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

Update main.cpp

Browse source
This commit is contained in:
Anujsharma002 2023-03-15 23:10:29 +05:30 • committed by GitHub
parent 47857e564c
commit b1a4a2a02a
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
1 changed files with 17 additions and 17 deletions
Download patch file Download diff file Expand all files Collapse all files

30
main.cpp
View file

@ -150,7 +150,7 @@ bool llama_model_load(const std::string & fname, llama_model & model, gpt_vocab
} }
std::string word; std::string word;
for (int i = 0; i < n_vocab; i++) { for (register int i = 0; i < n_vocab; i++) {
uint32_t len; uint32_t len;
fin.read((char *) &len, sizeof(len)); fin.read((char *) &len, sizeof(len));
@ -316,7 +316,7 @@ bool llama_model_load(const std::string & fname, llama_model & model, gpt_vocab
std::vector<uint8_t> tmp; std::vector<uint8_t> tmp;
for (int i = 0; i < n_parts; ++i) { for (register int i = 0; i < n_parts; ++i) {
const int part_id = i; const int part_id = i;
//const int part_id = n_parts - i - 1; //const int part_id = n_parts - i - 1;
@ -755,11 +755,11 @@ static bool is_interacting = false;
#if defined (__unix__) || (defined (__APPLE__) && defined (__MACH__)) #if defined (__unix__) || (defined (__APPLE__) && defined (__MACH__))
void sigint_handler(int signo) { void sigint_handler(int signo) {
if (signo == SIGINT) { if (!(signo == SIGINT)) {
_exit(130);
} else {
if (!is_interacting) { if (!is_interacting) {
is_interacting=true; is_interacting=true;
} else {
_exit(130);
} }
} }
} }
@ -929,7 +929,16 @@ int main(int argc, char ** argv) {
n_past += embd.size(); n_past += embd.size();
embd.clear(); embd.clear();
if (embd_inp.size() <= input_consumed) { if (!(embd_inp.size() <= input_consumed)) {
// some user input remains from prompt or interaction, forward it to processing
while (embd_inp.size() > input_consumed) {
embd.push_back(embd_inp[input_consumed]);
last_n_tokens.erase(last_n_tokens.begin());
last_n_tokens.push_back(embd_inp[input_consumed]);
++input_consumed;
if (embd.size() > params.n_batch) {
break;
} else {
// out of user input, sample next token // out of user input, sample next token
const float top_k = params.top_k; const float top_k = params.top_k;
const float top_p = params.top_p; const float top_p = params.top_p;
@ -959,15 +968,6 @@ int main(int argc, char ** argv) {
// decrement remaining sampling budget // decrement remaining sampling budget
--remaining_tokens; --remaining_tokens;
} else {
// some user input remains from prompt or interaction, forward it to processing
while (embd_inp.size() > input_consumed) {
embd.push_back(embd_inp[input_consumed]);
last_n_tokens.erase(last_n_tokens.begin());
last_n_tokens.push_back(embd_inp[input_consumed]);
++input_consumed;
if (embd.size() > params.n_batch) {
break;
} }
} }