Fixed somethings and activated cublas

llama_cpp/llama.py

@@ -390,7 +390,7 @@ class Llama:
             top_k=llama_cpp.c_int(top_k),
             top_p=llama_cpp.c_float(top_p),
             temp=llama_cpp.c_float(temp),
-            mirostat=llama_cpp.c_int(mirostat_mode),
+            mirostat_mode=llama_cpp.c_int(mirostat_mode),
             mirostat_mu=llama_cpp.c_float(mirostat_mu),
             mirostat_tau=llama_cpp.c_float(mirostat_tau),
             mirostat_eta=llama_cpp.c_float(mirostat_eta),
@@ -404,7 +404,7 @@ class Llama:
         top_k: int,
         top_p: float,
         temp: float,
-        mirostat: int,
+        mirostat_mode: int,
         mirostat_tau: float,
         mirostat_eta: float,
         mirostat_mu: float,

setup.py

@@ -22,6 +22,7 @@ setup(
     extras_require={
         "server": ["uvicorn>=0.21.1", "fastapi>=0.95.0", "sse-starlette>=1.3.3"],
     },
+    cmake_args=['-DLLAMA_CUBLAS=ON'],
     python_requires=">=3.7",
     classifiers=[
         "Programming Language :: Python :: 3",

vendor/llama.cpp

@@ -1 +1 @@
-Subproject commit 2edbdb0f99336cb41f0995061c7602ed54beb863
+Subproject commit 173d0e6419e8f8f3c1f4f13201b777f4c60629f3