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

Update llama.cpp

Browse Source
This commit is contained in:
Andrei Betlen
2023-05-07 00:12:47 -04:00
parent bc853e3742
commit 7c3743fe5f
3 changed files with 9 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
llama_cpp/llama.py
View File

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

8
llama_cpp/llama_cpp.py
View File

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

2
vendor/llama.cpp vendored

Submodule vendor/llama.cpp updated: 2edbdb0f99...1b0fd45465