 0;
state->train_samples = 0;
state->train_tokens = 0;
state->train_epochs = 0;
state->shuffle_samples_hash = 0;
state->shuffle_sample_count = 0;
state->shuffle_next_sample = 0;
state->shuffle_rng_state_current = "";
state->shuffle_rng_state_next = "";
state->opt = new struct ggml_opt_context;
state->opt->ctx = NULL;
state->opt->params = ggml_opt_default_params(GGML_OPT_TYPE_ADAM);
state->opt->params.graph_size = LLAMA_TRAIN_MAX_NODES;
state->opt->loss_after = 0.0f;
return state;
}
void free_train_state(struct train_state * state) {
delete state->opt;
delete state;
}
struct random_normal_distribution * init_random_normal_distribution(
int seed, float mean, float std, float min, float max
) {
struct random_normal_distribution * rnd = (struct random_normal_distribution *) malloc(sizeof(struct random_normal_distribution));
rnd->gen = std::mt19937(seed);
rnd->rd = std::normal_distribution<float>{mean, std};
rnd->min = min;
rnd->max = max;
return rnd;
}
struct random_uniform_distribution * init_random_uniform_distribution(int seed, float min, float max) {
struct random_uniform_distribution * rnd = (struct random_uniform_distribution *) malloc(sizeof(struct random_uniform_distribution));
rnd->gen = std::mt19937(seed);
rnd->rd = std::uniform_real_distribution<float>{min, max};
return rnd;
}
void free_random_normal_distribution (struct random_normal_distribution * rnd) {
free(rnd);
}
void free_random_uniform_distribution(struct random_uniform_distribution * rnd) {
free(rnd);
}
struct ggml_tensor * randomize_tensor_normal(struct ggml_tensor * tensor, struct random_normal_distribution * rnd) {
float scale = 1.0f; // xavier
switch (ggml_n_dims(tensor)) {
case 1:
scale /= sqrtf((float) tensor->ne[0]);
for (int i0 = 0; i0 < tensor->ne[0]; i0++) {
float * dst = (float *) ((char *) tensor->data + i0*tensor->nb[0]);
*dst = scale * frand_normal(rnd);
}
break;
case 2:
scale /= sqrtf((float) tensor->ne[0]+tensor->ne[1]);
for (int i1 = 0; i1 < tensor->ne[1]; i1++) {
for (int i0 = 0; i0 < tensor->ne[0]; i0++) {
float * dst = (float *) ((char *) tensor->data + i0*tensor->nb[0] + i1*tensor->nb[1]);
*dst = scale * frand_normal(rnd);
}
}
break;
case 3:
scale /= sqrtf((float) tensor->ne[0]+tensor->ne[1]);
for (int i2 = 0; i2 < tensor->ne[2]; i2++) {
for (int i1 = 0; i1 < tensor->ne[1]; i1++) {
for (int i0 = 0; i0 < tensor->ne[0]; i0++) {
float * dst = (float *) ((char *) tensor->data + i0*tensor->nb[0] + i1*tensor->nb[1] + i2*tensor->nb[2]);
*dst = scale * frand_normal(rnd);
}
}
}
break;
case 4:
scale /= sqrtf((float) tensor->ne[0]+tensor->ne[1]);
for (int i3 = 0; i3 < tensor->ne[3]; i3++) {
for (int i2 = 0; i2 < tensor->ne[2]; i2++) {
for (int i1 = 0; i1 < tensor->ne[1]; i1++) {
for (int i0 = 0; i0 < tensor->ne[0]; i0++) {
float * dst = (float *) ((char *) tensor->data + i0*tensor->nb[0] + i1*tensor->nb[1] + i2*tensor->nb[2] + i3*tensor->nb[3]);
*dst = scale * frand_normal(rnd);
}
}
}
}
break;
default:
die("Unsupported tensor->n_dims");
};
return tensor;
}
struct ggml_tensor * randomize_tensor_uniform(struct ggml_tensor * tensor, struct random_uniform_distribution * rnd) {
switch (ggml_n_dims(tensor)) {
case 1:
for (int i0 = 0; i0 < tensor->ne[0]; i0++) {
float * dst = (float *) ((char *) tensor->data + i0*tensor->nb[0]);
*dst = frand_uniform(rnd);
}
break;
case 2:
for (int i1 = 0; i1 < tensor->ne[1]; i1++) {
for (int i0 = 0; i0 < tensor->ne[0]; i0++) {
float * dst = (float *) ((char *) tensor->data + i0*tensor->nb[0] + i1*tensor->nb[1]);
*dst = frand_uniform(rnd);
}
}
break;
case 3:
for (int i2 = 0; i2 < tensor->ne[2]; i2++) {
for (int i1 = 0; i1 < tensor->ne[1]; i1++) {
for (int i0 = 0; i0 < tensor->ne[0]; i0++) {
float * dst = (float *) ((char *) tensor->data + i0*tensor->nb[0] + i1*tensor->nb[1] + i2*tensor->nb[2]);
*dst = frand_uniform(rnd);
}
}
}
break;
case 4:
for (int i3 = 0; i3 < tensor->ne[3]; i3++) {
for (int i2 = 0; i2 < tensor->ne[2]; i2++) {
for (int i1 = 0; i1 < tensor->ne[1]; i1++) {
for (int i0 = 0; i0 < tensor->ne[0]; i0++) {
float * dst = (float *) ((char *) tensor->data + i0*tensor->nb[0] + i1*tensor->nb[1] + i2*tensor->nb[2] + i3*tensor->nb[3]);
*dst = frand_uniform(rnd);
}
}
}
}
break;
default:
die("Unsupported tensor->n_dims");
};
return tensor;
}
float frand() {
return (float)rand()/((float)(RAND_MAX) + 1.0f);
}
float frand_normal(struct random_normal_distribution * rnd) {
return fclamp(rnd->rd(rnd->gen), rnd->min, rnd->max);
}
float frand_uniform(struct random_uniform_distribution * rnd) {
return rnd->rd(rnd->gen);
}
int clamp(const int v, const int min, const int max) {
return ((v < min) ? (min) : (v > max) ? (max) : v);
}
float fclamp(const float v, const float min, const float max) {
return ((v < min) ? (min) : (v > max) ? (max) : v);
}
void assert_shape_1d(struct ggml_tensor * tensor, int64_t ne0) {
GGML_ASSERT(tensor->ne[0] == ne0);
GGML_ASSERT(tensor->ne[1] == 1);
GGML_ASSERT(tensor->ne[2] == 1);
GGML_ASSERT(tensor->ne[3] == 1);
}
void assert_shape_2d(struct ggml_tensor * tensor, int64_t ne0, int64_t ne1) {
GGML_ASSERT(tensor->ne[0] == ne0);
GGML_ASSERT(tensor->ne[1] == ne1);
GGML_ASSERT(tensor->ne[2] == 1);
GGML_ASSERT(tensor->ne[3] == 1);
}
void assert_shape_3d(struct ggml_tensor * tensor, int64_t ne0, int64_t ne1, int64_t ne2) {
GGML_ASSERT(tensor->ne[0] == ne0);
GGML_ASSERT(tensor->ne[1] == ne1);
GGML_ASSERT(tensor->ne[2] == ne2);
GGML_ASSERT(tensor->ne[3] == 1);
}
void assert_shape_4d(struct ggml_tensor * tensor, int64_t ne0, int64_t ne1, int64_t ne2, int64_t ne3) {
GGML_ASSERT(tensor->ne[0] == ne0);
GGML_ASSERT(tensor->ne[1] == ne1);
GGML_ASSERT(tensor->ne[2] == ne2);
GGML_ASSERT(tensor->ne[3] == ne3);
}
int64_t get_example_targets_batch(
struct llama_context * lctx,
struct ggml_tensor * tokens_input,
struct ggml_tensor * target_probs,
int64_t example_id,
const size_t * samples_offs,
const size_t * samples_begin,
const size_t * samples_size,
size_t samples_count,
const llama_token * train_data,
size_t n_train_data,
bool separate_with_eos,
bool separate_with_bos,
bool fill_with_next_samples,
bool sample_random_offsets
) {
GGML_ASSERT(samples_count > 0);
GGML_ASSERT(ggml_is_matrix(tokens_input));
GGML_ASSERT(ggml_is_3d(target_probs));
int64_t n_vocab = target_probs->ne[0];
int64_t n_tokens = tokens_input->ne[0];
int64_t n_batch = tokens_input->ne[1];
GGML_ASSERT(n_vocab == target_probs->ne[0]);
GGML_ASSERT(n_tokens == target_probs->ne[1]);
GGML_ASSERT(n_batch == target_probs->ne[2]);
int64_t used_samples = 0;
ggml_set_f32(target_probs, 0.0f);
llama_token bos = llama_token_bos(llama_get_model(lctx));
llama_token eos = llama_token_eos(llama_get_model(lctx));
// printf("%s: example_id=%d n_batch=%d n_train_samples=%zu\n", __func__, example_id, n_batch, n_train_samples);
for (int k=0; k<n_batch; ++k) {
// printf("%s: batch %d\n", __func__, k);
size_t sample_idx = (example_id + used_samples) % samples_count;
size_t sample_offs = sample_random_offsets ? samples_offs[sample_idx] : 0;
size_t sample_begin = samples_begin[sample_idx];
size_t sample_size = samples_size[sample_idx];
++used_samples;
// printf("%s: sample_idx=%zu sample=%zu\n", __func__, sample_idx, sample);
GGML_ASSERT(sample_begin+sample_size-1 < n_train_data);
ggml_set_i32_nd(tokens_input, 0, k, 0, 0, bos);
bool sample_separation_eos = !separate_with_eos;
bool sample_separation_bos = !separate_with_bos;
for (int64_t i=0; i<n_tokens; ++i) {
llama_token token = eos;
if (sample_offs >= sample_size && fill_with_next_samples) {
if (!sample_separation_eos) {
// insert eos token to separate samples
sample_separation_eos = true;
} else if (!sample_separation_bos) {
// insert bos token to separate samples
sample_separation_bos = true;
token = bos;
} else {
// sample separation is done, continue with next sample
sample_separation_eos = !separate_with_eos;
sample_separation_bos = !separate_with_bos;
sample_offs = 0;
sample_idx = (example_id + used_samples) % samples_count;
sample_begin = samples_begin[sample_idx];
sample_size = samples_size[sample_idx];
++used_samples;
}
}
// note: no else-if here
if (sample_offs < sample_size) {
token = clamp(train_data[sample_begin+sample_offs], 0, (llama_token) (n_vocab - 1));
++sample_offs;
}
ggml_set_f32_nd(target_probs, token, (int) i, (int) k, 0, +1.0f);
if (i+1<n_tokens) {
ggml_set_i32_nd(tokens_input, (int) (i + 1), (int) k, 0, 0, token);
}
}
}
return used_samples;
}
void mt19937_set_state(std::mt19937& rng, const std::string& rng_state) {
std::stringstream s_rng_state;
s_rng_state.imbue(std::locale::classic());
s_rng_state.exceptions(std::stringstream::failbit);
s_rng_state.str(rng_state);
s_rng_state >> rng;
}
std::string mt19937_get_state(const std::mt19937& rng) {
std::stringstream s_rng_state;
s_rng_state.imbue(std::locale::classic());
s_rng_state << rng;
return s_rng_state.str();
}
std::string mt19937_seed_to_state(unsigned seed) {
std::mt19937 rng(seed);
return mt19937_get_state(rng);
}
std::string shuffle_samples(
const std::string & rng_state,
size_t * shuffled_offs,
size_t * shuffled_begins,
size_t * shuffled_sizes,
const size_t * begins,
const size_t * sizes,
size_t count) {
if (count == 0) return rng_state;
std::mt19937 rng;
mt19937_set_state(rng, rng_state);
// sort indices by random value for each index
std::vector<size_t> idcs;
{
std::vector<unsigned> rnd;
idcs.resize(count);
rnd.resize(count);
for (unsigned i=0; i<count; ++i) {
idcs[i] = i;
rnd[i] = rng();
}
std::sort(idcs.begin(), idcs.end(), [&rnd](size_t a, size_t b){
// stable sort for reproducibility
return (rnd[a] == rnd[b]) ? (a < b) : (rnd[a] < rnd[b]);
});
}
// create random offsets
for (unsigned i=0; i<count; ++i) {
shuffled_offs[i] = (size_t) ((sizes[idcs[i]] - 1) * ((double) rng() / (double) (rng.max()-1)));
}
// reorder begins and sizes by sorted indices
for (unsigned i=0; i<count; ++i) {
shuffled_begins[i] = begins[idcs[i]];
}
for (unsigned i=0; i<count; ++i) {
shuffled_sizes[i] = sizes[idcs[i]];
}
return mt19937_get_state(rng);
}
size_t hash_combine(size_t h1, size_t h2) {
return h1 ^ (h2 << 1);
}
size_t compute_samples_hash(const char* fn, const size_t* samples_begin, const size_t* samples_size, size_t sample_count) {
std::hash<std::string> h_string;
std::hash<unsigned long long> h_ull;
size_t h = h_string(std::string(fn));
h = hash_combine(h, h_ull((unsigned long long) sample_count));
for (size_t i=0; i< sample_count; ++i) {
h = hash_combine(h, h_ull((unsigned long long) samples_begin[i]));
h = hash_combine(h, h_ull((unsigned long long) samples_size[i]));
}
return h;
}
std::string replace_str(const char * s, const char * needle, const char * replacement) {
std::string str = s;
size_t pos = str.find(needle);
if (pos != std::string::npos) {
str.replace(pos, strlen(needle), replacement);
}
return str;
}
void print_duration(double fmillis) {
if (fmillis < 1000.0f) {
printf("%.1fms", (float) fmillis);
return;
}
const int64_t one_sec = 1000;
const int64_t one_min = one_sec * 60;
const int64_t one_hour = one_min * 60;
const int64_t one_day = one_hour * 24;
int64_t millis = (int64_t) fmillis;
int64_t days = millis/one_day;
int64_t hours = (millis - days*one_day)/one_hour;
int64_t minutes = (millis - days*one_day - hours*one_hour)/one_min;
int64_t seconds = (millis - days*one_day - hours*one_hour - minutes*one_min)/one_sec;
// to print int64_t either cast to (long long int) or use macro PRId64 from <inttypes.h>
if (days > 0) {
printf("%lldd ", (long long int) days);
}
printf("%02lld:%02lld:%02lld", (long long int) hours, (long long int) minutes, (long long int) seconds);
}
float cosine_decay(int64_t step, int64_t decay_steps, float minimum) {
if (step > decay_steps) {
step = decay_steps;
}
const float cosine_decay = 0.50f*(1.0f + cosf(3.14159265359f*step/decay_steps));
const float decay = (1 - minimum)*cosine_decay + minimum;
return decay;
}
float cosine_decay_restart(int64_t step, int64_t decay_steps, float minimum, float restart_step_mult) {
while (step > decay_steps) {
step -= decay_steps;
decay_steps = (int64_t) (restart_step_mult * decay_steps);
}
return cosine_decay(step, decay_steps, minimum);
}
float learning_schedule(
int64_t step,
int64_t warmup_steps,
int64_t cos_decay_steps,
float learning_rate,
float overall_minimum,
float cos_decay_minimum,
float cos_decay_restart_step_mult,
bool enable_restart) {
float result =
(step < warmup_steps)
? (float) step / (float) warmup_steps
: enable_restart
? cosine_decay_restart(
step - warmup_steps,
cos_decay_steps,
cos_decay_minimum,
cos_decay_restart_step_mult)
: cosine_decay(
step,
cos_decay_steps,
cos_decay_minimum);
float min = overall_minimum / learning_rate;
result = min + result * (1.0f - min);
return result;
}
static bool are_same_layout(struct ggml_tensor * a, struct ggml_tensor * b) {
GGML_ASSERT(a != NULL);
GGML_ASSERT(b != NULL);
GGML_ASSERT(a->type == b->type);
GGML_ASSERT(ggml_are_same_shape(a, b));
GGML_ASSERT(ggml_is_contiguous(a) && ggml_is_contiguous(b));
return true;
}
void copy_tensor_by_name(struct ggml_tensor * dst, struct ggml_context * ctx, const char * name) {
if (dst == NULL) {
return;
}
struct ggml_tensor * t = ggml_get_tensor(ctx, name);
GGML_ASSERT(are_same_layout(dst, t));
memcpy(dst->data, t->data, ggml_nbytes(t));
if (strlen(ggml_get_name(dst)) == 0) {
ggml_set_name(dst, name);
}
}
// gguf constants
static const char * LLM_KV_OPTIMIZER_TYPE = "optimizer.type";
static const char * LLM_KV_OPTIMIZER_TYPE_ADAM = "adam";
static const char * LLM_KV_OPTIMIZER_TYPE_LBFGS = "lbfgs";
static const char * LLM_KV_OPTIMIZER_FILE_VERSION = "optimizer.file_version";
static const char * LLM_KV_OPTIMIZER_CONVERGENCE_PAST_COUNT = "optimizer.convergence_past_count";
static const char * LLM_KV_OPTIMIZER_PARAMETER_COUNT = "optimizer.parameter_count";
static const char * LLM_KV_OPTIMIZER_ITERATION_COUNT = "optimizer.iteration_count";
static const char * LLM_KV_OPTIMIZER_JUST_INITIALIZED = "optimizer.just_initialized";
static const char * LLM_KV_OPTIMIZER_ADAM_BEST_LOSS = "optimizer.adam.best_loss";
static const char * LLM_KV_OPTIMIZER_ADAM_PREVIOUS_LOSS = "optimizer.adam.previous_loss";
static const char * LLM_KV_OPTIMIZER_ADAM_NO_IMPROVEMENT_COUNT = "optimizer.adam.no_improvement_count";
static const char * LLM_KV_OPTIMIZER_LBFGS_APPROX_HESSIAN_COUNT = "optimizer.lbfgs.approx_hessian_count";
static const char * LLM_KV_OPTIMIZER_LBFGS_BEST_LOSS = "optimizer.lbfgs.best_loss";
static const char * LLM_KV_OPTIMIZER_LBFGS_LINE_SEARCH_STEP = "optimizer.lbfgs.line_search_step";
static const char * LLM_KV_OPTIMIZER_LBFGS_LINE_SEARCH_J = "optimizer.lbfgs.line_search_j";
static const char * LLM_KV_OPTIMIZER_LBFGS_LINE_SEARCH_K = "optimizer.lbfgs.line_search_k";
static const char * LLM_KV_OPTIMIZER_LBFGS_LINE_SEARCH_END = "optimizer.lbfgs.line_search_end";
static const char * LLM_KV_OPTIMIZER_LBFGS_NO_IMPROVEMENT_COUNT = "optimizer.lbfgs.no_improvement_count";
static const char * LLM_TENSOR_OPTIMIZER_ADAM_FIRST_MOMENTS = "optimizer.adam.first_moments";
static const char * LLM_TENSOR_OPTIMIZER_ADAM_SECOND_MOMENTS = "optimizer.adam.second_moments";
static const char * LLM_TENSOR_OPTIMIZER_ADAM_PAST_LOSS_VALUES = "optimizer.adam.past_loss_values";
static const char * LLM_TENSOR_OPTIMIZER_LBFGS_CURRENT_PARAMETERS = "optimizer.lbfgs.current_parameters";
static const char * LLM_TENSOR_OPTIMIZER_LBFGS_PREVIOUS_PARAMETERS = "optimizer.lbfgs.previous_parameters";
static const char * LLM_TENSOR_OPTIMIZER_LBFGS_CURRENT_GRADIENTS = "optimizer.lbfgs.current_gradients";
static const char * LLM_TENSOR_OPTIMIZER_LBFGS_PREVIOUS_GRADIENTS = "optimizer.lbfgs.previous_gradients";
static const char * LLM_TENSOR_OPTIMIZER_LBFGS_SEARCH_DIRECTION = "optimizer.lbfgs.search_direction";
static const char * LLM_TENSOR_OPTIMIZER_LBFGS_PAST_LOSS_VALUES = "optimizer.lbfgs.past_loss_values";
static const char * LLM_TENSOR_OPTIMIZER_LBFGS_MEMORY_ALPHA = "optimizer.lbfgs.memory_alpha";
static const char * LLM_TENSOR_OPTIMIZER_LBFGS_MEMORY_YS = "optimizer.lbfgs.memory_ys";
static const char * LLM_TENSOR_OPTIMIZER_LBFGS_MEMORY_S = "optimizer.lbfgs.memory_s";
static const char * LLM_TENSOR_OPTIMIZER_LBFGS_MEMORY_Y = "optimizer.lbfgs.memory_y";
static const char * LLM_KV_TRAINING_FILE_VERSION = "training.file_version";
static const char * LLM_KV_TRAINING_ITERATION_COUNT = "training.iteration_count";
static const char * LLM_KV_TRAINING_SAMPLE_COUNT = "training.sample_count";
static const char * LLM_KV_TRAINING_TOKEN_COUNT = "training.token_count";
static const char * LLM_KV_TRAINING_EPOCH_COUNT = "training.epoch_count";
static const char * LLM_KV_TRAINING_SHUFFLE_SAMPLES_HASH = "training.shuffle.samples_hash";
static const char * LLM_KV_TRAINING_SHUFFLE_RNG_STATE = "training.shuffle.rng_state";
static const char * LLM_KV_TRAINING_SHUFFLE_SAMPLE_COUNT = "training.shuffle.sample_count";
static const char * LLM_KV_TRAINING_SHUFFLE_NEXT_SAMPLE = "training.shuffle.next_sample";
#define GGUF_GET_KEY(ctx, dst, func, type, req, key) \
{ \
const std::string skey(key); \
const int kid = gguf_find_key(ctx, skey.c_str()); \
if (kid >= 0) { \
enum gguf_type ktype = gguf_get_kv_type(ctx, kid); \
if (ktype != (type)) { \
die_fmt("key %s has wrong type: %s", skey.c_str(), gguf_type_name(ktype)); \
} \
(dst) = func(ctx, kid); \
} else if (req) { \
die_fmt("key not found in model: %s", skey.c_str()); \
} \
}
void load_opt_context_gguf(struct gguf_context * fctx, struct ggml_context * f_ggml_ctx, struct ggml_opt_context * opt) {
// NOTE: gguf_context must be initialized with f_ggml_ctx and no_alloc=false, otherwise tensor data can not be read
uint32_t file_version;
GGUF_GET_KEY(fctx, file_version, gguf_get_val_u32, GGUF_TYPE_UINT32, true, LLM_KV_OPTIMIZER_FILE_VERSION);
GGML_ASSERT(file_version == 0);
GGUF_GET_KEY(fctx, opt->params.past, gguf_get_val_u32, GGUF_TYPE_UINT32, true, LLM_KV_OPTIMIZER_CONVERGENCE_PAST_COUNT);
GGUF_GET_KEY(fctx, opt->iter, gguf_get_val_u32, GGUF_TYPE_UINT32, true, LLM_KV_OPTIMIZER_ITERATION_COUNT);
GGUF_GET_KEY(fctx, opt->just_initialized, gguf_get_val_bool, GGUF_TYPE_BOOL, true, LLM_KV_OPTIMIZER_JUST_INITIALIZED);
uint64_t nx;
GGUF_GET_KEY(fctx, nx, gguf_get_val_u64, GGUF_TYPE_UINT64, true, LLM_KV_OPTIMIZER_PARAMETER_COUNT);
opt->nx = (size_t) nx;
// don't call ggml_opt_init until optimizer type and optimizer specific parameters are know
std::string opt_type;
GGUF_GET_KEY(fctx, opt_type, gguf_get_val_str, GGUF_TYPE_STRING, true, LLM_KV_OPTIMIZER_TYPE);
if (opt_type == LLM_KV_OPTIMIZER_TYPE_ADAM) {
opt->params.type = GGML_OPT_TYPE_ADAM;
GGUF_GET_KEY(fctx, opt->adam.fx_best, gguf_get_val_f32, GGUF_TYPE_FLOAT32, true, LLM_KV_OPTIMIZER_ADAM_BEST_LOSS);
GGUF_GET_KEY(fctx, opt->adam.fx_prev, gguf_get_val_f32, GGUF_TYPE_FLOAT32, true, LLM_KV_OPTIMIZER_ADAM_PREVIOUS_LOSS);
GGUF_GET_KEY(fctx, opt->adam.n_no_improvement, gguf_get_val_u32, GGUF_TYPE_UINT32, true, LLM_KV_OPTIMIZER_ADAM_NO_IMPROVEMENT_COUNT);
ggml_opt_init(opt->ctx, opt, opt->params, opt->nx);
copy_tensor_by_name(opt->adam.m, f_ggml_ctx, LLM_TENSOR_OPTIMIZER_ADAM_FIRST_MOMENTS);
copy_tensor_by_name(opt->adam.v, f_ggml_ctx, LLM_TENSOR_OPTIMIZER_ADAM_SECOND_MOMENTS);
copy_tensor_by_name(opt->adam.pf, f_ggml_ctx, LLM_TENSOR_OPTIMIZER_ADAM_PAST_LOSS_VALUES);
} else if (opt_type == LLM_KV_OPTIMIZER_TYPE_LBFGS) {
opt->params.type = GGML_OPT_TYPE_LBFGS;
GGUF_GET_KEY(fctx, opt->params.lbfgs.m, gguf_get_val_u32, GGUF_TYPE_UINT32, true, LLM_KV_OPTIMIZER_LBFGS_APPROX_HESSIAN_COUNT);
GGUF_GET_KEY(fctx, opt->lbfgs.fx_best, gguf_get_val_f32, GGUF_TYPE_FLOAT32, true, LLM_KV_OPTIMIZER_LBFGS_BEST_LOSS);
GGUF_GET_KEY(fctx, opt->lbfgs.step, gguf_get_val_f32, GGUF_TYPE_FLOAT32, true, LLM_KV_OPTIMIZER_LBFGS_LINE_SEARCH_STEP);
GGUF_GET_KEY(fctx, opt->lbfgs.j, gguf_get_val_i32, GGUF_TYPE_INT32, true, LLM_KV_OPTIMIZER_LBFGS_LINE_SEARCH_J);
GGUF_GET_KEY(fctx, opt->lbfgs.k, gguf_get_val_i32, GGUF_TYPE_INT32, true, LLM_KV_OPTIMIZER_LBFGS_LINE_SEARCH_K);
GGUF_GET_KEY(fctx, opt->lbfgs.end, gguf_get_val_i32, GGUF_TYPE_INT32, true, LLM_KV_OPTIMIZER_LBFGS_LINE_SEARCH_END);
GGUF_GET_KEY(fctx, opt->lbfgs.n_no_improvement, gguf_get_val_u32, GGUF_TYPE_UINT32, true, LLM_KV_OPTIMIZER_LBFGS_NO_IMPROVEMENT_COUNT);
ggml_opt_init(opt->ctx, opt, opt->params, opt->nx);
copy_tensor_by_name(opt->lbfgs.x, f_ggml_ctx, LLM_TENSOR_OPTIMIZER_LBFGS_CURRENT_PARAMETERS);
copy_tensor_by_name(opt->lbfgs.xp, f_ggml_ctx, LLM_TENSOR_OPTIMIZER_LBFGS_PREVIOUS_PARAMETERS);
copy_tensor_by_name(opt->lbfgs.g, f_ggml_ctx, LLM_TENSOR_OPTIMIZER_LBFGS_CURRENT_GRADIENTS);
copy_tensor_by_name(opt->lbfgs.gp, f_ggml_ctx, LLM_TENSOR_OPTIMIZER_LBFGS_PREVIOUS_GRADIENTS);
copy_tensor_by_name(opt->lbfgs.d, f_ggml_ctx, LLM_TENSOR_OPTIMIZER_LBFGS_SEARCH_DIRECTION);
copy_tensor_by_name(opt->lbfgs.pf, f_ggml_ctx, LLM_TENSOR_OPTIMIZER_LBFGS_PAST_LOSS_VALUES);
copy_tensor_by_name(opt->lbfgs.lmal, f_ggml_ctx, LLM_TENSOR_OPTIMIZER_LBFGS_MEMORY_ALPHA);
copy_tensor_by_name(opt->lbfgs.lmys, f_ggml_ctx, LLM_TENSOR_OPTIMIZER_LBFGS_MEMORY_YS);
copy_tensor_by_name(opt->lbfgs.lms, f_ggml_ctx, LLM_TENSOR_OPTIMIZER_LBFGS_MEMORY_S);
copy_tensor_by_name(opt->lbfgs.lmy, f_ggml_ctx, LLM_TENSOR_OPTIMIZER_LBFGS_MEMORY_Y);
} else {
die("unknown optimizer type\n");
}
}
void save_opt_context_gguf(struct gguf_context * fctx, struct ggml_opt_context * opt) {
gguf_set_val_u32(fctx, LLM_KV_OPTIMIZER_FILE_VERSION, 0);
gguf_set_val_u32(fctx, LLM_KV_OPTIMIZER_CONVERGENCE_PAST_COUNT, opt->params.past);
gguf_set_val_u64(fctx, LLM_KV_OPTIMIZER_PARAMETER_COUNT, (uint64_t) opt->nx);
gguf_set_val_u32(fctx, LLM_KV_OPTIMIZER_ITERATION_COUNT, opt->iter);
gguf_set_val_bool(fctx, LLM_KV_OPTIMIZER_JUST_INITIALIZED, opt->just_initialized);
switch (opt->params.type) {
case GGML_OPT_TYPE_ADAM:
{
gguf_set_val_str(fctx, LLM_KV_OPTIMIZER_TYPE, LLM_KV_OPTIMIZER_TYPE_ADAM);
gguf_set_val_f32(fctx, LLM_KV_OPTIMIZER_ADAM_BEST_LOSS, opt->adam.fx_best);
gguf_set_val_f32(fctx, LLM_KV_OPTIMIZER_ADAM_PREVIOUS_LOSS, opt->adam.fx_prev);
gguf_set_val_u32(fctx, LLM_KV_OPTIMIZER_ADAM_NO_IMPROVEMENT_COUNT, opt->adam.n_no_improvement);
ggml_set_name(opt->adam.m, LLM_TENSOR_OPTIMIZER_ADAM_FIRST_MOMENTS);
ggml_set_name(opt->adam.v, LLM_TENSOR_OPTIMIZER_ADAM_SECOND_MOMENTS);
if (opt->adam.pf) {
ggml_set_name(opt->adam.pf, LLM_TENSOR_OPTIMIZER_ADAM_PAST_LOSS_VALUES);
}
gguf_add_tensor(fctx, opt->adam.m);
gguf_add_tensor(fctx, opt->adam.v);
if (opt->adam.pf) {
gguf_add_tensor(fctx, opt->adam.pf);
}
} break;
case GGML_OPT_TYPE_LBFGS:
{
gguf_set_val_str(fctx, LLM_KV_OPTIMIZER_TYPE, LLM_KV_OPTIMIZER_TYPE_LBFGS);
gguf_set_val_u32(fctx, LLM_KV_OPTIMIZER_LBFGS_APPROX_HESSIAN_COUNT, opt->params.lbfgs.m);
gguf_set_val_f32(fctx, LLM_KV_OPTIMIZER_LBFGS_BEST_LOSS, opt->lbfgs.fx_best);
gguf_set_val_f32(fctx, LLM_KV_OPTIMIZER_LBFGS_LINE_SEARCH_STEP, opt->lbfgs.step);
gguf_set_val_i32(fctx, LLM_KV_OPTIMIZER_LBFGS_LINE_SEARCH_J, opt->lbfgs.j);
gguf_set_val_i32(fctx, LLM_KV_OPTIMIZER_LBFGS_LINE_SEARCH_K, opt->lbfgs.k);
gguf_set_val_i32(fctx, LLM_KV_OPTIMIZER_LBFGS_LINE_SEARCH_END, opt->lbfgs.end);
gguf_set_val_u32(fctx, LLM_KV_OPTIMIZER_LBFGS_NO_IMPROVEMENT_COUNT, opt->lbfgs.n_no_improvement);
ggml_set_name(opt->lbfgs.x, LLM_TENSOR_OPTIMIZER_LBFGS_CURRENT_PARAMETERS);
ggml_set_name(opt->lbfgs.xp, LLM_TENSOR_OPTIMIZER_LBFGS_PREVIOUS_PARAMETERS);
ggml_set_name(opt->lbfgs.g, LLM_TENSOR_OPTIMIZER_LBFGS_CURRENT_GRADIENTS);
ggml_set_name(opt->lbfgs.gp, LLM_TENSOR_OPTIMIZER_LBFGS_PREVIOUS_GRADIENTS);
ggml_set_name(opt->lbfgs.d, LLM_TENSOR_OPTIMIZER_LBFGS_SEARCH_DIRECTION);
if (opt->lbfgs.pf) {
ggml_set_name(opt->lbfgs.pf, LLM_TENSOR_OPTIMIZER_LBFGS_PAST_LOSS_VALUES);
}
ggml_set_name(opt->lbfgs.lmal, LLM_TENSOR_OPTIMIZER_LBFGS_MEMORY_ALPHA);
ggml_set_name(opt->lbfgs.lmys, LLM_TENSOR_OPTIMIZER_LBFGS_MEMORY_YS);
ggml_set_name(opt->lbfgs.lms, LLM_TENSOR_OPTIMIZER_LBFGS_MEMORY_S);
ggml_set_name(opt->lbfgs.lmy, LLM_TENSOR_OPTIMIZER_LBFGS_MEMORY_Y);
gguf_add_tensor(fctx, opt->lbfgs.x);
gguf_add_tensor(fctx, opt->lbfgs.xp);
gguf_add_tensor(fctx, opt->lbfgs.g);
gguf_add_tensor(fctx, opt->lbfgs.gp);
gguf_add_tensor(fctx, opt->lbfgs.d);
if (opt->lbfgs.pf) {
gguf_add_tensor(fctx, opt->lbfgs.pf);
}
gguf_add_tensor(fctx, opt->lbfgs.lmal);
gguf_add_tensor(fctx, opt->lbfgs.lmys);
gguf_add_tensor(fctx, opt->lbfgs.lms);
gguf_add_tensor(fctx, opt->lbfgs.lmy);
} break;
}
}
bool load_train_state_gguf(struct gguf_context * fctx, struct ggml_context * f_ggml_ctx, struct train_state * train) {
if (gguf_find_key(fctx, LLM_KV_TRAINING_FILE_VERSION) < 0) {
return false;
}
uint32_t file_version;
GGUF_GET_KEY(fctx, file_version, gguf_get_val_u32, GGUF_TYPE_UINT32, true, LLM_KV_TRAINING_FILE_VERSION);
GGML_ASSERT(file_version <= 1);
if (file_version == 0) {
GGUF_GET_KEY(fctx, train->train_its, gguf_get_val_u32, GGUF_TYPE_UINT32, true, LLM_KV_TRAINING_ITERATION_COUNT);
GGUF_GET_KEY(fctx, train->train_samples, gguf_get_val_u32, GGUF_TYPE_UINT32, true, LLM_KV_TRAINING_SAMPLE_COUNT);
GGUF_GET_KEY(fctx, train->train_tokens, gguf_get_val_u32, GGUF_TYPE_UINT32, true, LLM_KV_TRAINING_TOKEN_COUNT);
} else if (file_version == 1) {
GGUF_GET_KEY(fctx, train->train_its, gguf_get_val_u64, GGUF_TYPE_UINT64, true, LLM_KV_TRAINING_ITERATION_COUNT);
GGUF_GET_KEY(fctx, train->train_samples, gguf_get_val_u64, GGUF_TYPE_UINT64, true, LLM_KV_TRAINING_SAMPLE_COUNT);
GGUF_GET_KEY(fctx, train->train_tokens, gguf_get_val_u64, GGUF_TYPE_UINT64, true, LLM_KV_TRAINING_TOKEN_COUNT);
GGUF_GET_KEY(fctx, train->train_epochs, gguf_get_val_u64, GGUF_TYPE_UINT64, true, LLM_KV_TRAINING_EPOCH_COUNT);
GGUF_GET_KEY(fctx, train->shuffle_samples_hash, gguf_get_val_u64, GGUF_TYPE_UINT64, false, LLM_KV_TRAINING_SHUFFLE_SAMPLES_HASH);
GGUF_GET_KEY(fctx, train->shuffle_rng_state_current, gguf_get_val_str, GGUF_TYPE_STRING, false, LLM_KV_TRAINING_SHUFFLE_RNG_STATE);
GGUF_GET_KEY(fctx, train->shuffle_sample_count, gguf_get_val_u64, GGUF_TYPE_UINT64, false, LLM_KV_TRAINING_SHUFFLE_SAMPLE_COUNT);
GGUF_GET_KEY(fctx, train->shuffle_next_sample, gguf_get_val_u64, GGUF_TYPE_UINT64, false, LLM_KV_TRAINING_SHUFFLE_NEXT_SAMPLE);
}
load_opt_context_gguf(fctx, f_ggml_ctx, train->opt);
return true;
}
void save_train_state_gguf(struct gguf_context * fctx, struct train_state * train) {
gguf_set_val_u32(fctx, LLM_KV_TRAINING_FILE_VERSION, 1);
gguf_set_val_u64(fctx, LLM_KV_TRAINING_ITERATION_COUNT, train->train_its);
gguf_set_val_u64(fctx, LLM_KV_TRAINING_SAMPLE_COUNT, train->train_samples);
gguf_set_val_u64(fctx, LLM_KV_TRAINING_TOKEN_COUNT, train->train_tokens);
gguf_set_val_u64(fctx, LLM_KV_TRAINING_EPOCH_COUNT, train->train_epochs);
gguf_set_val_u64(fctx, LLM_KV_TRAINING_SHUFFLE_SAMPLES_HASH, (uint64_t) train->shuffle_samples_hash);
gguf_set_val_str(fctx, LLM_KV_TRAINING_SHUFFLE_RNG_STATE, train->shuffle_rng_state_current.c_str());
gguf_set_val_u64(fctx, LLM_KV_TRAINING_SHUFFLE_SAMPLE_COUNT, (uint64_t) train->shuffle_sample_count);
gguf_set_val_u64(fctx, LLM_KV_TRAINING_SHUFFLE_NEXT_SAMPLE, (uint64_t) train->shuffle_next_sample);
save_opt_context_gguf(fctx, train->opt);
}
struct llama_file {
// use FILE * so we don't have to re-open the file to mmap
FILE * fp;
size_t size;
llama_file(const char * fname, const char * mode) {
fp = std::fopen(fname, mode);
if (fp == NULL) {
size = 0;
} else {
seek(0, SEEK_END);
size = tell();
seek(0, SEEK_SET);
}
}
size_t tell() const {
#ifdef _WIN32
__int64 ret = _ftelli64(fp);
#else
long ret = std::ftell(fp);
#endif
GGML_ASSERT(ret != -1); // this really shouldn't fail
return (size_t) ret;
}
void seek(size_t offset, int whence) {
#ifdef _WIN32
int ret = _fseeki64(fp, (__int64) offset, whence);
#else
int ret = std::fseek(fp, (long) offset, whence);
#endif
GGML_ASSERT(ret == 0); // same
}
void read_raw(void * ptr, size_t size) {
if (size == 0) {
return;
}
errno = 0;
std::size_t ret = std::fread(ptr, size, 1, fp);
if (ferror(fp)) {
die_fmt("read error: %s", strerror(errno));
}
if (ret != 1) {
die("unexpectedly reached end of file");
}
}
std::uint32_t read_u32() {
std::uint32_t ret;
read_raw(&ret, sizeof(ret));
return ret;
}
std::string read_string(std::uint32_t len) {
std::vector<char> chars(len);
read_raw(chars.data(), len);
return std::string(chars.data(), len);
}
void write_raw(const void * ptr, size_t size) {
if (size == 0) {
return;
}
errno = 0;
size_t ret = std::fwrite(ptr, size, 1, fp);
if (ret != 1) {
die_fmt("write error: %s", strerror(errno));
}
}
void write_u32(std::uint32_t val) {
write_raw(&val, sizeof(val));
}
~llama_file() {
if (fp) {
std::fclose(fp);
}
}
};
static size_t utf8_len(char src) {
const size_t lookup[] = { 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 3, 4 };
uint8_t highbits = static_cast<uint8_t>(src) >> 4;
return lookup[highbits];
}
// mark each byte with its utf8 unit number.
// returns the number of utf8 characters.
// e.g. when bytes == '\x61\xD0\xB0\x62',
// then utf8_units will become [0,0,1,0]
// utf8_nunits will become [1,2,2,1] and 3 is returned.
// bytes where utf8_units is zero, are the begin of an utf8 character.
static size_t mark_utf8_units(const char* bytes, int * utf8_units, int * utf8_nunits, size_t count) {
size_t offs = 0;
size_t count_utf8 = 0;
while(offs < count) {
int len = (int) utf8_len(bytes[offs]);
for (int i=0; i<len; ++i) {
utf8_units[offs+i] = i;
utf8_nunits[offs+i] = len;
}
offs += len;
++count_utf8;
}
return count_utf8;
}
size_t tokenize_file(
struct llama_context * lctx,
const char * filename,
const std::string & sample_start,
bool include_sample_start,
bool overlapping_samples,
unsigned context_length,
std::vector<llama_token> & out_tokens,
std::vector<size_t> & out_samples_begin,
std::vector<size_t> & out_samples_size) {
struct llama_file f(filename, "rb");
if (f.size == 0) {
out_tokens.clear();
out_samples_begin.clear();
out_samples_size.clear();
printf("%s: warning: empty or not existing training data file '%s'\n",
__func__, filename);
return out_tokens.size();
}
// account for possible leading whitespace that will be added by tokenizer
// e.g. '\t' will be tokenized by llama spm tokenizer to [29871, 12]
const int n_max_tokens_overhead = 1;
std::vector<char> buf;
buf.resize(f.size);
f.read_raw(buf.data(), f.size);
std::vector<int> utf8_units;
std::vector<int> utf8_nunits;
utf8_units.resize(buf.size());
utf8_nunits.resize(buf.size());
mark_utf8_units(buf.data(), utf8_units.data(), utf8_nunits.data(), buf.size());
if (sample_start.size() == 0) {
// tokenize all data at once
out_tokens.resize(buf.size() + n_max_tokens_overhead);
int n_tokens = llama_tokenize(
llama_get_model(lctx),
buf.data(),
(int) buf.size(),
out_tokens.data(),
(int) out_tokens.size(),
false, false);
if (n_tokens < 0) {
out_tokens.resize(-n_tokens);
n_tokens = llama_tokenize(
llama_get_model(lctx),
buf.data(),
(int) buf.size(),
out_tokens.data(),
(int) out_tokens.size(),
false, false);
}
if (n_tokens >= 0) {
out_tokens.resize(n_tokens);
}
// generate sample starts at all token positions
out_samples_begin.clear();
out_samples_begin.push_back(0);
out_samples_size.push_back(std::min((size_t) context_length, out_tokens.size()));
size_t end = (out_tokens.size() >= context_length) ? (out_tokens.size() - context_length) : 0;
for (size_t sample_begin = 1; sample_begin < end; ++sample_begin) {
out_samples_begin.push_back(sample_begin);
out_samples_size.push_back(context_length);
}
} else {
// split data into samples and tokenize each sample
std::string data_str(buf.data(), buf.size());
out_samples_begin.clear();
out_samples_size.clear();
out_tokens.clear();
// find all positions of pattern sample_start
size_t sample_begin = data_str.find(sample_start, 0);
while (sample_begin != std::string::npos) {
out_samples_begin.push_back(sample_begin);
const size_t search_start = sample_begin + sample_start.size();
sample_begin = data_str.find(sample_start, search_start);
}
if (out_samples_begin.size() == 0) {
printf("%s: warning: sample start pattern '%s' not found. inserting single sample at data begin\n",
__func__, sample_start.c_str());
out_samples_begin.push_back(0);
}
out_samples_size.resize(out_samples_begin.size(), 0);
std::vector<char> buf_sample;
std::vector<llama_token> tok_sample;
const size_t sample_begin_offset = (include_sample_start ? 0 : sample_start.size());
size_t found_too_big_sample = 0;
size_t found_too_small_sample = 0;
size_t found_empty_sample = 0;
size_t found_min_sample_size = SIZE_MAX;
size_t found_max_sample_size = 0;
size_t max_token_text_size = 0;
int n_vocab = llama_n_vocab(llama_get_model(lctx));
for (llama_token token=0; token < n_vocab; ++token) {
max_token_text_size = std::max(
max_token_text_size,
strlen(llama_token_get_text(llama_get_model(lctx), token)));
}
// upper bound of context byte length.
// strings with this byte length should always tokenize to at least context_length tokens.
size_t context_byte_len = max_token_text_size*context_length;
for (unsigned i=0; i<out_samples_begin.size(); ++i) {
// determine sample begin and end from pattern positions
size_t sample_begin = out_samples_begin[i] + sample_begin_offset;
size_t sample_end = overlapping_samples
? std::min(
data_str.size(),
sample_begin + context_byte_len)
: (i+1 < out_samples_begin.size()
? out_samples_begin[i+1]
: data_str.size());
if (sample_end < utf8_units.size() && utf8_units[sample_end] > 0) {
// sample end is in the middle of an utf8 character.
// advance sample_end to the begin of the next utf8 character.
sample_end += utf8_nunits[sample_end] - utf8_units[sample_end];
}
size_t sample_size = sample_end - sample_begin;
if (sample_size == 0) {
++found_empty_sample;
}
if (sample_size > 0) {
// llama_tokenize expects zero terminated string,
// copy sample into buffer and zero terminate it.
buf_sample.resize(sample_size);
memcpy(buf_sample.data(), data_str.data() + sample_begin, sample_size);
// printf("sample: '%s'\n", buf_sample.data());
// tokenize the sample
tok_sample.resize(buf_sample.size() + n_max_tokens_overhead);
int n_tokens = llama_tokenize(llama_get_model(lctx),
buf_sample.data(),
(int) buf_sample.size(),
tok_sample.data(),
(int) tok_sample.size(),
false, false);
if (n_tokens < 0) {
tok_sample.resize(-n_tokens);
n_tokens = llama_tokenize(llama_get_model(lctx),
buf_sample.data(),
(int) buf_sample.size(),
tok_sample.data(),
(int) tok_sample.size(),
false, false);
GGML_ASSERT(n_tokens >= 0);
}
GGML_ASSERT(n_tokens <= (int) tok_sample.size());
if ((size_t) n_tokens > context_length) {
++found_too_big_sample;
} else if ((size_t) n_tokens < context_length) {
++found_too_small_sample;
}
found_max_sample_size = std::max(found_max_sample_size, (size_t) n_tokens);
found_min_sample_size = std::min(found_min_sample_size, (size_t) n_tokens);
// write out tokens, start and size of sample
// overwrite the string start position with the token start position
out_samples_begin[i] = out_tokens.size();
out_samples_size[i] = (size_t) n_tokens;
out_tokens.insert(out_tokens.end(), tok_sample.begin(), tok_sample.begin() + n_tokens);
} else {
out_samples_begin[i] = out_tokens.size();
out_samples_size[i] = 0;
}
}
if (found_too_big_sample > 0) {
printf("%s: warning: found %zu samples (max length %zu) that exceed context length of %u. samples will be cut off.\n",
__func__, found_too_big_sample, found_max_sample_size, context_length);
}
if (found_too_small_sample > 0) {
printf("%s: warning: found %zu samples (min length %zu) that are shorter than context length of %u.\n",
__func__, found_too_small_sample, found_min_sample_size, context_length);
}
if (found_empty_sample) {
printf("%s: warning: found %zu empty samples.\n",
__func__, found_empty_sample);
}
}
printf("%s: total number of samples: %zu\n",
__func__, out_samples_begin.size());
GGML_ASSERT(out_samples_begin.size() == out_samples_size.size());
return out_tokens.size();
}
std::string get_train_filename(const char * filename, const char * pattern_it, const char * latest, int64_t iteration) {
std::string sit = (iteration >= 0) ? std::to_string(iteration) : std::string(latest);
return replace_str(filename, pattern_it, sit.c_str());
}
struct train_params_common get_default_train_params_common() {
struct train_params_common params;
params.fn_train_data = "shakespeare.txt";
params.fn_checkpoint_in = "checkpoint.gguf";
params.fn_checkpoint_out = "checkpoint-ITERATION.gguf";
params.pattern_fn_it = "ITERATION";
params.fn_latest = "LATEST";
params.print_usage = false;
params.save_every = 10;
params.seed = -1;
params.n_ctx = 128;
params.n_threads = 6;
params.n_batch = 8;
params.n_gradient_accumulation = 1;
params.n_epochs = -1;
params.n_gpu_layers = 0;
params.custom_n_ctx = false;
params.use_flash = false;
params.use_checkpointing = true;
params.sample_start = "";
params.include_sample_start = false;
params.escape = false;
params.overlapping_samples = false;
params.fill_with_next_samples = false;
params.separate_with_eos = false;
params.separate_with_bos = true;
params.sample_random_offsets = false;
params.force_reshuffle = false;
params.opt_past = 0;
params.opt_delta = 1e-5f;
params.opt_max_no_improvement = 0;
params.warmup = 100;
params.cos_decay_steps = 1000;
params.cos_decay_restart = 1.1f;
params.cos_decay_min = 0.1f;
params.enable_restart = false;
params.adam_n_iter = 256;
params.adam_alpha = 1e-3f;
params.adam_min_alpha = 0;
params.adam_decay = 1e-1f;
params.adam_decay_min_ndim = 2;
params.adam_beta1 = 0.9f;
params.adam_beta2 = 0.999f;
params.adam_gclip = 1.0f;
params.adam_eps_f = 0.0f;
return params;
}
void print_common_train_usage(int /*argc*/, char ** /*argv*/, const struct train_params_common * params) {
// fprintf(stderr, "usage: %s [options]\n", argv[0]);
// fprintf(stderr, "\n");
// fprintf(stderr, "options:\n");
// fprintf(stderr, " -h, --help show this help message and exit\n");
fprintf(stderr, " --train-data FNAME path from which to load training data (default '%s')\n", params->fn_train_data);
fprintf(stderr, " --checkpoint-in FNAME path from which to load training checkpoint (default '%s')\n", params->fn_checkpoint_in);
fprintf(stderr, " --checkpoint-out FNAME path to save training checkpoint (default '%s')\n", params->fn_checkpoint_out);
fprintf(stderr, " --pattern-fn-it STR pattern in output filenames to be replaced by iteration number (default '%s')\n", params->pattern_fn_it);
fprintf(stderr, " --fn-latest STR string to use instead of iteration number for saving latest output (default '%s')\n", params->fn_latest);
fprintf(stderr, " --save-every N save checkpoint and lora every N iterations. Disabled when N <= 0. (default '%d')\n", params->save_every);
fprintf(stderr, " -s SEED, --seed SEED RNG seed (default: -1, use random seed for -1)\n");
fprintf(stderr, " -c N, --ctx N Context size used during training (default %d)\n", params->n_ctx);
fprintf(stderr, " -t N, --threads N Number of threads (default %d)\n", params->n_threads);
fprintf(stderr, " -b N, --batch N Parallel batch size (default %d)\n", params->n_batch);
fprintf(stderr, " --grad-acc N Number of gradient accumulation steps (simulates larger batch size of batch*gradacc) (default %d)\n", params->n_gradient_accumulation);
fprintf(stderr, " --sample-start STR Sets the starting point for samples after the specified pattern. If empty use every token position as sample start. (default '%s')\n", params->sample_start.c_str());
fprintf(stderr, " --include-sample-start Include the sample start in the samples. (default off)\n");
fprintf(stderr, " --escape process sample start escapes sequences (\\n, \\r, \\t, \\', \\\", \\\\)\n");
fprintf(stderr, " --overlapping-samples Samples may overlap, will include sample-start of second and following samples. When off, samples will end at begin of next sample. (default off)\n");
fprintf(stderr, " --fill-with-next-samples Samples shorter than context length will be followed by the next (shuffled) samples. (default off)\n");
fprintf(stderr, " --separate-with-eos When fill-with-next-samples, insert end-of-sequence token between samples.%s\n", params->separate_with_eos ? " (default)" : "");
fprintf(stderr, " --separate-with-bos When fill-with-next-samples, insert begin-of-sequence token between samples.%s\n", params->separate_with_bos ? " (default)" : "");
fprintf(stderr, " --no-separate-with-eos When fill-with-next-samples, don't insert end-of-sequence token between samples.%s\n", !params->separate_with_eos ? " (default)" : "");
fprintf(stderr, " --no-separate-with-bos When fill-with-next-samples, don't insert begin-of-sequence token between samples.%s\n", !params->separate_with_bos ? " (default)" : "");
fprintf(stderr, " --sample-random-offsets Use samples beginning at random offsets. Together with fill-with-next-samples this may help for training endless text generation.%s\n", params->sample_random_offsets ? " (default)" : "");
fprintf(stderr, " --force-reshuffle Force a reshuffling of data at program start, otherwise the shuffling of loaded checkpoint is resumed.\n");
fprintf(stderr, " --no-flash Don't use flash attention \n");
fprintf(stderr, " --use-flash Use flash attention (default)\n");
fprintf(stderr, " --no-checkpointing Don't use gradient checkpointing\n");
fprintf(stderr, " --use-checkpointing Use gradient checkpointing (default)\n");
fprintf(stderr, " --warmup N Only for Adam optimizer. Number of warmup steps (default %d)\n", params->warmup);
fprintf(stderr, " --cos-decay-steps N Only for Adam optimizer. Number of cosine decay steps (default %d)\n", params->cos_decay_steps);
fprintf(stderr, " --cos-decay-restart N Only for Adam optimizer. Increase of cosine decay steps after restart (default %f)\n", params->cos_decay_restart);
fprintf(stderr, " --cos-decay-min N Only for Adam optimizer. Cosine decay minimum (default %f)\n", params->cos_decay_min);
fprintf(stderr, " --enable-restart N Only for Adam optimizer. Enable restarts of cos-decay %s\n", params->enable_restart ? "(default)" : "");
fprintf(stderr, " --disable-restart N Only for Adam optimizer. Disable restarts of cos-decay %s\n", !params->enable_restart ? "(default)" : "");
fprintf(stderr, " --opt-past N Number of optimization iterations to track for delta convergence test. Disabled when zero. (default %d)\n", params->opt_past);
fprintf(stderr, " --opt-delta N Maximum delta for delta convergence test. Disabled when <= zero. (default %f)\n", params->opt_delta);
fprintf(stderr, " --opt-max-no-improvement N Maximum number of optimization iterations with no improvement. Disabled when <= zero. (default %d)\n", params->opt_max_no_improvement);
fprintf(stderr, " --epochs N Maximum number epochs to process. (default %d)\n", params->n_epochs);
fprintf(stderr, " --adam-iter N Maximum number of Adam optimization iterations for each batch (default %d)\n", params->adam_n_iter);
fprintf(stderr, " --adam-alpha N Adam learning rate alpha (default %f)\n", params->adam_alpha);
fprintf(stderr, " --adam-min-alpha N Adam minimum learning rate alpha - including warmup phase (default %f)\n", params->adam_min_alpha);
fprintf(stderr, " --adam-decay N AdamW weight decay. Values greater zero enable AdamW instead of regular Adam. (default %f)\n", params->adam_decay);
fprintf(stderr, " --adam-decay-min-ndim N Minimum number of tensor dimensions to apply AdamW weight decay. Weight decay is not applied to tensors with less n_dims. (default %d)\n", params->adam_decay_min_ndim);
fprintf(stderr, " --adam-beta1 N AdamW beta1 in interval [0,1). How much to smooth the first moment of gradients. (default %f)\n", params->adam_beta1);
fprintf(stderr, " --adam-beta2 N AdamW beta2 in interval [0,1). How much to smooth the second moment of gradients. (default %f)\n", params->adam_beta2);
fprintf(stderr, " --adam-gclip N AdamW gradient clipping. Disabled when zero. (default %f)\n", params->adam_gclip);
fprintf(stderr, " --adam-epsf N AdamW epsilon for convergence test. Disabled when <= zero. (default %f)\n", params->adam_eps_f);
fprintf(stderr, " -ngl N, --n-gpu-layers N Number of model layers to offload to GPU (default %d)", params->n_gpu_layers);
fprintf(stderr, "\n");
}
bool consume_common_train_arg(
int argc, char ** argv, int * idx, struct train_params_common * params, bool * invalid_param
) {
int& i = *idx;
std::string arg = argv[i];
const std::string arg_prefix = "--";
if (arg.compare(0, arg_prefix.size(), arg_prefix) == 0) {
std::replace(arg.begin(), arg.end(), '_', '-');
}
if (arg == "--train-data") {
if (++i >= argc) {
*invalid_param = true;
return true;
}
params->fn_train_data = argv[i];
} else if (arg == "--checkpoint-in") {
if (++i >= argc) {
*invalid_param = true;
return true;
}
params->fn_checkpoint_in = argv[i];
} else if (arg == "--checkpoint-out") {
if (++i >= argc) {
*invalid_param = true;
return true;
}
params->fn_checkpoint_out = argv[i];
} else if (arg == "--pattern-fn-it") {
if (++i >= argc) {
*invalid_param = true;
return true;
}
params->pattern_fn_it = argv[i];
} else if (arg == "--fn-latest") {
if (++i >= argc) {
*invalid_param = true;
return true;
}
params->fn_latest = argv[i];
} else if (arg == "--save-every") {
if (++i >= argc) {
*invalid_param = true;
return true;
}
params->save_every = std::stoi(argv[i]);
} else if (arg == "-s" || arg == "--seed") {
if (++i >= argc) {
*invalid_param = true;
return true;
}
params->seed = std::stoi(argv[i]);
} else if (arg == "-c" || arg == "--ctx") {
if (++i >= argc) {
*invalid_param = true;
return true;
}
params->n_ctx = std::stoi(argv[i]);
params->custom_n_ctx = true;
} else if (arg == "-t" || arg == "--threads") {
if (++i >= argc) {
*invalid_param = true;
return true;
}
params->n_threads = std::stoi(argv[i]);
} else if (arg == "-b" || arg == "--batch") {
if (++i >= argc) {
*invalid_param = true;
return true;
}
params->n_batch = std::stoi(argv[i]);
} else if (arg == "--grad-acc") {
if (++i >= argc) {
*invalid_param = true;
return true;
}
params->n_gradient_accumulation = std::max(1, std::stoi(argv[i]));
} else if (arg == "--sample-start") {
if (++i >= argc) {
*invalid_param = true;
return true;
}
params->sample_start = std::string(argv[i]);
} else if (arg == "--escape") {
params->escape = true;
} else if (arg == "--include-sample-start") {
params->include_sample_start = true;
} else if (arg == "--overlapping-samples") {
params->overlapping_samples = true;
} else if (arg == "--fill-with-next-samples") {
params->fill_with_next_samples = true;
} else if (arg == "--separate-with-eos") {
params->separate_with_eos = true;
} else if (arg == "--separate-with-bos") {
params->separate_with_bos = true;
} else if (arg == "--no-separate-with-eos") {
params->separate_with_eos = false;
} else if (arg == "--no-separate-with-bos") {
params->separate_with_bos = false;
} else if (arg == "--sample-random-offsets") {
params->sample_random_offsets = true;
} else if (arg == "--force-reshuffle") {
params->force_reshuffle = true;
} else if (arg == "--no-flash") {
params->use_flash = false;
} else if (arg == "--use-flash") {
params->use_flash = true;
} else if (arg == "--no-checkpointing") {
params->use_checkpointing = false;
} else if (arg == "--use-checkpointing") {
params->use_checkpointing = true;
} else if (arg == "--warmup") {
if (++i >= argc) {
*invalid_param = true;
return true;
}
params->warmup = std::stoi(argv[i]);
} else if (arg == "--cos-decay-steps") {
if (++i >= argc) {
*invalid_param = true;
return true;
}
params->cos_decay_steps = std::stoi(argv[i]);
} else if (arg == "--cos-decay-restart") {
if (++i >= argc) {
*invalid_param = true;
return true;
}
params->cos_decay_restart = std::stof(argv[i]);
} else if (arg == "--cos-decay-min") {
if (++i >= argc) {
*invalid_param = true;
return true;
}
params->cos_decay_min = std::stof(argv[i]);
} else if (arg == "--enable-restart") {
params->enable_restart = true;
} else if (arg == "--disable-restart") {
params->enable_restart = false;
} else if (arg == "--opt-past") {
if (++i >= argc) {
*invalid_param = true;
return true;
}
params->opt_past = std::stoi(argv[i]);
} else if (arg == "--opt-delta") {
if (++i >= argc) {
*invalid_param = true;
return true;
}
params->opt_delta = std::stof(argv[i]);
} else if (arg == "--opt-max-no-improvement") {
if (++i >= argc) {
*invalid_param = true;
return true;
}
params->opt_max_no_improvement = std::stoi(argv[i]);
} else if (arg == "--adam-epsf") {
if (++i >= argc) {
*invalid_param = true;
return true;
}
params->adam_eps_f = std::stof(argv[i]);
} else if (arg == "--epochs") {
if (++i >= argc) {
*invalid_param = true;
return true;
}
params->n_epochs = std::stoi(argv[i]);
} else if (arg == "--adam-iter") {
if (++i >= argc) {
*invalid_param = true;
return true;
}
params->adam_n_iter = std::stoi(argv[i]);
} else if (arg == "--adam-alpha") {
if (++i >= argc) {
*invalid_param = true;
return true;
}
params->adam_alpha = std::stof(argv[i]);
} else if (arg == "--adam-min-alpha") {
if (++i >= argc) {
*invalid_param = true;
return true;
}
params->adam_min_alpha = std::stof(argv[i]);
} else if (arg == "--adam-decay") {
if (++i >= argc) {
*invalid_param = true;
return true;
}
params->adam_decay = std::stof(argv[i]);
} else if (arg == "--adam-decay-min-ndim") {
if (++i >= argc) {
*invalid_param = true;
return true;
}
params->adam_decay_min_ndim = std::stoi(argv[i]);
} else if (arg == "--adam-beta1") {
if (++i >= argc) {
*invalid_param = true;
return true;
}
params->adam_beta1 = std::stof(argv[i]);
} else if (arg == "--adam-beta2") {
if (++i >= argc) {
*invalid_param = true;
return true;
}
params->adam_beta2 = std::stof(argv[i]);
} else if (arg == "--adam-gclip") {
if (++i >= argc) {
*invalid_param = true;
return true;
}
params->adam_gclip = std::stof(argv[i]);
} else if (arg == "-ngl" || arg == "--n-gpu-layers") {
if (++i >= argc) {
*invalid_param = true;
return true;
}
if (llama_supports_gpu_offload()) {
params->n_gpu_layers = std::stoi(argv[i]);
} else {
fprintf(stderr, "warning: not compiled with GPU offload support, --n-gpu-layers option will be ignored\n");
fprintf(stderr, "warning: see main README.md for information on enabling GPU BLAS support\n");
}
} else if (arg == "-h" || arg == "--help") {
params->print_usage = true;
return true;
} else {
return false;
}
return true;
}
void finish_processing_train_args(struct train_params_common * params) {
if (params->escape) {
string_process_escapes(params->sample_start);
}
}
void train_opt_callback(void * vdata, int accum_step, float * sched, bool * cancel) {
struct train_opt_callback_data * data = (struct train_opt_callback_data *) vdata;
struct train_params_common * params = data->params;
struct train_state * train = data->train;
struct ggml_opt_context * opt = train->opt;
int n_batch = params->n_batch;
int n_ctx = params->n_ctx;
if (accum_step == 0) {
// time measurement
int64_t now = ggml_time_ms();
if (now > data->last_time && opt->iter > data->first_iter) {
double dt = (double) (now - data->last_time);
if (data->millis_per_iter == 0.0) {
data->millis_per_iter = dt;
} else {
const double gain = 0.7;
data->millis_per_iter = data->millis_per_iter*(1.0-gain) + dt*gain;
}
}
double remaining_millis = 0.0;
if (data->millis_per_iter > 0.0) {
const int n_iter = params->adam_n_iter;
const int done_iter = opt->iter - data->first_iter;
const int remaining_iter = n_iter - done_iter;
remaining_millis = remaining_iter * data->millis_per_iter;
}
// file saving
const bool save_now = (params->save_every > 0) && (opt->iter - data->last_save_iter >= params->save_every);
if (save_now) {
int new_iters = opt->iter - data->last_save_iter;
train->train_its += new_iters;
train->train_tokens += new_iters * opt->params.n_gradient_accumulation * n_batch * n_ctx;
if (data->save_cb) {
data->save_cb(data->save_data, train);
}
data->last_save_iter = opt->iter;
}
// exclude file saving from time measurement, by measuring last_time after saving
data->last_time = ggml_time_ms();
*sched = learning_schedule(
opt->iter,
params->warmup,
params->cos_decay_steps,
params->adam_alpha,
params->adam_min_alpha,
params->cos_decay_min,
params->cos_decay_restart,
params->enable_restart);
int impr_plot = -(int)(1 + (opt->loss_before - opt->loss_after) * 10.0f + 0.5f);
if (impr_plot > 0) impr_plot = 0;
if (std::isnan(opt->loss_before) || std::isnan(opt->loss_after)) impr_plot = 0;
printf("%s: iter=%6d sample=%zu/%zu sched=%f loss=%f",
__func__, opt->iter, std::min(1+train->shuffle_next_sample, train->shuffle_sample_count), train->shuffle_sample_count,
*sched, opt->loss_after);
if (data->millis_per_iter > 0) {
printf(" dt=");
print_duration(data->millis_per_iter);
printf(" eta=");
print_duration(remaining_millis);
}
float improvement = opt->loss_before - opt->loss_after;
const float plot_scale = 10.0f;
int bar_len = (int)(1 + improvement*plot_scale + 0.5);
printf(" |");
for (int i=0; i<bar_len; ++i) {
printf("-");
}
printf(">");
printf("\n");
}
int64_t used_samples = get_example_targets_batch(
data->lctx,
data->tokens_input,
data->target_probs,
train->shuffle_next_sample,
data->shuffled_samples_offs,
data->shuffled_samples_begin,
data->shuffled_samples_size,
data->samples_count,
data->tokens_data,
data->tokens_size,
params->separate_with_eos,
params->separate_with_bos,
params->fill_with_next_samples,
params->sample_random_offsets);
train->train_samples += used_samples;
train->shuffle_next_sample += used_samples;
if (train->shuffle_next_sample >= train->shuffle_sample_count) {
++train->train_epochs;
printf("%s: reshuffle samples. completed epochs: %llu\n", __func__, (long long unsigned) train->train_epochs);
// note: we may have used some samples from the current shuffling more than once
train->shuffle_rng_state_current = train->shuffle_rng_state_next;
train->shuffle_rng_state_next = shuffle_samples(
train->shuffle_rng_state_current,
data->shuffled_samples_offs,
data->shuffled_samples_begin,
data->shuffled_samples_size,
data->samples_begin,
data->samples_size,
data->samples_count);
train->shuffle_next_sample = 0;
}
const bool last_epoch_reached = (params->n_epochs > 0 && (int64_t) train->train_epochs - data->first_epoch >= params->n_epochs);
if (last_epoch_reached) {
// allow optimization iteration at last epoch to be completed before canceling
if (data->iter_at_last_epoch < 0) {
data->iter_at_last_epoch = opt->iter;
} else if (opt->iter > data->iter_at_last_epoch) {
*cancel = true;
}
}
}
ollama/llm/llama.cpp/common/train.h 0 → 100755
View file @ ff27a817
// Various helper functions and utilities for training
#pragma once
#include <string>
#include <random>
#include <vector>
#include "ggml.h"
#include "llama.h"
#define LLAMA_TRAIN_MAX_NODES 16384
typedef std::string mt19937_state;
struct train_state {
struct ggml_opt_context * opt;
uint64_t train_its;
uint64_t train_samples;
uint64_t train_tokens;
uint64_t train_epochs;
size_t shuffle_samples_hash; // fn, sample_count, *zip(sample_begins, sample_sizes)
mt19937_state shuffle_rng_state_current;
mt19937_state shuffle_rng_state_next;
size_t shuffle_sample_count;
size_t shuffle_next_sample;
};
struct train_params_common {
const char * fn_train_data;
const char * fn_checkpoint_in;
const char * fn_checkpoint_out;
const char * pattern_fn_it;
const char * fn_latest;
bool print_usage;
int save_every;
uint32_t seed;
int n_ctx;
int n_threads;
int n_batch;
int n_gradient_accumulation;
int n_epochs;
int n_gpu_layers;
bool custom_n_ctx;
bool use_flash;
bool use_checkpointing;
std::string sample_start;
bool include_sample_start;
bool escape;
bool overlapping_samples;
bool fill_with_next_samples;
bool separate_with_eos;
bool separate_with_bos;
bool sample_random_offsets;
bool force_reshuffle;
int warmup;
int cos_decay_steps;
float cos_decay_restart;
float cos_decay_min;
bool enable_restart;
int opt_past;
float opt_delta;
int opt_max_no_improvement;
int adam_n_iter;
float adam_alpha;
float adam_min_alpha;
float adam_decay;
int adam_decay_min_ndim;
float adam_beta1;
float adam_beta2;
float adam_gclip;
float adam_eps_f;
};
typedef void (*save_train_files_callback)(void * data, struct train_state * train);
struct train_opt_callback_data {
struct train_params_common * params;
struct train_state * train;
save_train_files_callback save_cb;
void * save_data;
struct llama_context * lctx;
int last_save_iter;
llama_token * tokens_data;
size_t tokens_size;
size_t * samples_begin;
size_t * samples_size;
size_t * shuffled_samples_offs;
size_t * shuffled_samples_begin;
size_t * shuffled_samples_size;
size_t samples_count;
struct ggml_tensor * tokens_input;
struct ggml_tensor * target_probs;
int first_iter;
int first_epoch;
int iter_at_last_epoch;
int64_t last_time;
double millis_per_iter;
};
struct train_state * init_train_state();
void free_train_state(struct train_state * state);
struct train_params_common get_default_train_params_common();
void print_common_train_usage(int /*argc*/, char ** argv, const struct train_params_common * params);
bool consume_common_train_arg(int argc, char ** argv, int * idx, struct train_params_common * params, bool * invalid_param);
void finish_processing_train_args(struct train_params_common * params);
struct random_normal_distribution;
struct random_uniform_distribution;
struct random_normal_distribution * init_random_normal_distribution (int seed, float mean, float std, float min, float max);
struct random_uniform_distribution * init_random_uniform_distribution(int seed, float min, float max);
void free_random_normal_distribution (struct random_normal_distribution * rnd);
void free_random_uniform_distribution(struct random_uniform_distribution * rnd);
struct ggml_tensor * randomize_tensor_normal (struct ggml_tensor * tensor, struct random_normal_distribution * rnd);
struct ggml_tensor * randomize_tensor_uniform(struct ggml_tensor * tensor, struct random_uniform_distribution * rnd);
// generate random float in interval [0,1)
float frand();
float frand_normal (struct random_normal_distribution * rnd);
float frand_uniform(struct random_uniform_distribution * rnd);
int clamp (const int v, const int min, const int max);
float fclamp(const float v, const float min, const float max);
void assert_shape_1d(struct ggml_tensor * tensor, int64_t ne0);
void assert_shape_2d(struct ggml_tensor * tensor, int64_t ne0, int64_t ne1);
void assert_shape_3d(struct ggml_tensor * tensor, int64_t ne0, int64_t ne1, int64_t ne2);
void assert_shape_4d(struct ggml_tensor * tensor, int64_t ne0, int64_t ne1, int64_t ne2, int64_t ne3);
size_t tokenize_file(
struct llama_context * lctx,
const char * filename,
const std::string & sample_start,
bool include_sample_start,
bool overlapping_samples,
unsigned context_length,
std::vector<llama_token> & out_tokens,
std::vector<size_t> & out_samples_begin,
std::vector<size_t> & out_samples_size);
int64_t get_example_targets_batch(
struct llama_context * lctx,
struct ggml_tensor * tokens_input,
struct ggml_tensor * target_probs,
int64_t example_id,
const size_t * samples_offs,
const size_t * samples_begin,
const size_t * samples_size,
size_t samples_count,
const llama_token * train_data,
size_t n_train_data,
bool separate_with_eos,
bool separate_with_bos,
bool fill_with_next_samples,
bool sample_random_offsets);
void mt19937_set_state(std::mt19937& rng, const mt19937_state& rng_state);
mt19937_state mt19937_get_state(const std::mt19937& rng);
mt19937_state mt19937_seed_to_state(unsigned seed);
mt19937_state shuffle_samples(
const mt19937_state & rng_state,
size_t * shuffled_offs,
size_t * shuffled_begins,
size_t * shuffled_sizes,
const size_t * begins,
const size_t * sizes,
size_t count);
size_t hash_combine(size_t h1, size_t h2);
size_t compute_samples_hash(
const char* fn,
const size_t* samples_begin,
const size_t* samples_size,
size_t sample_count);
std::string replace_str(const char * s, const char * needle, const char * replacement);
void print_duration(double milliseconds);
float cosine_decay(
int64_t step,
int64_t decay_steps,
float minimum);
float cosine_decay_restart(
int64_t step,
int64_t decay_steps,
float minimum,
float restart_step_mult);
float learning_schedule(
int64_t step,
int64_t warmup_steps,
int64_t decay_steps,
float learning_rate,
float overall_minimum,
float cos_decay_minimum,
float cos_decay_restart_step_mult,
bool enable_restart);
void copy_tensor_by_name(struct ggml_tensor * dst, struct ggml_context * ctx, const char * name);
void load_opt_context_gguf(struct gguf_context * fctx, struct ggml_context * f_ggml_ctx, struct ggml_opt_context * opt);
void save_opt_context_gguf(struct gguf_context * fctx, struct ggml_opt_context * opt);
bool load_train_state_gguf(struct gguf_context * fctx, struct ggml_context * f_ggml_ctx, struct train_state * train);
void save_train_state_gguf(struct gguf_context * fctx, struct train_state * train);
std::string get_train_filename(const char * filename, const char * pattern_it, const char * latest, int64_t iteration);
void train_opt_callback(void * vdata, int accum_step, float * sched, bool * cancel);
ollama/llm/llama.cpp/convert-hf-to-gguf-update.py 0 → 100755
View file @ ff27a817
#!/usr/bin/env python3
# -*- coding: utf-8 -*-
# This script downloads the tokenizer models of the specified models from Huggingface and
# generates the get_vocab_base_pre() function for convert-hf-to-gguf.py
#
# This is necessary in order to analyze the type of pre-tokenizer used by the model and
# provide the necessary information to llama.cpp via the GGUF header in order to implement
# the same pre-tokenizer.
#
# ref: https://github.com/ggerganov/llama.cpp/pull/6920
#
# Instructions:
#
# - Add a new model to the "models" list
# - Run the script with your huggingface token:
#
# python3 convert-hf-to-gguf-update.py <huggingface_token>
#
# - Copy-paste the generated get_vocab_base_pre() function into convert-hf-to-gguf.py
# - Update llama.cpp with the new pre-tokenizer if necessary
#
# TODO: generate tokenizer tests for llama.cpp
#
import logging
import os
import pathlib
import re
import requests
import sys
import json
from hashlib import sha256
from enum import IntEnum, auto
from transformers import AutoTokenizer
logging.basicConfig(level=logging.DEBUG)
logger = logging.getLogger("convert-hf-to-gguf-update")
sess = requests.Session()
class TOKENIZER_TYPE(IntEnum):
SPM = auto()
BPE = auto()
WPM = auto()
# TODO: this string has to exercise as much pre-tokenizer functionality as possible
# will be updated with time - contributions welcome
chktxt = '\n \n\n \n\n\n \t \t\t \t\n \n \n \n \n🚀 (normal) 😶‍🌫️ (multiple emojis concatenated) ✅ 🦙🦙 3 33 333 3333 33333 333333 3333333 33333333 3.3 3..3 3...3 កាន់តែពិសេសអាច😁 ?我想在apple工作1314151天～ ------======= нещо на Български \'\'\'\'\'\'```````\"\"\"\"......!!!!!!?????? I\'ve been \'told he\'s there, \'RE you sure? \'M not sure I\'ll make it, \'D you like some tea? We\'Ve a\'lL'
if len(sys.argv) == 2:
token = sys.argv[1]
if not token.startswith("hf_"):
logger.info("Huggingface token seems invalid")
logger.info("Usage: python convert-hf-to-gguf-update.py <huggingface_token>")
sys.exit(1)
else:
logger.info("Usage: python convert-hf-to-gguf-update.py <huggingface_token>")
sys.exit(1)
# TODO: add models here, base models preferred
models = [
{"name": "llama-spm", "tokt": TOKENIZER_TYPE.SPM, "repo": "https://huggingface.co/meta-llama/Llama-2-7b-hf", },
{"name": "llama-bpe", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/meta-llama/Meta-Llama-3-8B", },
{"name": "phi-3", "tokt": TOKENIZER_TYPE.SPM, "repo": "https://huggingface.co/microsoft/Phi-3-mini-4k-instruct", },
{"name": "deepseek-llm", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/deepseek-ai/deepseek-llm-7b-base", },
{"name": "deepseek-coder", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/deepseek-ai/deepseek-coder-6.7b-base", },
{"name": "falcon", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/tiiuae/falcon-7b", },
{"name": "bert-bge", "tokt": TOKENIZER_TYPE.WPM, "repo": "https://huggingface.co/BAAI/bge-small-en-v1.5", },
{"name": "mpt", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/mosaicml/mpt-7b", },
{"name": "starcoder", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/bigcode/starcoder2-3b", },
{"name": "gpt-2", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/openai-community/gpt2", },
{"name": "stablelm2", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/stabilityai/stablelm-2-zephyr-1_6b", },
{"name": "refact", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/smallcloudai/Refact-1_6-base", },
{"name": "command-r", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/CohereForAI/c4ai-command-r-v01", },
{"name": "qwen2", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/Qwen/Qwen1.5-7B", },
{"name": "olmo", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/allenai/OLMo-1.7-7B-hf", },
{"name": "dbrx", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/databricks/dbrx-base", },
{"name": "jina-v2-en", "tokt": TOKENIZER_TYPE.WPM, "repo": "https://huggingface.co/jinaai/jina-embeddings-v2-base-en", }, # WPM!
{"name": "jina-v2-es", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/jinaai/jina-embeddings-v2-base-es", },
{"name": "jina-v2-de", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/jinaai/jina-embeddings-v2-base-de", },
{"name": "smaug-bpe", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/abacusai/Smaug-Llama-3-70B-Instruct", },
{"name": "poro-chat", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/LumiOpen/Poro-34B-chat", },
{"name": "jina-v2-code", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/jinaai/jina-embeddings-v2-base-code", },
]
def download_file_with_auth(url, token, save_path):
headers = {"Authorization": f"Bearer {token}"}
response = sess.get(url, headers=headers)
response.raise_for_status()
os.makedirs(os.path.dirname(save_path), exist_ok=True)
with open(save_path, 'wb') as f:
f.write(response.content)
logger.info(f"File {save_path} downloaded successfully")
def download_model(model):
name = model["name"]
repo = model["repo"]
tokt = model["tokt"]
os.makedirs(f"models/tokenizers/{name}", exist_ok=True)
files = ["config.json", "tokenizer.json", "tokenizer_config.json"]
if tokt == TOKENIZER_TYPE.SPM:
files.append("tokenizer.model")
for file in files:
save_path = f"models/tokenizers/{name}/{file}"
if os.path.isfile(save_path):
logger.info(f"{name}: File {save_path} already exists - skipping")
continue
download_file_with_auth(f"{repo}/resolve/main/{file}", token, save_path)
for model in models:
try:
download_model(model)
except Exception as e:
logger.error(f"Failed to download model {model['name']}. Error: {e}")
# generate the source code for the convert-hf-to-gguf.py:get_vocab_base_pre() function:
src_ifs = ""
for model in models:
name = model["name"]
tokt = model["tokt"]
if tokt == TOKENIZER_TYPE.SPM:
continue
# Skip if the tokenizer folder does not exist or there are other download issues previously
if not os.path.exists(f"models/tokenizers/{name}"):
logger.warning(f"Directory for tokenizer {name} not found. Skipping...")
continue
# create the tokenizer
try:
tokenizer = AutoTokenizer.from_pretrained(f"models/tokenizers/{name}")
except OSError as e:
logger.error(f"Error loading tokenizer for model {name}. The model may not exist or is not accessible with the provided token. Error: {e}")
continue # Skip to the next model if the tokenizer can't be loaded
chktok = tokenizer.encode(chktxt)
chkhsh = sha256(str(chktok).encode()).hexdigest()
logger.info(f"model: {name}")
logger.info(f"tokt: {tokt}")
logger.info(f"repo: {model['repo']}")
logger.info(f"chktok: {chktok}")
logger.info(f"chkhsh: {chkhsh}")
# print the "pre_tokenizer" content from the tokenizer.json
with open(f"models/tokenizers/{name}/tokenizer.json", "r", encoding="utf-8") as f:
cfg = json.load(f)
normalizer = cfg["normalizer"]
logger.info("normalizer: " + json.dumps(normalizer, indent=4))
pre_tokenizer = cfg["pre_tokenizer"]
logger.info("pre_tokenizer: " + json.dumps(pre_tokenizer, indent=4))
if "ignore_merges" in cfg["model"]:
logger.info("ignore_merges: " + json.dumps(cfg["model"]["ignore_merges"], indent=4))
logger.info("")
src_ifs += f" if chkhsh == \"{chkhsh}\":\n"
src_ifs += f" # ref: {model['repo']}\n"
src_ifs += f" res = \"{name}\"\n"
src_func = f"""
def get_vocab_base_pre(self, tokenizer) -> str:
# encoding this string and hashing the resulting tokens would (hopefully) give us a unique identifier that
# is specific for the BPE pre-tokenizer used by the model
# we will use this unique identifier to write a "tokenizer.ggml.pre" entry in the GGUF file which we can
# use in llama.cpp to implement the same pre-tokenizer
chktxt = {repr(chktxt)}
chktok = tokenizer.encode(chktxt)
chkhsh = sha256(str(chktok).encode()).hexdigest()
logger.debug(f"chktok: {{chktok}}")
logger.debug(f"chkhsh: {{chkhsh}}")
res = None
# NOTE: if you get an error here, you need to update the convert-hf-to-gguf-update.py script
# or pull the latest version of the model from Huggingface
# don't edit the hashes manually!
{src_ifs}
if res is None:
logger.warning("\\n")
logger.warning("**************************************************************************************")
logger.warning("** WARNING: The BPE pre-tokenizer was not recognized!")
logger.warning("** There are 2 possible reasons for this:")
logger.warning("** - the model has not been added to convert-hf-to-gguf-update.py yet")
logger.warning("** - the pre-tokenization config has changed upstream")
logger.warning("** Check your model files and convert-hf-to-gguf-update.py and update them accordingly.")
logger.warning("** ref: https://github.com/ggerganov/llama.cpp/pull/6920")
logger.warning("**")
logger.warning(f"** chkhsh: {{chkhsh}}")
logger.warning("**************************************************************************************")
logger.warning("\\n")
raise NotImplementedError("BPE pre-tokenizer was not recognized - update get_vocab_base_pre()")
logger.debug(f"tokenizer.ggml.pre: {{repr(res)}}")
logger.debug(f"chkhsh: {{chkhsh}}")
return res
"""
convert_py_pth = pathlib.Path("convert-hf-to-gguf.py")
convert_py = convert_py_pth.read_text()
convert_py = re.sub(
r"(# Marker: Start get_vocab_base_pre)(.+?)( +# Marker: End get_vocab_base_pre)",
lambda m: m.group(1) + src_func + m.group(3),
convert_py,
flags=re.DOTALL | re.MULTILINE,
)
convert_py_pth.write_text(convert_py)
logger.info("+++ convert-hf-to-gguf.py was updated")
# generate tests for each tokenizer model
tests = [
"ied 4 ½ months",
"Führer",
"",
" ",
" ",
" ",
"\t",
"\n",
"\n\n",
"\n\n\n",
"\t\n",
"Hello world",
" Hello world",
"Hello World",
" Hello World",
" Hello World!",
"Hello, world!",
" Hello, world!",
" this is 🦙.cpp",
"w048 7tuijk dsdfhu",
"нещо на Български",
"កាន់តែពិសេសអាចខលចេញ",
"🚀 (normal) 😶‍🌫️ (multiple emojis concatenated) ✅ (only emoji that has its own token)",
"Hello",
" Hello",
" Hello",
" Hello",
" Hello",
" Hello\n Hello",
" (",
"\n =",
"' era",
"Hello, y'all! How are you 😁 ?我想在apple工作1314151天～",
"3",
"33",
"333",
"3333",
"33333",
"333333",
"3333333",
"33333333",
"333333333",
# "Cửa Việt", # llama-bpe fails on this
chktxt,
]
# write the tests to ./models/ggml-vocab-{name}.gguf.inp
# the format is:
#
# test0
# __ggml_vocab_test__
# test1
# __ggml_vocab_test__
# ...
#
# with each model, encode all tests and write the results in ./models/ggml-vocab-{name}.gguf.out
# for each test, write the resulting tokens on a separate line
for model in models:
name = model["name"]
tokt = model["tokt"]
# Skip if the tokenizer folder does not exist or there are other download issues previously
if not os.path.exists(f"models/tokenizers/{name}"):
logger.warning(f"Directory for tokenizer {name} not found. Skipping...")
continue
# create the tokenizer
try:
tokenizer = AutoTokenizer.from_pretrained(f"models/tokenizers/{name}")
except OSError as e:
logger.error(f"Failed to load tokenizer for model {name}. Error: {e}")
continue # Skip this model and continue with the next one in the loop
with open(f"models/ggml-vocab-{name}.gguf.inp", "w", encoding="utf-8") as f:
for text in tests:
f.write(f"{text}")
f.write("\n__ggml_vocab_test__\n")
with open(f"models/ggml-vocab-{name}.gguf.out", "w") as f:
for text in tests:
res = tokenizer.encode(text, add_special_tokens=False)
for r in res:
f.write(f" {r}")
f.write("\n")
logger.info(f"Tests for {name} written in ./models/ggml-vocab-{name}.gguf.*")
# generate commands for creating vocab files
logger.info("\nRun the following commands to generate the vocab files for testing:\n")
for model in models:
name = model["name"]
print(f"python3 convert-hf-to-gguf.py models/tokenizers/{name}/ --outfile models/ggml-vocab-{name}.gguf --vocab-only") # noqa: NP100
logger.info("\n")
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment