Merge branch 'abetlen:main' into th-neu-dockerfile-slim

llama_cpp/llama.py

@@ -639,7 +639,7 @@ class Llama:
                 self.detokenize([token]).decode("utf-8", errors="ignore")
                 for token in all_tokens
             ]
-            all_logprobs = [Llama._logits_to_logprobs(row) for row in self.eval_logits]
+            all_logprobs = [Llama.logits_to_logprobs(list(map(float, row))) for row in self.eval_logits]
             for token, token_str, logprobs_token in zip(
                 all_tokens, all_token_strs, all_logprobs
             ):
@@ -985,7 +985,7 @@ class Llama:
         return llama_cpp.llama_token_bos()
 
     @staticmethod
-    def logits_to_logprobs(logits: List[llama_cpp.c_float]) -> List[llama_cpp.c_float]:
+    def logits_to_logprobs(logits: List[float]) -> List[float]:
         exps = [math.exp(float(x)) for x in logits]
         sum_exps = sum(exps)
-        return [llama_cpp.c_float(math.log(x / sum_exps)) for x in exps]
+        return [math.log(x / sum_exps) for x in exps]

pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "llama_cpp_python"
-version = "0.1.41"
+version = "0.1.42"
 description = "Python bindings for the llama.cpp library"
 authors = ["Andrei Betlen <abetlen@gmail.com>"]
 license = "MIT"

setup.py

@@ -10,7 +10,7 @@ setup(
     description="A Python wrapper for llama.cpp",
     long_description=long_description,
     long_description_content_type="text/markdown",
-    version="0.1.41",
+    version="0.1.42",
     author="Andrei Betlen",
     author_email="abetlen@gmail.com",
     license="MIT",

vendor/llama.cpp

@@ -1 +1 @@
-Subproject commit e216aa04633892b972d013719e38b59fd4917341
+Subproject commit 2edbdb0f99336cb41f0995061c7602ed54beb863