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

keep processed from the beginning

Browse source 
this means no limit to the input prompt,
it will just get reset again as normal
This commit is contained in:
Henri Vasserman 2023-05-28 12:08:37 +03:00
parent df0e0d094c
commit 549291fe61
No known key found for this signature in database
GPG key ID: 2995FC0F58B1A986
1 changed files with 1 additions and 5 deletions
Download patch file Download diff file Expand all files Collapse all files

6
examples/server/server.cpp
View file

@ -100,7 +100,7 @@ struct llama_server_context
// Reset context
const int n_left = n_past - params.n_keep;
n_past = std::max(1, params.n_keep);
processed_tokens.erase(processed_tokens.begin() + n_past, processed_tokens.end());
//processed_tokens.erase(processed_tokens.begin() + n_past, processed_tokens.end());
embd.insert(embd.begin(), last_n_tokens.begin() + params.n_ctx - n_left / 2 - embd.size(), last_n_tokens.end() - embd.size());
}
for (int i = 0; i < (int)embd.size(); i += params.n_batch)
@ -499,10 +499,6 @@ bool parse_options_completion(json body, llama_server_context& llama, Response &
{
llama.params.n_keep = body["n_keep"].get<int>();
}
if (!body["as_loop"].is_null())
{
llama.as_loop = body["as_loop"].get<bool>();
}
if (!body["prompt"].is_null())
{
llama.params.prompt = body["prompt"].get<std::string>();