mirror of
https://github.com/mudler/LocalAI.git
synced 2025-05-31 16:05:00 +00:00
feat: add initial AutoGPTQ backend implementation
This commit is contained in:
Ettore Di Giacinto
parent
91d49cfe9f
commit
a843e64fc2
37 changed files with 660 additions and 148 deletions
94
extra/grpc/autogptq/autogptq.py
Executable file
94
extra/grpc/autogptq/autogptq.py
Executable file
|
|
@ -0,0 +1,94 @@
|
|||
#!/usr/bin/env python3
|
||||
import grpc
|
||||
from concurrent import futures
|
||||
import time
|
||||
import backend_pb2
|
||||
import backend_pb2_grpc
|
||||
import argparse
|
||||
import signal
|
||||
import sys
|
||||
import os
|
||||
from auto_gptq import AutoGPTQForCausalLM, BaseQuantizeConfig
|
||||
from pathlib import Path
|
||||
from transformers import AutoTokenizer
|
||||
from transformers import TextGenerationPipeline
|
||||
|
||||
_ONE_DAY_IN_SECONDS = 60 * 60 * 24
|
||||
|
||||
# Implement the BackendServicer class with the service methods
|
||||
class BackendServicer(backend_pb2_grpc.BackendServicer):
|
||||
def Health(self, request, context):
|
||||
return backend_pb2.Reply(message=bytes("OK", 'utf-8'))
|
||||
def LoadModel(self, request, context):
|
||||
try:
|
||||
device = "cuda:0"
|
||||
if request.Device != "":
|
||||
device = request.Device
|
||||
|
||||
tokenizer = AutoTokenizer.from_pretrained(request.Model, use_fast=True)
|
||||
|
||||
model = AutoGPTQForCausalLM.from_quantized(request.Model,
|
||||
model_basename=request.ModelBaseName,
|
||||
use_safetensors=True,
|
||||
trust_remote_code=True,
|
||||
device=device,
|
||||
use_triton=request.UseTriton,
|
||||
quantize_config=None)
|
||||
|
||||
self.model = model
|
||||
self.tokenizer = tokenizer
|
||||
except Exception as err:
|
||||
return backend_pb2.Result(success=False, message=f"Unexpected {err=}, {type(err)=}")
|
||||
return backend_pb2.Result(message="Model loaded successfully", success=True)
|
||||
|
||||
def Predict(self, request, context):
|
||||
# Implement Predict RPC
|
||||
pipeline = TextGenerationPipeline(
|
||||
model=self.model,
|
||||
tokenizer=self.tokenizer,
|
||||
max_new_tokens=request.Tokens,
|
||||
temperature=request.Temperature,
|
||||
top_p=request.TopP,
|
||||
repetition_penalty=request.Penalty,
|
||||
)
|
||||
return backend_pb2.Result(message=bytes(pipeline(request.Prompt)[0]["generated_text"]))
|
||||
|
||||
def PredictStream(self, request, context):
|
||||
# Implement PredictStream RPC
|
||||
#for reply in some_data_generator():
|
||||
# yield reply
|
||||
# Not implemented yet
|
||||
return self.Predict(request, context)
|
||||
|
||||
|
||||
def serve(address):
|
||||
server = grpc.server(futures.ThreadPoolExecutor(max_workers=10))
|
||||
backend_pb2_grpc.add_BackendServicer_to_server(BackendServicer(), server)
|
||||
server.add_insecure_port(address)
|
||||
server.start()
|
||||
print("Server started. Listening on: " + address, file=sys.stderr)
|
||||
|
||||
# Define the signal handler function
|
||||
def signal_handler(sig, frame):
|
||||
print("Received termination signal. Shutting down...")
|
||||
server.stop(0)
|
||||
sys.exit(0)
|
||||
|
||||
# Set the signal handlers for SIGINT and SIGTERM
|
||||
signal.signal(signal.SIGINT, signal_handler)
|
||||
signal.signal(signal.SIGTERM, signal_handler)
|
||||
|
||||
try:
|
||||
while True:
|
||||
time.sleep(_ONE_DAY_IN_SECONDS)
|
||||
except KeyboardInterrupt:
|
||||
server.stop(0)
|
||||
|
||||
if __name__ == "__main__":
|
||||
parser = argparse.ArgumentParser(description="Run the gRPC server.")
|
||||
parser.add_argument(
|
||||
"--addr", default="localhost:50051", help="The address to bind the server to."
|
||||
)
|
||||
args = parser.parse_args()
|
||||
|
||||
serve(args.addr)
|
||||
49
extra/grpc/autogptq/backend_pb2.py
Normal file
49
extra/grpc/autogptq/backend_pb2.py
Normal file
|
|
@ -0,0 +1,49 @@
|
|||
# -*- coding: utf-8 -*-
|
||||
# Generated by the protocol buffer compiler. DO NOT EDIT!
|
||||
# source: backend.proto
|
||||
"""Generated protocol buffer code."""
|
||||
from google.protobuf import descriptor as _descriptor
|
||||
from google.protobuf import descriptor_pool as _descriptor_pool
|
||||
from google.protobuf import symbol_database as _symbol_database
|
||||
from google.protobuf.internal import builder as _builder
|
||||
# @@protoc_insertion_point(imports)
|
||||
|
||||
_sym_db = _symbol_database.Default()
|
||||
|
||||
|
||||
|
||||
|
||||
DESCRIPTOR = _descriptor_pool.Default().AddSerializedFile(b'\n\rbackend.proto\x12\x07\x62\x61\x63kend\"\x0f\n\rHealthMessage\"\x86\x06\n\x0ePredictOptions\x12\x0e\n\x06Prompt\x18\x01 \x01(\t\x12\x0c\n\x04Seed\x18\x02 \x01(\x05\x12\x0f\n\x07Threads\x18\x03 \x01(\x05\x12\x0e\n\x06Tokens\x18\x04 \x01(\x05\x12\x0c\n\x04TopK\x18\x05 \x01(\x05\x12\x0e\n\x06Repeat\x18\x06 \x01(\x05\x12\r\n\x05\x42\x61tch\x18\x07 \x01(\x05\x12\r\n\x05NKeep\x18\x08 \x01(\x05\x12\x13\n\x0bTemperature\x18\t \x01(\x02\x12\x0f\n\x07Penalty\x18\n \x01(\x02\x12\r\n\x05\x46\x31\x36KV\x18\x0b \x01(\x08\x12\x11\n\tDebugMode\x18\x0c \x01(\x08\x12\x13\n\x0bStopPrompts\x18\r \x03(\t\x12\x11\n\tIgnoreEOS\x18\x0e \x01(\x08\x12\x19\n\x11TailFreeSamplingZ\x18\x0f \x01(\x02\x12\x10\n\x08TypicalP\x18\x10 \x01(\x02\x12\x18\n\x10\x46requencyPenalty\x18\x11 \x01(\x02\x12\x17\n\x0fPresencePenalty\x18\x12 \x01(\x02\x12\x10\n\x08Mirostat\x18\x13 \x01(\x05\x12\x13\n\x0bMirostatETA\x18\x14 \x01(\x02\x12\x13\n\x0bMirostatTAU\x18\x15 \x01(\x02\x12\x12\n\nPenalizeNL\x18\x16 \x01(\x08\x12\x11\n\tLogitBias\x18\x17 \x01(\t\x12\r\n\x05MLock\x18\x19 \x01(\x08\x12\x0c\n\x04MMap\x18\x1a \x01(\x08\x12\x16\n\x0ePromptCacheAll\x18\x1b \x01(\x08\x12\x15\n\rPromptCacheRO\x18\x1c \x01(\x08\x12\x0f\n\x07Grammar\x18\x1d \x01(\t\x12\x0f\n\x07MainGPU\x18\x1e \x01(\t\x12\x13\n\x0bTensorSplit\x18\x1f \x01(\t\x12\x0c\n\x04TopP\x18 \x01(\x02\x12\x17\n\x0fPromptCachePath\x18! \x01(\t\x12\r\n\x05\x44\x65\x62ug\x18\" \x01(\x08\x12\x17\n\x0f\x45mbeddingTokens\x18# \x03(\x05\x12\x12\n\nEmbeddings\x18$ \x01(\t\x12\x14\n\x0cRopeFreqBase\x18% \x01(\x02\x12\x15\n\rRopeFreqScale\x18& \x01(\x02\x12\x1b\n\x13NegativePromptScale\x18\' \x01(\x02\x12\x16\n\x0eNegativePrompt\x18( \x01(\t\"\x18\n\x05Reply\x12\x0f\n\x07message\x18\x01 \x01(\x0c\"\xc8\x03\n\x0cModelOptions\x12\r\n\x05Model\x18\x01 \x01(\t\x12\x13\n\x0b\x43ontextSize\x18\x02 \x01(\x05\x12\x0c\n\x04Seed\x18\x03 \x01(\x05\x12\x0e\n\x06NBatch\x18\x04 \x01(\x05\x12\x11\n\tF16Memory\x18\x05 \x01(\x08\x12\r\n\x05MLock\x18\x06 \x01(\x08\x12\x0c\n\x04MMap\x18\x07 \x01(\x08\x12\x11\n\tVocabOnly\x18\x08 \x01(\x08\x12\x0f\n\x07LowVRAM\x18\t \x01(\x08\x12\x12\n\nEmbeddings\x18\n \x01(\x08\x12\x0c\n\x04NUMA\x18\x0b \x01(\x08\x12\x12\n\nNGPULayers\x18\x0c \x01(\x05\x12\x0f\n\x07MainGPU\x18\r \x01(\t\x12\x13\n\x0bTensorSplit\x18\x0e \x01(\t\x12\x0f\n\x07Threads\x18\x0f \x01(\x05\x12\x19\n\x11LibrarySearchPath\x18\x10 \x01(\t\x12\x14\n\x0cRopeFreqBase\x18\x11 \x01(\x02\x12\x15\n\rRopeFreqScale\x18\x12 \x01(\x02\x12\x12\n\nRMSNormEps\x18\x13 \x01(\x02\x12\x0c\n\x04NGQA\x18\x14 \x01(\x05\x12\x11\n\tModelFile\x18\x15 \x01(\t\x12\x0e\n\x06\x44\x65vice\x18\x16 \x01(\t\x12\x11\n\tUseTriton\x18\x17 \x01(\x08\x12\x15\n\rModelBaseName\x18\x18 \x01(\t\"*\n\x06Result\x12\x0f\n\x07message\x18\x01 \x01(\t\x12\x0f\n\x07success\x18\x02 \x01(\x08\"%\n\x0f\x45mbeddingResult\x12\x12\n\nembeddings\x18\x01 \x03(\x02\"C\n\x11TranscriptRequest\x12\x0b\n\x03\x64st\x18\x02 \x01(\t\x12\x10\n\x08language\x18\x03 \x01(\t\x12\x0f\n\x07threads\x18\x04 \x01(\r\"N\n\x10TranscriptResult\x12,\n\x08segments\x18\x01 \x03(\x0b\x32\x1a.backend.TranscriptSegment\x12\x0c\n\x04text\x18\x02 \x01(\t\"Y\n\x11TranscriptSegment\x12\n\n\x02id\x18\x01 \x01(\x05\x12\r\n\x05start\x18\x02 \x01(\x03\x12\x0b\n\x03\x65nd\x18\x03 \x01(\x03\x12\x0c\n\x04text\x18\x04 \x01(\t\x12\x0e\n\x06tokens\x18\x05 \x03(\x05\"\x9e\x01\n\x14GenerateImageRequest\x12\x0e\n\x06height\x18\x01 \x01(\x05\x12\r\n\x05width\x18\x02 \x01(\x05\x12\x0c\n\x04mode\x18\x03 \x01(\x05\x12\x0c\n\x04step\x18\x04 \x01(\x05\x12\x0c\n\x04seed\x18\x05 \x01(\x05\x12\x17\n\x0fpositive_prompt\x18\x06 \x01(\t\x12\x17\n\x0fnegative_prompt\x18\x07 \x01(\t\x12\x0b\n\x03\x64st\x18\x08 \x01(\t\"6\n\nTTSRequest\x12\x0c\n\x04text\x18\x01 \x01(\t\x12\r\n\x05model\x18\x02 \x01(\t\x12\x0b\n\x03\x64st\x18\x03 \x01(\t2\xeb\x03\n\x07\x42\x61\x63kend\x12\x32\n\x06Health\x12\x16.backend.HealthMessage\x1a\x0e.backend.Reply\"\x00\x12\x34\n\x07Predict\x12\x17.backend.PredictOptions\x1a\x0e.backend.Reply\"\x00\x12\x35\n\tLoadModel\x12\x15.backend.ModelOptions\x1a\x0f.backend.Result\"\x00\x12<\n\rPredictStream\x12\x17.backend.PredictOptions\x1a\x0e.backend.Reply\"\x00\x30\x01\x12@\n\tEmbedding\x12\x17.backend.PredictOptions\x1a\x18.backend.EmbeddingResult\"\x00\x12\x41\n\rGenerateImage\x12\x1d.backend.GenerateImageRequest\x1a\x0f.backend.Result\"\x00\x12M\n\x12\x41udioTranscription\x12\x1a.backend.TranscriptRequest\x1a\x19.backend.TranscriptResult\"\x00\x12-\n\x03TTS\x12\x13.backend.TTSRequest\x1a\x0f.backend.Result\"\x00\x42Z\n\x19io.skynet.localai.backendB\x0eLocalAIBackendP\x01Z+github.com/go-skynet/LocalAI/pkg/grpc/protob\x06proto3')
|
||||
|
||||
_globals = globals()
|
||||
_builder.BuildMessageAndEnumDescriptors(DESCRIPTOR, _globals)
|
||||
_builder.BuildTopDescriptorsAndMessages(DESCRIPTOR, 'backend_pb2', _globals)
|
||||
if _descriptor._USE_C_DESCRIPTORS == False:
|
||||
|
||||
DESCRIPTOR._options = None
|
||||
DESCRIPTOR._serialized_options = b'\n\031io.skynet.localai.backendB\016LocalAIBackendP\001Z+github.com/go-skynet/LocalAI/pkg/grpc/proto'
|
||||
_globals['_HEALTHMESSAGE']._serialized_start=26
|
||||
_globals['_HEALTHMESSAGE']._serialized_end=41
|
||||
_globals['_PREDICTOPTIONS']._serialized_start=44
|
||||
_globals['_PREDICTOPTIONS']._serialized_end=818
|
||||
_globals['_REPLY']._serialized_start=820
|
||||
_globals['_REPLY']._serialized_end=844
|
||||
_globals['_MODELOPTIONS']._serialized_start=847
|
||||
_globals['_MODELOPTIONS']._serialized_end=1303
|
||||
_globals['_RESULT']._serialized_start=1305
|
||||
_globals['_RESULT']._serialized_end=1347
|
||||
_globals['_EMBEDDINGRESULT']._serialized_start=1349
|
||||
_globals['_EMBEDDINGRESULT']._serialized_end=1386
|
||||
_globals['_TRANSCRIPTREQUEST']._serialized_start=1388
|
||||
_globals['_TRANSCRIPTREQUEST']._serialized_end=1455
|
||||
_globals['_TRANSCRIPTRESULT']._serialized_start=1457
|
||||
_globals['_TRANSCRIPTRESULT']._serialized_end=1535
|
||||
_globals['_TRANSCRIPTSEGMENT']._serialized_start=1537
|
||||
_globals['_TRANSCRIPTSEGMENT']._serialized_end=1626
|
||||
_globals['_GENERATEIMAGEREQUEST']._serialized_start=1629
|
||||
_globals['_GENERATEIMAGEREQUEST']._serialized_end=1787
|
||||
_globals['_TTSREQUEST']._serialized_start=1789
|
||||
_globals['_TTSREQUEST']._serialized_end=1843
|
||||
_globals['_BACKEND']._serialized_start=1846
|
||||
_globals['_BACKEND']._serialized_end=2337
|
||||
# @@protoc_insertion_point(module_scope)
|
||||
297
extra/grpc/autogptq/backend_pb2_grpc.py
Normal file
297
extra/grpc/autogptq/backend_pb2_grpc.py
Normal file
|
|
@ -0,0 +1,297 @@
|
|||
# Generated by the gRPC Python protocol compiler plugin. DO NOT EDIT!
|
||||
"""Client and server classes corresponding to protobuf-defined services."""
|
||||
import grpc
|
||||
|
||||
import backend_pb2 as backend__pb2
|
||||
|
||||
|
||||
class BackendStub(object):
|
||||
"""Missing associated documentation comment in .proto file."""
|
||||
|
||||
def __init__(self, channel):
|
||||
"""Constructor.
|
||||
|
||||
Args:
|
||||
channel: A grpc.Channel.
|
||||
"""
|
||||
self.Health = channel.unary_unary(
|
||||
'/backend.Backend/Health',
|
||||
request_serializer=backend__pb2.HealthMessage.SerializeToString,
|
||||
response_deserializer=backend__pb2.Reply.FromString,
|
||||
)
|
||||
self.Predict = channel.unary_unary(
|
||||
'/backend.Backend/Predict',
|
||||
request_serializer=backend__pb2.PredictOptions.SerializeToString,
|
||||
response_deserializer=backend__pb2.Reply.FromString,
|
||||
)
|
||||
self.LoadModel = channel.unary_unary(
|
||||
'/backend.Backend/LoadModel',
|
||||
request_serializer=backend__pb2.ModelOptions.SerializeToString,
|
||||
response_deserializer=backend__pb2.Result.FromString,
|
||||
)
|
||||
self.PredictStream = channel.unary_stream(
|
||||
'/backend.Backend/PredictStream',
|
||||
request_serializer=backend__pb2.PredictOptions.SerializeToString,
|
||||
response_deserializer=backend__pb2.Reply.FromString,
|
||||
)
|
||||
self.Embedding = channel.unary_unary(
|
||||
'/backend.Backend/Embedding',
|
||||
request_serializer=backend__pb2.PredictOptions.SerializeToString,
|
||||
response_deserializer=backend__pb2.EmbeddingResult.FromString,
|
||||
)
|
||||
self.GenerateImage = channel.unary_unary(
|
||||
'/backend.Backend/GenerateImage',
|
||||
request_serializer=backend__pb2.GenerateImageRequest.SerializeToString,
|
||||
response_deserializer=backend__pb2.Result.FromString,
|
||||
)
|
||||
self.AudioTranscription = channel.unary_unary(
|
||||
'/backend.Backend/AudioTranscription',
|
||||
request_serializer=backend__pb2.TranscriptRequest.SerializeToString,
|
||||
response_deserializer=backend__pb2.TranscriptResult.FromString,
|
||||
)
|
||||
self.TTS = channel.unary_unary(
|
||||
'/backend.Backend/TTS',
|
||||
request_serializer=backend__pb2.TTSRequest.SerializeToString,
|
||||
response_deserializer=backend__pb2.Result.FromString,
|
||||
)
|
||||
|
||||
|
||||
class BackendServicer(object):
|
||||
"""Missing associated documentation comment in .proto file."""
|
||||
|
||||
def Health(self, request, context):
|
||||
"""Missing associated documentation comment in .proto file."""
|
||||
context.set_code(grpc.StatusCode.UNIMPLEMENTED)
|
||||
context.set_details('Method not implemented!')
|
||||
raise NotImplementedError('Method not implemented!')
|
||||
|
||||
def Predict(self, request, context):
|
||||
"""Missing associated documentation comment in .proto file."""
|
||||
context.set_code(grpc.StatusCode.UNIMPLEMENTED)
|
||||
context.set_details('Method not implemented!')
|
||||
raise NotImplementedError('Method not implemented!')
|
||||
|
||||
def LoadModel(self, request, context):
|
||||
"""Missing associated documentation comment in .proto file."""
|
||||
context.set_code(grpc.StatusCode.UNIMPLEMENTED)
|
||||
context.set_details('Method not implemented!')
|
||||
raise NotImplementedError('Method not implemented!')
|
||||
|
||||
def PredictStream(self, request, context):
|
||||
"""Missing associated documentation comment in .proto file."""
|
||||
context.set_code(grpc.StatusCode.UNIMPLEMENTED)
|
||||
context.set_details('Method not implemented!')
|
||||
raise NotImplementedError('Method not implemented!')
|
||||
|
||||
def Embedding(self, request, context):
|
||||
"""Missing associated documentation comment in .proto file."""
|
||||
context.set_code(grpc.StatusCode.UNIMPLEMENTED)
|
||||
context.set_details('Method not implemented!')
|
||||
raise NotImplementedError('Method not implemented!')
|
||||
|
||||
def GenerateImage(self, request, context):
|
||||
"""Missing associated documentation comment in .proto file."""
|
||||
context.set_code(grpc.StatusCode.UNIMPLEMENTED)
|
||||
context.set_details('Method not implemented!')
|
||||
raise NotImplementedError('Method not implemented!')
|
||||
|
||||
def AudioTranscription(self, request, context):
|
||||
"""Missing associated documentation comment in .proto file."""
|
||||
context.set_code(grpc.StatusCode.UNIMPLEMENTED)
|
||||
context.set_details('Method not implemented!')
|
||||
raise NotImplementedError('Method not implemented!')
|
||||
|
||||
def TTS(self, request, context):
|
||||
"""Missing associated documentation comment in .proto file."""
|
||||
context.set_code(grpc.StatusCode.UNIMPLEMENTED)
|
||||
context.set_details('Method not implemented!')
|
||||
raise NotImplementedError('Method not implemented!')
|
||||
|
||||
|
||||
def add_BackendServicer_to_server(servicer, server):
|
||||
rpc_method_handlers = {
|
||||
'Health': grpc.unary_unary_rpc_method_handler(
|
||||
servicer.Health,
|
||||
request_deserializer=backend__pb2.HealthMessage.FromString,
|
||||
response_serializer=backend__pb2.Reply.SerializeToString,
|
||||
),
|
||||
'Predict': grpc.unary_unary_rpc_method_handler(
|
||||
servicer.Predict,
|
||||
request_deserializer=backend__pb2.PredictOptions.FromString,
|
||||
response_serializer=backend__pb2.Reply.SerializeToString,
|
||||
),
|
||||
'LoadModel': grpc.unary_unary_rpc_method_handler(
|
||||
servicer.LoadModel,
|
||||
request_deserializer=backend__pb2.ModelOptions.FromString,
|
||||
response_serializer=backend__pb2.Result.SerializeToString,
|
||||
),
|
||||
'PredictStream': grpc.unary_stream_rpc_method_handler(
|
||||
servicer.PredictStream,
|
||||
request_deserializer=backend__pb2.PredictOptions.FromString,
|
||||
response_serializer=backend__pb2.Reply.SerializeToString,
|
||||
),
|
||||
'Embedding': grpc.unary_unary_rpc_method_handler(
|
||||
servicer.Embedding,
|
||||
request_deserializer=backend__pb2.PredictOptions.FromString,
|
||||
response_serializer=backend__pb2.EmbeddingResult.SerializeToString,
|
||||
),
|
||||
'GenerateImage': grpc.unary_unary_rpc_method_handler(
|
||||
servicer.GenerateImage,
|
||||
request_deserializer=backend__pb2.GenerateImageRequest.FromString,
|
||||
response_serializer=backend__pb2.Result.SerializeToString,
|
||||
),
|
||||
'AudioTranscription': grpc.unary_unary_rpc_method_handler(
|
||||
servicer.AudioTranscription,
|
||||
request_deserializer=backend__pb2.TranscriptRequest.FromString,
|
||||
response_serializer=backend__pb2.TranscriptResult.SerializeToString,
|
||||
),
|
||||
'TTS': grpc.unary_unary_rpc_method_handler(
|
||||
servicer.TTS,
|
||||
request_deserializer=backend__pb2.TTSRequest.FromString,
|
||||
response_serializer=backend__pb2.Result.SerializeToString,
|
||||
),
|
||||
}
|
||||
generic_handler = grpc.method_handlers_generic_handler(
|
||||
'backend.Backend', rpc_method_handlers)
|
||||
server.add_generic_rpc_handlers((generic_handler,))
|
||||
|
||||
|
||||
# This class is part of an EXPERIMENTAL API.
|
||||
class Backend(object):
|
||||
"""Missing associated documentation comment in .proto file."""
|
||||
|
||||
@staticmethod
|
||||
def Health(request,
|
||||
target,
|
||||
options=(),
|
||||
channel_credentials=None,
|
||||
call_credentials=None,
|
||||
insecure=False,
|
||||
compression=None,
|
||||
wait_for_ready=None,
|
||||
timeout=None,
|
||||
metadata=None):
|
||||
return grpc.experimental.unary_unary(request, target, '/backend.Backend/Health',
|
||||
backend__pb2.HealthMessage.SerializeToString,
|
||||
backend__pb2.Reply.FromString,
|
||||
options, channel_credentials,
|
||||
insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
|
||||
|
||||
@staticmethod
|
||||
def Predict(request,
|
||||
target,
|
||||
options=(),
|
||||
channel_credentials=None,
|
||||
call_credentials=None,
|
||||
insecure=False,
|
||||
compression=None,
|
||||
wait_for_ready=None,
|
||||
timeout=None,
|
||||
metadata=None):
|
||||
return grpc.experimental.unary_unary(request, target, '/backend.Backend/Predict',
|
||||
backend__pb2.PredictOptions.SerializeToString,
|
||||
backend__pb2.Reply.FromString,
|
||||
options, channel_credentials,
|
||||
insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
|
||||
|
||||
@staticmethod
|
||||
def LoadModel(request,
|
||||
target,
|
||||
options=(),
|
||||
channel_credentials=None,
|
||||
call_credentials=None,
|
||||
insecure=False,
|
||||
compression=None,
|
||||
wait_for_ready=None,
|
||||
timeout=None,
|
||||
metadata=None):
|
||||
return grpc.experimental.unary_unary(request, target, '/backend.Backend/LoadModel',
|
||||
backend__pb2.ModelOptions.SerializeToString,
|
||||
backend__pb2.Result.FromString,
|
||||
options, channel_credentials,
|
||||
insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
|
||||
|
||||
@staticmethod
|
||||
def PredictStream(request,
|
||||
target,
|
||||
options=(),
|
||||
channel_credentials=None,
|
||||
call_credentials=None,
|
||||
insecure=False,
|
||||
compression=None,
|
||||
wait_for_ready=None,
|
||||
timeout=None,
|
||||
metadata=None):
|
||||
return grpc.experimental.unary_stream(request, target, '/backend.Backend/PredictStream',
|
||||
backend__pb2.PredictOptions.SerializeToString,
|
||||
backend__pb2.Reply.FromString,
|
||||
options, channel_credentials,
|
||||
insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
|
||||
|
||||
@staticmethod
|
||||
def Embedding(request,
|
||||
target,
|
||||
options=(),
|
||||
channel_credentials=None,
|
||||
call_credentials=None,
|
||||
insecure=False,
|
||||
compression=None,
|
||||
wait_for_ready=None,
|
||||
timeout=None,
|
||||
metadata=None):
|
||||
return grpc.experimental.unary_unary(request, target, '/backend.Backend/Embedding',
|
||||
backend__pb2.PredictOptions.SerializeToString,
|
||||
backend__pb2.EmbeddingResult.FromString,
|
||||
options, channel_credentials,
|
||||
insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
|
||||
|
||||
@staticmethod
|
||||
def GenerateImage(request,
|
||||
target,
|
||||
options=(),
|
||||
channel_credentials=None,
|
||||
call_credentials=None,
|
||||
insecure=False,
|
||||
compression=None,
|
||||
wait_for_ready=None,
|
||||
timeout=None,
|
||||
metadata=None):
|
||||
return grpc.experimental.unary_unary(request, target, '/backend.Backend/GenerateImage',
|
||||
backend__pb2.GenerateImageRequest.SerializeToString,
|
||||
backend__pb2.Result.FromString,
|
||||
options, channel_credentials,
|
||||
insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
|
||||
|
||||
@staticmethod
|
||||
def AudioTranscription(request,
|
||||
target,
|
||||
options=(),
|
||||
channel_credentials=None,
|
||||
call_credentials=None,
|
||||
insecure=False,
|
||||
compression=None,
|
||||
wait_for_ready=None,
|
||||
timeout=None,
|
||||
metadata=None):
|
||||
return grpc.experimental.unary_unary(request, target, '/backend.Backend/AudioTranscription',
|
||||
backend__pb2.TranscriptRequest.SerializeToString,
|
||||
backend__pb2.TranscriptResult.FromString,
|
||||
options, channel_credentials,
|
||||
insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
|
||||
|
||||
@staticmethod
|
||||
def TTS(request,
|
||||
target,
|
||||
options=(),
|
||||
channel_credentials=None,
|
||||
call_credentials=None,
|
||||
insecure=False,
|
||||
compression=None,
|
||||
wait_for_ready=None,
|
||||
timeout=None,
|
||||
metadata=None):
|
||||
return grpc.experimental.unary_unary(request, target, '/backend.Backend/TTS',
|
||||
backend__pb2.TTSRequest.SerializeToString,
|
||||
backend__pb2.Result.FromString,
|
||||
options, channel_credentials,
|
||||
insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
|
||||
|
|
@ -13,7 +13,7 @@ _sym_db = _symbol_database.Default()
|
|||
|
||||
|
||||
|
||||
DESCRIPTOR = _descriptor_pool.Default().AddSerializedFile(b'\n\rbackend.proto\x12\x07\x62\x61\x63kend\"\x0f\n\rHealthMessage\"\x86\x06\n\x0ePredictOptions\x12\x0e\n\x06Prompt\x18\x01 \x01(\t\x12\x0c\n\x04Seed\x18\x02 \x01(\x05\x12\x0f\n\x07Threads\x18\x03 \x01(\x05\x12\x0e\n\x06Tokens\x18\x04 \x01(\x05\x12\x0c\n\x04TopK\x18\x05 \x01(\x05\x12\x0e\n\x06Repeat\x18\x06 \x01(\x05\x12\r\n\x05\x42\x61tch\x18\x07 \x01(\x05\x12\r\n\x05NKeep\x18\x08 \x01(\x05\x12\x13\n\x0bTemperature\x18\t \x01(\x02\x12\x0f\n\x07Penalty\x18\n \x01(\x02\x12\r\n\x05\x46\x31\x36KV\x18\x0b \x01(\x08\x12\x11\n\tDebugMode\x18\x0c \x01(\x08\x12\x13\n\x0bStopPrompts\x18\r \x03(\t\x12\x11\n\tIgnoreEOS\x18\x0e \x01(\x08\x12\x19\n\x11TailFreeSamplingZ\x18\x0f \x01(\x02\x12\x10\n\x08TypicalP\x18\x10 \x01(\x02\x12\x18\n\x10\x46requencyPenalty\x18\x11 \x01(\x02\x12\x17\n\x0fPresencePenalty\x18\x12 \x01(\x02\x12\x10\n\x08Mirostat\x18\x13 \x01(\x05\x12\x13\n\x0bMirostatETA\x18\x14 \x01(\x02\x12\x13\n\x0bMirostatTAU\x18\x15 \x01(\x02\x12\x12\n\nPenalizeNL\x18\x16 \x01(\x08\x12\x11\n\tLogitBias\x18\x17 \x01(\t\x12\r\n\x05MLock\x18\x19 \x01(\x08\x12\x0c\n\x04MMap\x18\x1a \x01(\x08\x12\x16\n\x0ePromptCacheAll\x18\x1b \x01(\x08\x12\x15\n\rPromptCacheRO\x18\x1c \x01(\x08\x12\x0f\n\x07Grammar\x18\x1d \x01(\t\x12\x0f\n\x07MainGPU\x18\x1e \x01(\t\x12\x13\n\x0bTensorSplit\x18\x1f \x01(\t\x12\x0c\n\x04TopP\x18 \x01(\x02\x12\x17\n\x0fPromptCachePath\x18! \x01(\t\x12\r\n\x05\x44\x65\x62ug\x18\" \x01(\x08\x12\x17\n\x0f\x45mbeddingTokens\x18# \x03(\x05\x12\x12\n\nEmbeddings\x18$ \x01(\t\x12\x14\n\x0cRopeFreqBase\x18% \x01(\x02\x12\x15\n\rRopeFreqScale\x18& \x01(\x02\x12\x1b\n\x13NegativePromptScale\x18\' \x01(\x02\x12\x16\n\x0eNegativePrompt\x18( \x01(\t\"\x18\n\x05Reply\x12\x0f\n\x07message\x18\x01 \x01(\x0c\"\xfb\x02\n\x0cModelOptions\x12\r\n\x05Model\x18\x01 \x01(\t\x12\x13\n\x0b\x43ontextSize\x18\x02 \x01(\x05\x12\x0c\n\x04Seed\x18\x03 \x01(\x05\x12\x0e\n\x06NBatch\x18\x04 \x01(\x05\x12\x11\n\tF16Memory\x18\x05 \x01(\x08\x12\r\n\x05MLock\x18\x06 \x01(\x08\x12\x0c\n\x04MMap\x18\x07 \x01(\x08\x12\x11\n\tVocabOnly\x18\x08 \x01(\x08\x12\x0f\n\x07LowVRAM\x18\t \x01(\x08\x12\x12\n\nEmbeddings\x18\n \x01(\x08\x12\x0c\n\x04NUMA\x18\x0b \x01(\x08\x12\x12\n\nNGPULayers\x18\x0c \x01(\x05\x12\x0f\n\x07MainGPU\x18\r \x01(\t\x12\x13\n\x0bTensorSplit\x18\x0e \x01(\t\x12\x0f\n\x07Threads\x18\x0f \x01(\x05\x12\x19\n\x11LibrarySearchPath\x18\x10 \x01(\t\x12\x14\n\x0cRopeFreqBase\x18\x11 \x01(\x02\x12\x15\n\rRopeFreqScale\x18\x12 \x01(\x02\x12\x12\n\nRMSNormEps\x18\x13 \x01(\x02\x12\x0c\n\x04NGQA\x18\x14 \x01(\x05\"*\n\x06Result\x12\x0f\n\x07message\x18\x01 \x01(\t\x12\x0f\n\x07success\x18\x02 \x01(\x08\"%\n\x0f\x45mbeddingResult\x12\x12\n\nembeddings\x18\x01 \x03(\x02\"C\n\x11TranscriptRequest\x12\x0b\n\x03\x64st\x18\x02 \x01(\t\x12\x10\n\x08language\x18\x03 \x01(\t\x12\x0f\n\x07threads\x18\x04 \x01(\r\"N\n\x10TranscriptResult\x12,\n\x08segments\x18\x01 \x03(\x0b\x32\x1a.backend.TranscriptSegment\x12\x0c\n\x04text\x18\x02 \x01(\t\"Y\n\x11TranscriptSegment\x12\n\n\x02id\x18\x01 \x01(\x05\x12\r\n\x05start\x18\x02 \x01(\x03\x12\x0b\n\x03\x65nd\x18\x03 \x01(\x03\x12\x0c\n\x04text\x18\x04 \x01(\t\x12\x0e\n\x06tokens\x18\x05 \x03(\x05\"\x9e\x01\n\x14GenerateImageRequest\x12\x0e\n\x06height\x18\x01 \x01(\x05\x12\r\n\x05width\x18\x02 \x01(\x05\x12\x0c\n\x04mode\x18\x03 \x01(\x05\x12\x0c\n\x04step\x18\x04 \x01(\x05\x12\x0c\n\x04seed\x18\x05 \x01(\x05\x12\x17\n\x0fpositive_prompt\x18\x06 \x01(\t\x12\x17\n\x0fnegative_prompt\x18\x07 \x01(\t\x12\x0b\n\x03\x64st\x18\x08 \x01(\t\"6\n\nTTSRequest\x12\x0c\n\x04text\x18\x01 \x01(\t\x12\r\n\x05model\x18\x02 \x01(\t\x12\x0b\n\x03\x64st\x18\x03 \x01(\t2\xeb\x03\n\x07\x42\x61\x63kend\x12\x32\n\x06Health\x12\x16.backend.HealthMessage\x1a\x0e.backend.Reply\"\x00\x12\x34\n\x07Predict\x12\x17.backend.PredictOptions\x1a\x0e.backend.Reply\"\x00\x12\x35\n\tLoadModel\x12\x15.backend.ModelOptions\x1a\x0f.backend.Result\"\x00\x12<\n\rPredictStream\x12\x17.backend.PredictOptions\x1a\x0e.backend.Reply\"\x00\x30\x01\x12@\n\tEmbedding\x12\x17.backend.PredictOptions\x1a\x18.backend.EmbeddingResult\"\x00\x12\x41\n\rGenerateImage\x12\x1d.backend.GenerateImageRequest\x1a\x0f.backend.Result\"\x00\x12M\n\x12\x41udioTranscription\x12\x1a.backend.TranscriptRequest\x1a\x19.backend.TranscriptResult\"\x00\x12-\n\x03TTS\x12\x13.backend.TTSRequest\x1a\x0f.backend.Result\"\x00\x42Z\n\x19io.skynet.localai.backendB\x0eLocalAIBackendP\x01Z+github.com/go-skynet/LocalAI/pkg/grpc/protob\x06proto3')
|
||||
DESCRIPTOR = _descriptor_pool.Default().AddSerializedFile(b'\n\rbackend.proto\x12\x07\x62\x61\x63kend\"\x0f\n\rHealthMessage\"\x86\x06\n\x0ePredictOptions\x12\x0e\n\x06Prompt\x18\x01 \x01(\t\x12\x0c\n\x04Seed\x18\x02 \x01(\x05\x12\x0f\n\x07Threads\x18\x03 \x01(\x05\x12\x0e\n\x06Tokens\x18\x04 \x01(\x05\x12\x0c\n\x04TopK\x18\x05 \x01(\x05\x12\x0e\n\x06Repeat\x18\x06 \x01(\x05\x12\r\n\x05\x42\x61tch\x18\x07 \x01(\x05\x12\r\n\x05NKeep\x18\x08 \x01(\x05\x12\x13\n\x0bTemperature\x18\t \x01(\x02\x12\x0f\n\x07Penalty\x18\n \x01(\x02\x12\r\n\x05\x46\x31\x36KV\x18\x0b \x01(\x08\x12\x11\n\tDebugMode\x18\x0c \x01(\x08\x12\x13\n\x0bStopPrompts\x18\r \x03(\t\x12\x11\n\tIgnoreEOS\x18\x0e \x01(\x08\x12\x19\n\x11TailFreeSamplingZ\x18\x0f \x01(\x02\x12\x10\n\x08TypicalP\x18\x10 \x01(\x02\x12\x18\n\x10\x46requencyPenalty\x18\x11 \x01(\x02\x12\x17\n\x0fPresencePenalty\x18\x12 \x01(\x02\x12\x10\n\x08Mirostat\x18\x13 \x01(\x05\x12\x13\n\x0bMirostatETA\x18\x14 \x01(\x02\x12\x13\n\x0bMirostatTAU\x18\x15 \x01(\x02\x12\x12\n\nPenalizeNL\x18\x16 \x01(\x08\x12\x11\n\tLogitBias\x18\x17 \x01(\t\x12\r\n\x05MLock\x18\x19 \x01(\x08\x12\x0c\n\x04MMap\x18\x1a \x01(\x08\x12\x16\n\x0ePromptCacheAll\x18\x1b \x01(\x08\x12\x15\n\rPromptCacheRO\x18\x1c \x01(\x08\x12\x0f\n\x07Grammar\x18\x1d \x01(\t\x12\x0f\n\x07MainGPU\x18\x1e \x01(\t\x12\x13\n\x0bTensorSplit\x18\x1f \x01(\t\x12\x0c\n\x04TopP\x18 \x01(\x02\x12\x17\n\x0fPromptCachePath\x18! \x01(\t\x12\r\n\x05\x44\x65\x62ug\x18\" \x01(\x08\x12\x17\n\x0f\x45mbeddingTokens\x18# \x03(\x05\x12\x12\n\nEmbeddings\x18$ \x01(\t\x12\x14\n\x0cRopeFreqBase\x18% \x01(\x02\x12\x15\n\rRopeFreqScale\x18& \x01(\x02\x12\x1b\n\x13NegativePromptScale\x18\' \x01(\x02\x12\x16\n\x0eNegativePrompt\x18( \x01(\t\"\x18\n\x05Reply\x12\x0f\n\x07message\x18\x01 \x01(\x0c\"\xc8\x03\n\x0cModelOptions\x12\r\n\x05Model\x18\x01 \x01(\t\x12\x13\n\x0b\x43ontextSize\x18\x02 \x01(\x05\x12\x0c\n\x04Seed\x18\x03 \x01(\x05\x12\x0e\n\x06NBatch\x18\x04 \x01(\x05\x12\x11\n\tF16Memory\x18\x05 \x01(\x08\x12\r\n\x05MLock\x18\x06 \x01(\x08\x12\x0c\n\x04MMap\x18\x07 \x01(\x08\x12\x11\n\tVocabOnly\x18\x08 \x01(\x08\x12\x0f\n\x07LowVRAM\x18\t \x01(\x08\x12\x12\n\nEmbeddings\x18\n \x01(\x08\x12\x0c\n\x04NUMA\x18\x0b \x01(\x08\x12\x12\n\nNGPULayers\x18\x0c \x01(\x05\x12\x0f\n\x07MainGPU\x18\r \x01(\t\x12\x13\n\x0bTensorSplit\x18\x0e \x01(\t\x12\x0f\n\x07Threads\x18\x0f \x01(\x05\x12\x19\n\x11LibrarySearchPath\x18\x10 \x01(\t\x12\x14\n\x0cRopeFreqBase\x18\x11 \x01(\x02\x12\x15\n\rRopeFreqScale\x18\x12 \x01(\x02\x12\x12\n\nRMSNormEps\x18\x13 \x01(\x02\x12\x0c\n\x04NGQA\x18\x14 \x01(\x05\x12\x11\n\tModelFile\x18\x15 \x01(\t\x12\x0e\n\x06\x44\x65vice\x18\x16 \x01(\t\x12\x11\n\tUseTriton\x18\x17 \x01(\x08\x12\x15\n\rModelBaseName\x18\x18 \x01(\t\"*\n\x06Result\x12\x0f\n\x07message\x18\x01 \x01(\t\x12\x0f\n\x07success\x18\x02 \x01(\x08\"%\n\x0f\x45mbeddingResult\x12\x12\n\nembeddings\x18\x01 \x03(\x02\"C\n\x11TranscriptRequest\x12\x0b\n\x03\x64st\x18\x02 \x01(\t\x12\x10\n\x08language\x18\x03 \x01(\t\x12\x0f\n\x07threads\x18\x04 \x01(\r\"N\n\x10TranscriptResult\x12,\n\x08segments\x18\x01 \x03(\x0b\x32\x1a.backend.TranscriptSegment\x12\x0c\n\x04text\x18\x02 \x01(\t\"Y\n\x11TranscriptSegment\x12\n\n\x02id\x18\x01 \x01(\x05\x12\r\n\x05start\x18\x02 \x01(\x03\x12\x0b\n\x03\x65nd\x18\x03 \x01(\x03\x12\x0c\n\x04text\x18\x04 \x01(\t\x12\x0e\n\x06tokens\x18\x05 \x03(\x05\"\x9e\x01\n\x14GenerateImageRequest\x12\x0e\n\x06height\x18\x01 \x01(\x05\x12\r\n\x05width\x18\x02 \x01(\x05\x12\x0c\n\x04mode\x18\x03 \x01(\x05\x12\x0c\n\x04step\x18\x04 \x01(\x05\x12\x0c\n\x04seed\x18\x05 \x01(\x05\x12\x17\n\x0fpositive_prompt\x18\x06 \x01(\t\x12\x17\n\x0fnegative_prompt\x18\x07 \x01(\t\x12\x0b\n\x03\x64st\x18\x08 \x01(\t\"6\n\nTTSRequest\x12\x0c\n\x04text\x18\x01 \x01(\t\x12\r\n\x05model\x18\x02 \x01(\t\x12\x0b\n\x03\x64st\x18\x03 \x01(\t2\xeb\x03\n\x07\x42\x61\x63kend\x12\x32\n\x06Health\x12\x16.backend.HealthMessage\x1a\x0e.backend.Reply\"\x00\x12\x34\n\x07Predict\x12\x17.backend.PredictOptions\x1a\x0e.backend.Reply\"\x00\x12\x35\n\tLoadModel\x12\x15.backend.ModelOptions\x1a\x0f.backend.Result\"\x00\x12<\n\rPredictStream\x12\x17.backend.PredictOptions\x1a\x0e.backend.Reply\"\x00\x30\x01\x12@\n\tEmbedding\x12\x17.backend.PredictOptions\x1a\x18.backend.EmbeddingResult\"\x00\x12\x41\n\rGenerateImage\x12\x1d.backend.GenerateImageRequest\x1a\x0f.backend.Result\"\x00\x12M\n\x12\x41udioTranscription\x12\x1a.backend.TranscriptRequest\x1a\x19.backend.TranscriptResult\"\x00\x12-\n\x03TTS\x12\x13.backend.TTSRequest\x1a\x0f.backend.Result\"\x00\x42Z\n\x19io.skynet.localai.backendB\x0eLocalAIBackendP\x01Z+github.com/go-skynet/LocalAI/pkg/grpc/protob\x06proto3')
|
||||
|
||||
_globals = globals()
|
||||
_builder.BuildMessageAndEnumDescriptors(DESCRIPTOR, _globals)
|
||||
|
|
@ -29,21 +29,21 @@ if _descriptor._USE_C_DESCRIPTORS == False:
|
|||
_globals['_REPLY']._serialized_start=820
|
||||
_globals['_REPLY']._serialized_end=844
|
||||
_globals['_MODELOPTIONS']._serialized_start=847
|
||||
_globals['_MODELOPTIONS']._serialized_end=1226
|
||||
_globals['_RESULT']._serialized_start=1228
|
||||
_globals['_RESULT']._serialized_end=1270
|
||||
_globals['_EMBEDDINGRESULT']._serialized_start=1272
|
||||
_globals['_EMBEDDINGRESULT']._serialized_end=1309
|
||||
_globals['_TRANSCRIPTREQUEST']._serialized_start=1311
|
||||
_globals['_TRANSCRIPTREQUEST']._serialized_end=1378
|
||||
_globals['_TRANSCRIPTRESULT']._serialized_start=1380
|
||||
_globals['_TRANSCRIPTRESULT']._serialized_end=1458
|
||||
_globals['_TRANSCRIPTSEGMENT']._serialized_start=1460
|
||||
_globals['_TRANSCRIPTSEGMENT']._serialized_end=1549
|
||||
_globals['_GENERATEIMAGEREQUEST']._serialized_start=1552
|
||||
_globals['_GENERATEIMAGEREQUEST']._serialized_end=1710
|
||||
_globals['_TTSREQUEST']._serialized_start=1712
|
||||
_globals['_TTSREQUEST']._serialized_end=1766
|
||||
_globals['_BACKEND']._serialized_start=1769
|
||||
_globals['_BACKEND']._serialized_end=2260
|
||||
_globals['_MODELOPTIONS']._serialized_end=1303
|
||||
_globals['_RESULT']._serialized_start=1305
|
||||
_globals['_RESULT']._serialized_end=1347
|
||||
_globals['_EMBEDDINGRESULT']._serialized_start=1349
|
||||
_globals['_EMBEDDINGRESULT']._serialized_end=1386
|
||||
_globals['_TRANSCRIPTREQUEST']._serialized_start=1388
|
||||
_globals['_TRANSCRIPTREQUEST']._serialized_end=1455
|
||||
_globals['_TRANSCRIPTRESULT']._serialized_start=1457
|
||||
_globals['_TRANSCRIPTRESULT']._serialized_end=1535
|
||||
_globals['_TRANSCRIPTSEGMENT']._serialized_start=1537
|
||||
_globals['_TRANSCRIPTSEGMENT']._serialized_end=1626
|
||||
_globals['_GENERATEIMAGEREQUEST']._serialized_start=1629
|
||||
_globals['_GENERATEIMAGEREQUEST']._serialized_end=1787
|
||||
_globals['_TTSREQUEST']._serialized_start=1789
|
||||
_globals['_TTSREQUEST']._serialized_end=1843
|
||||
_globals['_BACKEND']._serialized_start=1846
|
||||
_globals['_BACKEND']._serialized_end=2337
|
||||
# @@protoc_insertion_point(module_scope)
|
||||
|
|
|
|||
|
|
@ -18,7 +18,6 @@ class BackendServicer(backend_pb2_grpc.BackendServicer):
|
|||
return backend_pb2.Reply(message=bytes("OK", 'utf-8'))
|
||||
def LoadModel(self, request, context):
|
||||
model_name = request.Model
|
||||
model_name = os.path.basename(model_name)
|
||||
try:
|
||||
self.model = SentenceTransformer(model_name)
|
||||
except Exception as err:
|
||||
|
|
|
|||
|
|
@ -1,4 +1,6 @@
|
|||
sentence_transformers
|
||||
grpcio
|
||||
google
|
||||
protobuf
|
||||
protobuf
|
||||
https://github.com/PanQiWei/AutoGPTQ/releases/download/v0.3.0/auto_gptq-0.3.0+cu117-cp310-cp310-win_amd64.whl; platform_system == "Windows"
|
||||
https://github.com/PanQiWei/AutoGPTQ/releases/download/v0.3.0/auto_gptq-0.3.0+cu117-cp310-cp310-linux_x86_64.whl; platform_system == "Linux" and platform_machine == "x86_64"
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue