Update llama.cpp

Author: abetlen

llama_cpp/llama.py

@@ -297,21 +297,25 @@ def _sample_top_p_top_k(
                 ctx=self.ctx,
                 candidates=llama_cpp.ctypes.pointer(candidates),
                 k=top_k,
+                min_keep=llama_cpp.c_size_t(1),
             )
             llama_cpp.llama_sample_tail_free(
                 ctx=self.ctx,
                 candidates=llama_cpp.ctypes.pointer(candidates),
                 z=llama_cpp.c_float(1.0),
+                min_keep=llama_cpp.c_size_t(1),
             )
             llama_cpp.llama_sample_typical(
                 ctx=self.ctx,
                 candidates=llama_cpp.ctypes.pointer(candidates),
                 p=llama_cpp.c_float(1.0),
+                min_keep=llama_cpp.c_size_t(1),
             )
             llama_cpp.llama_sample_top_p(
                 ctx=self.ctx,
                 candidates=llama_cpp.ctypes.pointer(candidates),
                 p=top_p,
+                min_keep=llama_cpp.c_size_t(1),
             )
             llama_cpp.llama_sample_temperature(
                 ctx=self.ctx,

llama_cpp/llama_cpp.py

@@ -515,7 +515,7 @@ def llama_sample_top_k(
     ctx: llama_context_p,
     candidates,  # type: _Pointer[llama_token_data_array]
     k: c_int,
-    min_keep: c_size_t = c_size_t(1),
+    min_keep: c_size_t,
 ):
     return _lib.llama_sample_top_k(ctx, candidates, k, min_keep)
 
@@ -534,7 +534,7 @@ def llama_sample_top_p(
     ctx: llama_context_p,
     candidates,  # type: _Pointer[llama_token_data_array]
     p: c_float,
-    min_keep: c_size_t = c_size_t(1),
+    min_keep: c_size_t,
 ):
     return _lib.llama_sample_top_p(ctx, candidates, p, min_keep)
 
@@ -553,7 +553,7 @@ def llama_sample_tail_free(
     ctx: llama_context_p,
     candidates,  # type: _Pointer[llama_token_data_array]
     z: c_float,
-    min_keep: c_size_t = c_size_t(1),
+    min_keep: c_size_t,
 ):
     return _lib.llama_sample_tail_free(ctx, candidates, z, min_keep)
 
@@ -572,7 +572,7 @@ def llama_sample_typical(
     ctx: llama_context_p,
     candidates,  # type: _Pointer[llama_token_data_array]
     p: c_float,
-    min_keep: c_size_t = c_size_t(1),
+    min_keep: c_size_t,
 ):
     return _lib.llama_sample_typical(ctx, candidates, p, min_keep)