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

small fixes

Browse source
This commit is contained in:
anon 2023-06-02 05:57:20 -03:00
parent 49dce94885
commit a8a9f19689
1 changed files with 5 additions and 8 deletions
Download patch file Download diff file Expand all files Collapse all files

13
examples/server/server.cpp
View file

@ -137,8 +137,6 @@ struct llama_server_context
void beginCompletion() void beginCompletion()
{ {
// number of tokens to keep when resetting context // number of tokens to keep when resetting context
n_remain = params.n_predict; n_remain = params.n_predict;
llama_set_rng_seed(ctx, params.seed); llama_set_rng_seed(ctx, params.seed);
} }
@ -192,9 +190,8 @@ struct llama_server_context
auto n_vocab = llama_n_vocab(ctx); auto n_vocab = llama_n_vocab(ctx);
// Apply params.logit_bias map // Apply params.logit_bias map
for (auto it = params.logit_bias.begin(); it != params.logit_bias.end(); it++) for (const auto &it : params.logit_bias) {
{ logits[it.first] += it.second;
logits[it->first] += it->second;
} }
std::vector<llama_token_data> candidates; std::vector<llama_token_data> candidates;
@ -271,7 +268,7 @@ struct llama_server_context
return result; return result;
} }
has_next_token = params.n_predict == -1 ? true : n_remain != 0; has_next_token = params.n_predict == -1 || n_remain != 0;
return result; return result;
} }
@ -330,7 +327,7 @@ struct llama_server_context
std::vector<float> embedding(std::string content, int threads) { std::vector<float> embedding(std::string content, int threads) {
content.insert(0, 1, ' '); content.insert(0, 1, ' ');
std::vector<llama_token> tokens = ::llama_tokenize(ctx, content, true); std::vector<llama_token> tokens = ::llama_tokenize(ctx, content, true);
if (tokens.size() > 0) if (!tokens.empty())
{ {
if (llama_eval(ctx, tokens.data(), tokens.size(), 0, threads)) if (llama_eval(ctx, tokens.data(), tokens.size(), 0, threads))
{ {
@ -340,7 +337,7 @@ struct llama_server_context
} }
} }
const int n_embd = llama_n_embd(ctx); const int n_embd = llama_n_embd(ctx);
const auto embeddings = llama_get_embeddings(ctx); auto *const embeddings = llama_get_embeddings(ctx);
std::vector<float> embeddings_(embeddings, embeddings + n_embd); std::vector<float> embeddings_(embeddings, embeddings + n_embd);
return embeddings_; return embeddings_;
} }