tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-08-18 14:18:50 -04:00
Code Issues Packages Projects Releases Wiki Activity

apply to the rest

Browse Source
This commit is contained in:
Xuan Son Nguyen
2025-03-13 22:36:27 +01:00
parent 4aabf4e8f4
commit 47086fa82d
18 changed files with 242 additions and 323 deletions
Download Patch File Download Diff File Expand all files Collapse all files

37
common/common.cpp
View File

@@ -582,43 +582,6 @@ std::string string_from(const struct llama_context * ctx, const std::vector<llam
return buf.str();
}
/*
std::string string_from(const struct llama_context * ctx, const struct llama_batch & 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 = common_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();
}
*/
void string_process_escapes(std::string & input) {
std::size_t input_len = input.length();
std::size_t output_idx = 0;

18
common/common.h
View File

@@ -516,7 +516,6 @@ void string_process_escapes(std::string & input);
std::string string_from(bool value);
std::string string_from(const std::vector<int> & values);
std::string string_from(const struct llama_context * ctx, const std::vector<llama_token> & tokens);
std::string string_from(const struct llama_context * ctx, const struct llama_batch & batch);
//
// Filesystem utils
@@ -587,10 +586,10 @@ struct common_batch {
llama_batch_ext_ptr batch;
struct batch_token {
llama_token token;
llama_seq_id seq_id;
bool logits;
};
std::vector<batch_token> tokens;
int n_outputs = 0;
common_batch() = default;
common_batch(int32_t n_tokens, int32_t n_seq_max) {
batch.reset(llama_batch_ext_init(n_tokens, n_seq_max));
@@ -602,7 +601,17 @@ struct common_batch {
}
void add_text(llama_token token, llama_pos pos, llama_seq_id seq_id, bool logits) {
llama_batch_ext_add_text(batch.get(), token, pos, &seq_id, 1, logits);
tokens.push_back({token, seq_id, logits});
tokens.push_back({token, logits});
if (logits) {
n_outputs++;
}
}
void add_text(llama_token token, llama_pos pos, std::vector<llama_seq_id> seq_ids, bool logits) {
llama_batch_ext_add_text(batch.get(), token, pos, seq_ids.data(), seq_ids.size(), logits);
tokens.push_back({token, logits});
if (logits) {
n_outputs++;
}
}
void set_logits_last() {
if (!tokens.empty()) {
@@ -622,6 +631,9 @@ struct common_batch {
view.tokens.reserve(n_tokens);
for (int32_t i = 0; i < n_tokens; i++) {
view.tokens.push_back(tokens[offset + i]);
if (tokens[offset + i].logits) {
view.n_outputs++;
}
}
return view;
}