alihan/llama-cpp-python
1
0
Fork 0
mirror of https://github.com/abetlen/llama-cpp-python.git synced 2023-09-07 17:34:22 +03:00
Code Issues Packages Projects Releases Wiki Activity

Format

Browse Source
This commit is contained in:
Andrei Betlen
2023-05-17 02:00:39 -04:00
parent d28b753ed2
commit f5c2f998ab
1 changed files with 5 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
llama_cpp/llama.py
View File

@@ -298,7 +298,11 @@ class Llama:
n_vocab = int(llama_cpp.llama_n_vocab(self.ctx)) n_vocab = int(llama_cpp.llama_n_vocab(self.ctx))
n_ctx = int(llama_cpp.llama_n_ctx(self.ctx)) n_ctx = int(llama_cpp.llama_n_ctx(self.ctx))
top_k = llama_cpp.c_int(n_vocab) if top_k.value <= 0 else top_k top_k = llama_cpp.c_int(n_vocab) if top_k.value <= 0 else top_k
last_n_tokens_size = llama_cpp.c_int(n_ctx) if last_n_tokens_size.value < 0 else last_n_tokens_size last_n_tokens_size = (
llama_cpp.c_int(n_ctx)
if last_n_tokens_size.value < 0
else last_n_tokens_size
)
logits = self.eval_logits[-1] logits = self.eval_logits[-1]
nl_logit = logits[llama_cpp.llama_token_nl().value] nl_logit = logits[llama_cpp.llama_token_nl().value]
data = (llama_cpp.llama_token_data * n_vocab)( data = (llama_cpp.llama_token_data * n_vocab)(