vrr/koboldcpp
2
0
Fork 0
mirror of https://github.com/LostRuins/koboldcpp.git synced 2025-09-10 17:14:36 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

update classpy

Browse source
This commit is contained in:
Concedo 2024-09-15 23:12:49 +08:00
parent de0c96818e
commit 5ab2fd6118
1 changed files with 4 additions and 4 deletions
Download patch file Download diff file Expand all files Collapse all files

8
class.py
View file

@ -57,7 +57,7 @@ class model_backend(InferenceModel):
self.kcpp_ropebase = 10000.0 self.kcpp_ropebase = 10000.0
self.kcpp_useclblast = None self.kcpp_useclblast = None
self.kcpp_usecublas = None self.kcpp_usecublas = None
self.kcpp_noblas = False self.kcpp_usecpu = False
self.kcpp_noavx2 = False self.kcpp_noavx2 = False
self.kcpp_nommap = False self.kcpp_nommap = False
self.kcpp_usevulkan = None self.kcpp_usevulkan = None
@ -236,7 +236,7 @@ class model_backend(InferenceModel):
accel = parameters["kcpp_accelerator"] accel = parameters["kcpp_accelerator"]
if accel==0: if accel==0:
self.kcpp_noblas = True self.kcpp_usecpu = True
elif accel==1: elif accel==1:
self.kcpp_usecublas = ["normal"] self.kcpp_usecublas = ["normal"]
elif accel==2: elif accel==2:
@ -249,7 +249,7 @@ class model_backend(InferenceModel):
self.kcpp_noavx2 = True self.kcpp_noavx2 = True
elif accel==6: elif accel==6:
self.kcpp_noavx2 = True self.kcpp_noavx2 = True
self.kcpp_noblas = True self.kcpp_usecpu = True
self.kcpp_nommap = True self.kcpp_nommap = True
elif accel==7: elif accel==7:
self.kcpp_usevulkan = [0] self.kcpp_usevulkan = [0]
@ -268,7 +268,7 @@ class model_backend(InferenceModel):
port=5001, port_param=5001, host='', launch=False, lora=None, threads=self.kcpp_threads, blasthreads=self.kcpp_threads, port=5001, port_param=5001, host='', launch=False, lora=None, threads=self.kcpp_threads, blasthreads=self.kcpp_threads,
psutil_set_threads=False, highpriority=False, contextsize=self.kcpp_ctxsize, blasbatchsize=self.kcpp_blasbatchsize, psutil_set_threads=False, highpriority=False, contextsize=self.kcpp_ctxsize, blasbatchsize=self.kcpp_blasbatchsize,
ropeconfig=[self.kcpp_ropescale, self.kcpp_ropebase], stream=False, smartcontext=self.kcpp_smartcontext, forceversion=0, ropeconfig=[self.kcpp_ropescale, self.kcpp_ropebase], stream=False, smartcontext=self.kcpp_smartcontext, forceversion=0,
nommap=self.kcpp_nommap, usemlock=False, noavx2=self.kcpp_noavx2, debugmode=self.kcpp_debugmode, skiplauncher=True, noblas=self.kcpp_noblas, nommap=self.kcpp_nommap, usemlock=False, noavx2=self.kcpp_noavx2, debugmode=self.kcpp_debugmode, skiplauncher=True, usecpu=self.kcpp_usecpu,
useclblast=self.kcpp_useclblast, usecublas=self.kcpp_usecublas, usevulkan=self.kcpp_usevulkan, gpulayers=self.kcpp_gpulayers, useclblast=self.kcpp_useclblast, usecublas=self.kcpp_usecublas, usevulkan=self.kcpp_usevulkan, gpulayers=self.kcpp_gpulayers,
tensor_split=self.kcpp_tensor_split, config=None, onready='', multiuser=False, foreground=False, preloadstory=None, noshift=False, tensor_split=self.kcpp_tensor_split, config=None, onready='', multiuser=False, foreground=False, preloadstory=None, noshift=False,
remotetunnel=False, ssl=False, benchmark=None, nocertify=False, mmproj=None, password=None, chatcompletionsadapter=None) remotetunnel=False, ssl=False, benchmark=None, nocertify=False, mmproj=None, password=None, chatcompletionsadapter=None)