mirror of
https://github.com/LostRuins/koboldcpp.git
synced 2025-09-11 01:24:36 +00:00
fixed compile errors, made mmap automatic when lora is selected, added updated quantizers and quantization handling for gpt neox gpt 2 and gptj
This commit is contained in:
Concedo
parent
3962eb39c7
commit
59fb174678
11 changed files with 297 additions and 590 deletions
|
|
@ -1,6 +1,7 @@
|
|||
#include "ggml.h"
|
||||
|
||||
#include "otherarch/utils.h"
|
||||
#include "common-ggml.h"
|
||||
|
||||
#include <cassert>
|
||||
#include <cmath>
|
||||
|
|
@ -23,20 +24,7 @@ struct gpt2_hparams {
|
|||
};
|
||||
|
||||
// quantize a model
|
||||
bool gpt2_model_quantize(const std::string & fname_inp, const std::string & fname_out, int itype) {
|
||||
ggml_type type = GGML_TYPE_Q4_1;
|
||||
|
||||
switch (itype) {
|
||||
case 2: type = GGML_TYPE_Q4_0; break;
|
||||
case 3: type = GGML_TYPE_Q4_1; break;
|
||||
default: fprintf(stderr, "%s: invalid quantization type %d\n", __func__, itype); return 1;
|
||||
};
|
||||
|
||||
if (type != GGML_TYPE_Q4_0 && type != GGML_TYPE_Q4_1) {
|
||||
fprintf(stderr, "%s: invalid quantization type %d\n", __func__, type);
|
||||
return false;
|
||||
}
|
||||
|
||||
bool gpt2_model_quantize(const std::string & fname_inp, const std::string & fname_out, ggml_mtype mtype) {
|
||||
gpt_vocab vocab;
|
||||
|
||||
printf("%s: loading model from '%s'\n", __func__, fname_inp.c_str());
|
||||
|
|
@ -88,7 +76,7 @@ bool gpt2_model_quantize(const std::string & fname_inp, const std::string & fnam
|
|||
fout.write((char *) &hparams.n_embd, sizeof(hparams.n_embd));
|
||||
fout.write((char *) &hparams.n_head, sizeof(hparams.n_head));
|
||||
fout.write((char *) &hparams.n_layer, sizeof(hparams.n_layer));
|
||||
fout.write((char *) &itype, sizeof(hparams.f16));
|
||||
fout.write((char *) &mtype, sizeof(hparams.f16));
|
||||
}
|
||||
|
||||
// load vocab
|
||||
|
|
@ -118,158 +106,19 @@ bool gpt2_model_quantize(const std::string & fname_inp, const std::string & fnam
|
|||
}
|
||||
}
|
||||
|
||||
// load weights
|
||||
{
|
||||
size_t total_size_org = 0;
|
||||
size_t total_size_new = 0;
|
||||
// regexes of tensor names to be quantized
|
||||
const std::vector<std::string> to_quant = {
|
||||
"model/wte",
|
||||
"model/lm_head",
|
||||
"model/h.*/attn/c_attn/w",
|
||||
"model/h.*/attn/c_proj/w",
|
||||
"model/h.*/mlp/c_fc/w",
|
||||
"model/h.*/mlp/c_proj/w",
|
||||
};
|
||||
|
||||
std::vector<float> work;
|
||||
|
||||
std::vector<uint8_t> data_u8;
|
||||
std::vector<ggml_fp16_t> data_f16;
|
||||
std::vector<float> data_f32;
|
||||
|
||||
std::vector<int64_t> hist_all(1 << 4, 0);
|
||||
|
||||
while (true) {
|
||||
int32_t n_dims;
|
||||
int32_t length;
|
||||
int32_t ftype;
|
||||
|
||||
finp.read(reinterpret_cast<char *>(&n_dims), sizeof(n_dims));
|
||||
finp.read(reinterpret_cast<char *>(&length), sizeof(length));
|
||||
finp.read(reinterpret_cast<char *>(&ftype), sizeof(ftype));
|
||||
|
||||
if (finp.eof()) {
|
||||
break;
|
||||
}
|
||||
|
||||
int32_t nelements = 1;
|
||||
int32_t ne[2] = { 1, 1 };
|
||||
for (int i = 0; i < n_dims; ++i) {
|
||||
finp.read (reinterpret_cast<char *>(&ne[i]), sizeof(ne[i]));
|
||||
nelements *= ne[i];
|
||||
}
|
||||
|
||||
std::string name(length, 0);
|
||||
finp.read (&name[0], length);
|
||||
|
||||
{
|
||||
static const char * ftype_str[] = { "f32", "f16", "q4_0", "q4_1", };
|
||||
printf("%24s - [%5d, %5d], type = %6s ", name.data(), ne[0], ne[1], ftype_str[ftype]);
|
||||
}
|
||||
|
||||
// regexes of tensor names to be quantized
|
||||
const std::vector<std::string> k_names = {
|
||||
"model/wte",
|
||||
"model/lm_head",
|
||||
"model/h.*/attn/c_attn/w",
|
||||
"model/h.*/attn/c_proj/w",
|
||||
"model/h.*/mlp/c_fc/w",
|
||||
"model/h.*/mlp/c_proj/w",
|
||||
};
|
||||
|
||||
bool quantize = false;
|
||||
for (const auto & s : k_names) {
|
||||
if (std::regex_match(name, std::regex(s))) {
|
||||
quantize = true;
|
||||
break;
|
||||
}
|
||||
}
|
||||
|
||||
if (quantize) {
|
||||
if (ftype != 0 && ftype != 1) {
|
||||
fprintf(stderr, "%s: unsupported ftype %d for integer quantization\n", __func__, ftype);
|
||||
return false;
|
||||
}
|
||||
|
||||
if (ftype == 1) {
|
||||
data_f16.resize(nelements);
|
||||
finp.read(reinterpret_cast<char *>(data_f16.data()), nelements * sizeof(ggml_fp16_t));
|
||||
data_f32.resize(nelements);
|
||||
for (int i = 0; i < nelements; ++i) {
|
||||
data_f32[i] = ggml_fp16_to_fp32(data_f16[i]);
|
||||
}
|
||||
} else {
|
||||
data_f32.resize(nelements);
|
||||
finp.read(reinterpret_cast<char *>(data_f32.data()), nelements * sizeof(float));
|
||||
}
|
||||
|
||||
ftype = itype;
|
||||
} else {
|
||||
const int bpe = (ftype == 0) ? sizeof(float) : sizeof(uint16_t);
|
||||
|
||||
data_u8.resize(nelements*bpe);
|
||||
finp.read(reinterpret_cast<char *>(data_u8.data()), nelements * bpe);
|
||||
}
|
||||
|
||||
fout.write(reinterpret_cast<char *>(&n_dims), sizeof(n_dims));
|
||||
fout.write(reinterpret_cast<char *>(&length), sizeof(length));
|
||||
fout.write(reinterpret_cast<char *>(&ftype), sizeof(ftype));
|
||||
for (int i = 0; i < n_dims; ++i) {
|
||||
fout.write(reinterpret_cast<char *>(&ne[i]), sizeof(ne[i]));
|
||||
}
|
||||
fout.write(&name[0], length);
|
||||
|
||||
if (quantize) {
|
||||
printf("quantizing .. ");
|
||||
work.resize(nelements); // for quantization
|
||||
|
||||
size_t cur_size = 0;
|
||||
std::vector<int64_t> hist_cur(1 << 4, 0);
|
||||
|
||||
switch (type) {
|
||||
case GGML_TYPE_Q4_0:
|
||||
{
|
||||
cur_size = ggml_quantize_q4_0(data_f32.data(), work.data(), nelements, ne[0], hist_cur.data());
|
||||
} break;
|
||||
case GGML_TYPE_Q4_1:
|
||||
{
|
||||
cur_size = ggml_quantize_q4_1(data_f32.data(), work.data(), nelements, ne[0], hist_cur.data());
|
||||
} break;
|
||||
default:
|
||||
{
|
||||
fprintf(stderr, "%s: unsupported quantization type %d\n", __func__, type);
|
||||
return false;
|
||||
}
|
||||
}
|
||||
|
||||
fout.write(reinterpret_cast<char *>(work.data()), cur_size);
|
||||
total_size_new += cur_size;
|
||||
|
||||
printf("size = %8.2f MB -> %8.2f MB | hist: ", nelements * sizeof(float)/1024.0/1024.0, cur_size/1024.0/1024.0);
|
||||
for (int i = 0; i < hist_cur.size(); ++i) {
|
||||
hist_all[i] += hist_cur[i];
|
||||
}
|
||||
|
||||
for (int i = 0; i < hist_cur.size(); ++i) {
|
||||
printf("%5.3f ", hist_cur[i] / (float)nelements);
|
||||
}
|
||||
printf("\n");
|
||||
} else {
|
||||
printf("size = %8.3f MB\n", data_u8.size()/1024.0/1024.0);
|
||||
fout.write(reinterpret_cast<char *>(data_u8.data()), data_u8.size());
|
||||
total_size_new += data_u8.size();
|
||||
}
|
||||
|
||||
total_size_org += nelements * sizeof(float);
|
||||
}
|
||||
|
||||
printf("%s: model size = %8.2f MB\n", __func__, total_size_org/1024.0/1024.0);
|
||||
printf("%s: quant size = %8.2f MB\n", __func__, total_size_new/1024.0/1024.0);
|
||||
|
||||
{
|
||||
int64_t sum_all = 0;
|
||||
for (int i = 0; i < hist_all.size(); ++i) {
|
||||
sum_all += hist_all[i];
|
||||
}
|
||||
|
||||
printf("%s: hist: ", __func__);
|
||||
for (int i = 0; i < hist_all.size(); ++i) {
|
||||
printf("%5.3f ", hist_all[i] / (float)sum_all);
|
||||
}
|
||||
printf("\n");
|
||||
}
|
||||
if (!ggml_common_quantize_0(finp, fout, mtype, to_quant, {})) {
|
||||
fprintf(stderr, "%s: failed to quantize model '%s'\n", __func__, fname_inp.c_str());
|
||||
return false;
|
||||
}
|
||||
|
||||
finp.close();
|
||||
|
|
@ -287,6 +136,8 @@ int main(int argc, char ** argv) {
|
|||
fprintf(stderr, "usage: %s model-f32.bin model-quant.bin type\n", argv[0]);
|
||||
fprintf(stderr, " type = 2 - q4_0\n");
|
||||
fprintf(stderr, " type = 3 - q4_1\n");
|
||||
fprintf(stderr, " type = 5 - q4_2\n");
|
||||
fprintf(stderr, " type = 6 - q4_3\n");
|
||||
return 1;
|
||||
}
|
||||
|
||||
|
|
@ -300,7 +151,7 @@ int main(int argc, char ** argv) {
|
|||
const std::string fname_inp = argv[1];
|
||||
const std::string fname_out = argv[2];
|
||||
|
||||
const int itype = atoi(argv[3]);
|
||||
const int mtype = atoi(argv[3]);
|
||||
|
||||
const int64_t t_main_start_us = ggml_time_us();
|
||||
|
||||
|
|
@ -310,7 +161,7 @@ int main(int argc, char ** argv) {
|
|||
{
|
||||
const int64_t t_start_us = ggml_time_us();
|
||||
|
||||
if (!gpt2_model_quantize(fname_inp, fname_out, itype)) {
|
||||
if (!gpt2_model_quantize(fname_inp, fname_out, ggml_mtype(mtype))) {
|
||||
fprintf(stderr, "%s: failed to quantize model from '%s'\n", __func__, fname_inp.c_str());
|
||||
return 1;
|
||||
}
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue