llama.cpp/examples/common.h

141 lines
5.6 KiB
C
Raw Normal View History

2023-03-10 19:40:58 +01:00
// Various helper functions and utilities
#pragma once
#include "llama.h"
2023-03-10 19:40:58 +01:00
#include <string>
#include <vector>
#include <random>
#include <thread>
#include <unordered_map>
#include <tuple>
2023-03-10 19:40:58 +01:00
#if !defined (_WIN32)
#include <stdio.h>
#include <termios.h>
#endif
2023-03-10 19:40:58 +01:00
//
// CLI argument parsing
//
int32_t get_num_physical_cores();
2023-03-10 19:40:58 +01:00
struct gpt_params {
uint32_t seed = -1; // RNG seed
int32_t n_threads = get_num_physical_cores();
int32_t n_predict = -1; // new tokens to predict
int32_t n_ctx = 512; // context size
int32_t n_batch = 512; // batch size for prompt processing (must be >=32 to use BLAS)
int32_t n_keep = 0; // number of tokens to keep from initial prompt
int32_t n_gpu_layers = 0; // number of layers to store in VRAM
int32_t main_gpu = 0; // the GPU that is used for scratch and small tensors
float tensor_split[LLAMA_MAX_DEVICES] = {0}; // how split tensors should be distributed across GPUs
int32_t n_probs = 0; // if greater than 0, output the probabilities of top n_probs tokens.
2023-03-17 20:46:46 +01:00
2023-03-10 19:40:58 +01:00
// sampling parameters
std::unordered_map<llama_token, float> logit_bias; // logit bias for specific tokens
int32_t top_k = 40; // <= 0 to use vocab size
float top_p = 0.95f; // 1.0 = disabled
float tfs_z = 1.00f; // 1.0 = disabled
float typical_p = 1.00f; // 1.0 = disabled
float temp = 0.80f; // 1.0 = disabled
float repeat_penalty = 1.10f; // 1.0 = disabled
int32_t repeat_last_n = 64; // last n tokens to penalize (0 = disable penalty, -1 = context size)
float frequency_penalty = 0.00f; // 0.0 = disabled
float presence_penalty = 0.00f; // 0.0 = disabled
int mirostat = 0; // 0 = disabled, 1 = mirostat, 2 = mirostat 2.0
float mirostat_tau = 5.00f; // target entropy
float mirostat_eta = 0.10f; // learning rate
2023-03-10 19:40:58 +01:00
2023-05-19 19:14:51 +02:00
std::string model = "models/7B/ggml-model.bin"; // model path
std::string model_alias = "unknown"; // model alias
2023-05-19 19:14:51 +02:00
std::string prompt = "";
std::string path_prompt_cache = ""; // path to file for saving/loading prompt eval state
std::string input_prefix = ""; // string to prefix user inputs with
std::string input_suffix = ""; // string to suffix user inputs with
std::vector<std::string> antiprompt; // string upon seeing which more user input is prompted
2023-04-17 17:28:55 +02:00
std::string lora_adapter = ""; // lora adapter path
2023-05-19 19:14:51 +02:00
std::string lora_base = ""; // base model path for the lora adapter
2023-04-17 17:28:55 +02:00
bool low_vram = false; // if true, reduce VRAM usage at the cost of performance
bool memory_f16 = true; // use f16 instead of f32 for memory kv
bool random_prompt = false; // do not randomize prompt if none provided
bool use_color = false; // use color to distinguish generations and inputs
bool interactive = false; // interactive mode
bool prompt_cache_all = false; // save user input and generations to prompt cache
bool prompt_cache_ro = false; // open the prompt cache read-only and do not update it
bool embedding = false; // get only sentence embedding
bool interactive_first = false; // wait for user input immediately
bool multiline_input = false; // reverse the usage of `\`
bool instruct = false; // instruction mode (used for Alpaca models)
bool penalize_nl = true; // consider newlines as a repeatable token
bool perplexity = false; // compute perplexity over the prompt
Rewrite loading code to try to satisfy everyone: - Support all three formats (ggml, ggmf, ggjt). (However, I didn't include the hack needed to support GPT4All files without conversion. Those can still be used after converting them with convert.py from my other PR.) - Support both mmap and read (mmap is used by default, but can be disabled with `--no-mmap`, and is automatically disabled for pre-ggjt files or on platforms where mmap is not supported). - Support multi-file models like before, but automatically determine the number of parts rather than requiring `--n_parts`. - Improve validation and error checking. - Stop using the per-file type field (f16) entirely in favor of just relying on the per-tensor type/size fields. This has no immediate benefit, but makes it easier to experiment with different formats, and should make it easier to support the new GPTQ-for-LLaMa models in the future (I have some work in progress on that front). - Support VirtualLock on Windows (using the same `--mlock` option as on Unix). - Indicate loading progress when using mmap + mlock. (Which led me to the interesting observation that on my Linux machine, with a warm file cache, mlock actually takes some time, whereas mmap without mlock starts almost instantly...) - To help implement this, move mlock support from ggml to the loading code. - madvise/PrefetchVirtualMemory support (based on #740) - Switch from ifstream to the `fopen` family of functions to avoid unnecessary copying and, when mmap is enabled, allow reusing the same file descriptor for both metadata reads and mmap (whereas the existing implementation opens the file a second time to mmap). - Quantization now produces a single-file output even with multi-file inputs (not really a feature as much as 'it was easier this way'). Implementation notes: I tried to factor the code into more discrete pieces than before. Regarding code style: I tried to follow the code style, but I'm naughty and used a few advanced C++ features repeatedly: - Destructors to make it easier to ensure everything gets cleaned up. - Exceptions. I don't even usually use exceptions when writing C++, and I can remove them if desired... but here they make the loading code much more succinct while still properly handling a variety of errors, ranging from API calls failing to integer overflow and allocation failure. The exceptions are converted to error codes at the API boundary.) Co-authored-by: Pavol Rusnak <pavol@rusnak.io> (for the bit I copied from #740)
2023-04-08 21:24:37 +02:00
bool use_mmap = true; // use mmap for faster loads
bool use_mlock = false; // use mlock to keep model in memory
bool mem_test = false; // compute maximum memory usage
bool numa = false; // attempt optimizations that help on some NUMA systems
llama : Metal inference (#1642) * mtl : export the LLaMA computation graph * ci : disable temporary * mtl : adapt the MNIST example as starter * mtl : no need for mtl-export tool, add cli arg for main instead * mtl : export just a small part of the graph for now to make it easier * mtl : move MSL code into separate file for easy editing * mtl : initial get_rows_q4_0 kernel * mtl : confirmed get_rows_q4_0 is working correctly * mtl : add rms_norm kernel + confirm working * mtl : add mul kernel + confirm working * mtl : initial mul_mat Q4 kernel (wrong results) * mtl : mul_mat fixes (still wrong) * mtl : another mul_mat Q4 (still does not work) * mtl : working mul_mat q4 * ggml : fix handling of "view" ops in ggml_graph_import() * mtl : add rope kernel * mtl : add reshape and transpose handling * ggml : store offset as opt arg for ggml_view_xd() operators * mtl : add cpy kernel + handle view ops * mtl : confirm f16 x f32 attention mul mat * mtl : add scale kernel * mtl : add diag_mask_inf kernel * mtl : fix soft_max kernel * ggml : update ggml_nbytes() to handle non-contiguous tensors * mtl : verify V tensor contents * mtl : add f32 -> f32 cpy kernel * mtl : add silu kernel * mtl : add non-broadcast mul kernel * mtl : full GPU inference of the computation graph * mtl : optimize rms_norm and soft_max kernels * mtl : add f16 mat x f32 vec multiplication kernel * mtl : fix bug in f16 x f32 mul mat + speed-up computation * mtl : faster mul_mat_q4_0_f32 kernel * mtl : fix kernel signature + roll inner loop * mtl : more threads for rms_norm + better timing * mtl : remove printfs from inner loop * mtl : simplify implementation * mtl : add save/load vocab to ggml file * mtl : plug Metal inference into llama.cpp (very quick-n-dirty) * mtl : make it work with main example Lots of hacks but at least now it generates text * mtl : preparing for merge * mtl : clean-up ggml mtl interface + suport scratch / inplace * mtl : remove temp / debug code * metal : final refactoring and simplification * Revert "ci : disable temporary" This reverts commit 98c267fc77fe811082f672538fc91bcfc9072d63. * metal : add comments * metal : clean-up stuff, fix typos * readme : add Metal instructions * readme : add example for main
2023-06-04 22:34:30 +02:00
bool export_cgraph = false; // export the computation graph
bool verbose_prompt = false; // print prompt tokens before generation
2023-03-10 19:40:58 +01:00
};
bool gpt_params_parse(int argc, char ** argv, gpt_params & params);
void gpt_print_usage(int argc, char ** argv, const gpt_params & params);
2023-03-10 19:40:58 +01:00
std::string gpt_random_prompt(std::mt19937 & rng);
//
// Vocab utils
//
std::vector<llama_token> llama_tokenize(struct llama_context * ctx, const std::string & text, bool add_bos);
//
// Model utils
//
std::tuple<struct llama_model *, struct llama_context *> llama_init_from_gpt_params(const gpt_params & params);
//
// Console utils
//
#define ANSI_COLOR_RED "\x1b[31m"
#define ANSI_COLOR_GREEN "\x1b[32m"
#define ANSI_COLOR_YELLOW "\x1b[33m"
#define ANSI_COLOR_BLUE "\x1b[34m"
#define ANSI_COLOR_MAGENTA "\x1b[35m"
#define ANSI_COLOR_CYAN "\x1b[36m"
#define ANSI_COLOR_RESET "\x1b[0m"
#define ANSI_BOLD "\x1b[1m"
enum console_color_t {
CONSOLE_COLOR_DEFAULT=0,
CONSOLE_COLOR_PROMPT,
CONSOLE_COLOR_USER_INPUT,
CONSOLE_COLOR_ERROR
};
struct console_state {
bool multiline_input = false;
bool use_color = false;
console_color_t color = CONSOLE_COLOR_DEFAULT;
FILE* out = stdout;
#if defined (_WIN32)
void* hConsole;
#else
FILE* tty = nullptr;
termios prev_state;
#endif
};
void console_init(console_state & con_st);
void console_cleanup(console_state & con_st);
void console_set_color(console_state & con_st, console_color_t color);
bool console_readline(console_state & con_st, std::string & line);