CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-10-29 08:41:22 +00:00
Code Issues Packages Projects Releases Wiki Activity

llava-cli : always tokenize special tokens (#5382)

Browse Source 
* llava-cli: tokenize special tokens in prompt

* llava-cli: use the escape CLI argument, remove incomplete separate escaping process
This commit is contained in:
Xiao-Yong Jin
2024-02-07 02:17:25 -06:00
committed by GitHub
parent ee1628bdfe
commit 0ef46da632
1 changed files with 1 additions and 13 deletions
Download Patch File Download Diff File Expand all files Collapse all files

14
examples/llava/llava-cli.cpp
View File

@@ -34,7 +34,7 @@ static bool eval_id(struct llama_context * ctx_llama, int id, int * n_past) {
static bool eval_string(struct llama_context * ctx_llama, const char* str, int n_batch, int * n_past, bool add_bos){ static bool eval_string(struct llama_context * ctx_llama, const char* str, int n_batch, int * n_past, bool add_bos){
std::string str2 = str; std::string str2 = str;
std::vector<llama_token> embd_inp = ::llama_tokenize(ctx_llama, str2, add_bos); std::vector<llama_token> embd_inp = ::llama_tokenize(ctx_llama, str2, add_bos, true);
eval_tokens(ctx_llama, embd_inp, n_batch, n_past); eval_tokens(ctx_llama, embd_inp, n_batch, n_past);
return true; return true;
} }
@@ -152,20 +152,8 @@ static void process_prompt(struct llava_context * ctx_llava, struct llava_image_
size_t image_pos = prompt.find("<image>"); size_t image_pos = prompt.find("<image>");
if (image_pos != std::string::npos) { if (image_pos != std::string::npos) {
// new templating mode: Provide the full prompt including system message and use <image> as a placeholder for the image // new templating mode: Provide the full prompt including system message and use <image> as a placeholder for the image
system_prompt = prompt.substr(0, image_pos); system_prompt = prompt.substr(0, image_pos);
user_prompt = prompt.substr(image_pos + std::string("<image>").length()); user_prompt = prompt.substr(image_pos + std::string("<image>").length());
// We replace \n with actual newlines in user_prompt, just in case -e was not used in templating string
size_t pos = 0;
while ((pos = user_prompt.find("\\n", pos)) != std::string::npos) {
user_prompt.replace(pos, 2, "\n");
pos += 1; // Advance past the replaced newline
}
while ((pos = system_prompt.find("\\n", pos)) != std::string::npos) {
system_prompt.replace(pos, 2, "\n");
pos += 1; // Advance past the replaced newline
}
printf("system_prompt: %s\n", system_prompt.c_str()); printf("system_prompt: %s\n", system_prompt.c_str());
printf("user_prompt: %s\n", user_prompt.c_str()); printf("user_prompt: %s\n", user_prompt.c_str());
} else { } else {