vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Add tokenizer test + revert to C++11 (#355)

Browse source 
* Add test-tokenizer-0 to do a few tokenizations - feel free to expand
* Added option to convert-pth-to-ggml.py script to dump just the vocabulary
* Added ./models/ggml-vocab.bin containing just LLaMA vocab data (used for tests)
* Added utility to load vocabulary file from previous point (temporary implementation)
* Avoid using std::string_view and drop back to C++11 (hope I didn't break something)
* Rename gpt_vocab -> llama_vocab
* All CMake binaries go into ./bin/ now
This commit is contained in:
Georgi Gerganov 2023-03-21 17:29:41 +02:00 • committed by GitHub
parent 2e664f1ff4
commit eb34620aec
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
11 changed files with 249 additions and 148 deletions
Download patch file Download diff file Expand all files Collapse all files

28
main.cpp
View file

@ -90,7 +90,7 @@ struct llama_model {
};
// load the model's weights from a file
bool llama_model_load(const std::string & fname, llama_model & model, gpt_vocab & vocab, int n_ctx, ggml_type memory_type = GGML_TYPE_F32) {
bool llama_model_load(const std::string & fname, llama_model & model, llama_vocab & vocab, int n_ctx, ggml_type memory_type = GGML_TYPE_F32) {
fprintf(stderr, "%s: loading model from '%s' - please wait ...\n", __func__, fname.c_str());
std::vector<char> f_buf(1024*1024);
@ -544,9 +544,9 @@ bool llama_eval(
const llama_model & model,
const int n_threads,
const int n_past,
const std::vector<gpt_vocab::id> & embd_inp,
std::vector<float> & embd_w,
size_t & mem_per_token) {
const std::vector<llama_vocab::id> & embd_inp,
std::vector<float> & embd_w,
size_t & mem_per_token) {
const int N = embd_inp.size();
const auto & hparams = model.hparams;
@ -832,7 +832,7 @@ int main(int argc, char ** argv) {
int64_t t_load_us = 0;
gpt_vocab vocab;
llama_vocab vocab;
llama_model model;
// load the model
@ -864,13 +864,13 @@ int main(int argc, char ** argv) {
// Add a space in front of the first character to match OG llama tokenizer behavior
params.prompt.insert(0, 1, ' ');
// tokenize the prompt
std::vector<gpt_vocab::id> embd_inp = ::llama_tokenize(vocab, params.prompt, true);
std::vector<llama_vocab::id> embd_inp = ::llama_tokenize(vocab, params.prompt, true);
params.n_predict = std::min(params.n_predict, model.hparams.n_ctx - (int) embd_inp.size());
// prefix & suffix for instruct mode
const std::vector<gpt_vocab::id> inp_pfx = ::llama_tokenize(vocab, "\n\n### Instruction:\n\n", true);
const std::vector<gpt_vocab::id> inp_sfx = ::llama_tokenize(vocab, "\n\n### Response:\n\n", false);
const std::vector<llama_vocab::id> inp_pfx = ::llama_tokenize(vocab, "\n\n### Instruction:\n\n", true);
const std::vector<llama_vocab::id> inp_sfx = ::llama_tokenize(vocab, "\n\n### Response:\n\n", false);
// in instruct mode, we inject a prefix and a suffix to each input by the user
if (params.instruct) {
@ -879,8 +879,8 @@ int main(int argc, char ** argv) {
}
// tokenize the reverse prompt
std::vector<std::vector<gpt_vocab::id>> antipromptv_inp;
std::vector<std::vector<llama_vocab::id>> antipromptv_inp;
for (auto antiprompt : params.antiprompt) {
antipromptv_inp.push_back(::llama_tokenize(vocab, antiprompt, false));
}
@ -925,14 +925,14 @@ int main(int argc, char ** argv) {
fprintf(stderr, "sampling parameters: temp = %f, top_k = %d, top_p = %f, repeat_last_n = %i, repeat_penalty = %f\n", params.temp, params.top_k, params.top_p, params.repeat_last_n, params.repeat_penalty);
fprintf(stderr, "\n\n");
std::vector<gpt_vocab::id> embd;
std::vector<llama_vocab::id> embd;
// determine the required inference memory per token:
size_t mem_per_token = 0;
llama_eval(model, params.n_threads, 0, { 0, 1, 2, 3 }, logits, mem_per_token);
int last_n_size = params.repeat_last_n;
std::vector<gpt_vocab::id> last_n_tokens(last_n_size);
std::vector<llama_vocab::id> last_n_tokens(last_n_size);
std::fill(last_n_tokens.begin(), last_n_tokens.end(), 0);
if (params.interactive) {
@ -980,7 +980,7 @@ int main(int argc, char ** argv) {
const int n_vocab = model.hparams.n_vocab;
gpt_vocab::id id = 0;
llama_vocab::id id = 0;
{
const int64_t t_start_sample_us = ggml_time_us();
@ -1066,7 +1066,7 @@ int main(int argc, char ** argv) {
} while (another_line);
if (params.use_color) printf(ANSI_COLOR_RESET);
std::vector<gpt_vocab::id> line_inp = ::llama_tokenize(vocab, buffer, false);
std::vector<llama_vocab::id> line_inp = ::llama_tokenize(vocab, buffer, false);
embd_inp.insert(embd_inp.end(), line_inp.begin(), line_inp.end());
if (params.instruct) {