dearwolf/LocalAI
1
0
Fork 0
mirror of https://github.com/mudler/LocalAI.git synced 2025-05-20 10:35:01 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 21

refactor: move backends into the backends directory (#1279)

Browse source 
* refactor: move backends into the backends directory

Signed-off-by: Ettore Di Giacinto <mudler@localai.io>

* refactor: move main close to implementation for every backend

Signed-off-by: Ettore Di Giacinto <mudler@localai.io>

---------

Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
This commit is contained in:
Ettore Di Giacinto 2023-11-13 22:40:16 +01:00 • committed by GitHub
parent 55461188a4
commit ad0e30bca5
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
102 changed files with 156 additions and 190 deletions
Download patch file Download diff file Expand all files Collapse all files

11
backend/python/vllm/Makefile Normal file
View file

@ -0,0 +1,11 @@
.PONY: vllm
vllm:
@echo "Creating virtual environment..."
@conda env create --name vllm --file vllm.yml
@echo "Virtual environment created."
.PONY: run
run:
@echo "Running vllm..."
bash run.sh
@echo "vllm run."

5
backend/python/vllm/README.md Normal file
View file

@ -0,0 +1,5 @@
# Creating a separate environment for the vllm project
```
make vllm
```

61
backend/python/vllm/backend_pb2.py Normal file
View file

File diff suppressed because one or more lines are too long

363
backend/python/vllm/backend_pb2_grpc.py Normal file
View file

@ -0,0 +1,363 @@
# 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,
)
self.TokenizeString = channel.unary_unary(
'/backend.Backend/TokenizeString',
request_serializer=backend__pb2.PredictOptions.SerializeToString,
response_deserializer=backend__pb2.TokenizationResponse.FromString,
)
self.Status = channel.unary_unary(
'/backend.Backend/Status',
request_serializer=backend__pb2.HealthMessage.SerializeToString,
response_deserializer=backend__pb2.StatusResponse.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 TokenizeString(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 Status(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,
),
'TokenizeString': grpc.unary_unary_rpc_method_handler(
servicer.TokenizeString,
request_deserializer=backend__pb2.PredictOptions.FromString,
response_serializer=backend__pb2.TokenizationResponse.SerializeToString,
),
'Status': grpc.unary_unary_rpc_method_handler(
servicer.Status,
request_deserializer=backend__pb2.HealthMessage.FromString,
response_serializer=backend__pb2.StatusResponse.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)
@staticmethod
def TokenizeString(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/TokenizeString',
backend__pb2.PredictOptions.SerializeToString,
backend__pb2.TokenizationResponse.FromString,
options, channel_credentials,
insecure, call_credentials, compression, wait_for_ready, timeout, metadata)
@staticmethod
def Status(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/Status',
backend__pb2.HealthMessage.SerializeToString,
backend__pb2.StatusResponse.FromString,
options, channel_credentials,
insecure, call_credentials, compression, wait_for_ready, timeout, metadata)

162
backend/python/vllm/backend_vllm.py Normal file
View file

@ -0,0 +1,162 @@
#!/usr/bin/env python3
from concurrent import futures
import time
import argparse
import signal
import sys
import os
import backend_pb2
import backend_pb2_grpc
import grpc
from vllm import LLM, SamplingParams
_ONE_DAY_IN_SECONDS = 60 * 60 * 24
# If MAX_WORKERS are specified in the environment use it, otherwise default to 1
MAX_WORKERS = int(os.environ.get('PYTHON_GRPC_MAX_WORKERS', '1'))
# Implement the BackendServicer class with the service methods
class BackendServicer(backend_pb2_grpc.BackendServicer):
"""
A gRPC servicer that implements the Backend service defined in backend.proto.
"""
def generate(self,prompt, max_new_tokens):
"""
Generates text based on the given prompt and maximum number of new tokens.
Args:
prompt (str): The prompt to generate text from.
max_new_tokens (int): The maximum number of new tokens to generate.
Returns:
str: The generated text.
"""
self.generator.end_beam_search()
# Tokenizing the input
ids = self.generator.tokenizer.encode(prompt)
self.generator.gen_begin_reuse(ids)
initial_len = self.generator.sequence[0].shape[0]
has_leading_space = False
decoded_text = ''
for i in range(max_new_tokens):
token = self.generator.gen_single_token()
if i == 0 and self.generator.tokenizer.tokenizer.IdToPiece(int(token)).startswith('▁'):
has_leading_space = True
decoded_text = self.generator.tokenizer.decode(self.generator.sequence[0][initial_len:])
if has_leading_space:
decoded_text = ' ' + decoded_text
if token.item() == self.generator.tokenizer.eos_token_id:
break
return decoded_text
def Health(self, request, context):
"""
Returns a health check message.
Args:
request: The health check request.
context: The gRPC context.
Returns:
backend_pb2.Reply: The health check reply.
"""
return backend_pb2.Reply(message=bytes("OK", 'utf-8'))
def LoadModel(self, request, context):
"""
Loads a language model.
Args:
request: The load model request.
context: The gRPC context.
Returns:
backend_pb2.Result: The load model result.
"""
try:
if request.Quantization != "":
self.llm = LLM(model=request.Model, quantization=request.Quantization)
else:
self.llm = LLM(model=request.Model)
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):
"""
Generates text based on the given prompt and sampling parameters.
Args:
request: The predict request.
context: The gRPC context.
Returns:
backend_pb2.Result: The predict result.
"""
if request.TopP == 0:
request.TopP = 0.9
sampling_params = SamplingParams(temperature=request.Temperature, top_p=request.TopP)
outputs = self.llm.generate([request.Prompt], sampling_params)
generated_text = outputs[0].outputs[0].text
# Remove prompt from response if present
if request.Prompt in generated_text:
generated_text = generated_text.replace(request.Prompt, "")
return backend_pb2.Result(message=bytes(generated_text, encoding='utf-8'))
def PredictStream(self, request, context):
"""
Generates text based on the given prompt and sampling parameters, and streams the results.
Args:
request: The predict stream request.
context: The gRPC context.
Returns:
backend_pb2.Result: The predict stream result.
"""
# 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=MAX_WORKERS))
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)

14
backend/python/vllm/run.sh Executable file
View file

@ -0,0 +1,14 @@
#!/bin/bash
##
## A bash script wrapper that runs the diffusers server with conda
export PATH=$PATH:/opt/conda/bin
# Activate conda environment
source activate vllm
# get the directory where the bash script is located
DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )" >/dev/null 2>&1 && pwd )"
python $DIR/backend_vllm.py $@

41
backend/python/vllm/test_backend_vllm.py Normal file
View file

@ -0,0 +1,41 @@
import unittest
import subprocess
import time
import backend_pb2
import backend_pb2_grpc
import grpc
import unittest
import subprocess
import time
import grpc
import backend_pb2_grpc
import backend_pb2
class TestBackendServicer(unittest.TestCase):
"""
TestBackendServicer is the class that tests the gRPC service.
This class contains methods to test the startup and shutdown of the gRPC service.
"""
def setUp(self):
self.service = subprocess.Popen(["python", "backend_vllm.py", "--addr", "localhost:50051"])
def tearDown(self) -> None:
self.service.terminate()
self.service.wait()
def test_server_startup(self):
time.sleep(2)
try:
self.setUp()
with grpc.insecure_channel("localhost:50051") as channel:
stub = backend_pb2_grpc.BackendStub(channel)
response = stub.Health(backend_pb2.HealthMessage())
self.assertEqual(response.message, b'OK')
except Exception as err:
print(err)
self.fail("Server failed to start")
finally:
self.tearDown()

99
backend/python/vllm/vllm.yml Normal file
View file

@ -0,0 +1,99 @@
name: vllm
channels:
- defaults
dependencies:
- _libgcc_mutex=0.1=main
- _openmp_mutex=5.1=1_gnu
- bzip2=1.0.8=h7b6447c_0
- ca-certificates=2023.08.22=h06a4308_0
- ld_impl_linux-64=2.38=h1181459_1
- libffi=3.4.4=h6a678d5_0
- libgcc-ng=11.2.0=h1234567_1
- libgomp=11.2.0=h1234567_1
- libstdcxx-ng=11.2.0=h1234567_1
- libuuid=1.41.5=h5eee18b_0
- ncurses=6.4=h6a678d5_0
- openssl=3.0.11=h7f8727e_2
- pip=23.2.1=py311h06a4308_0
- python=3.11.5=h955ad1f_0
- readline=8.2=h5eee18b_0
- setuptools=68.0.0=py311h06a4308_0
- sqlite=3.41.2=h5eee18b_0
- tk=8.6.12=h1ccaba5_0
- wheel=0.41.2=py311h06a4308_0
- xz=5.4.2=h5eee18b_0
- zlib=1.2.13=h5eee18b_0
- pip:
- aiosignal==1.3.1
- anyio==3.7.1
- attrs==23.1.0
- certifi==2023.7.22
- charset-normalizer==3.3.0
- click==8.1.7
- cmake==3.27.6
- fastapi==0.103.2
- filelock==3.12.4
- frozenlist==1.4.0
- fsspec==2023.9.2
- grpcio==1.59.0
- h11==0.14.0
- httptools==0.6.0
- huggingface-hub==0.17.3
- idna==3.4
- jinja2==3.1.2
- jsonschema==4.19.1
- jsonschema-specifications==2023.7.1
- lit==17.0.2
- markupsafe==2.1.3
- mpmath==1.3.0
- msgpack==1.0.7
- networkx==3.1
- ninja==1.11.1
- numpy==1.26.0
- nvidia-cublas-cu11==11.10.3.66
- nvidia-cuda-cupti-cu11==11.7.101
- nvidia-cuda-nvrtc-cu11==11.7.99
- nvidia-cuda-runtime-cu11==11.7.99
- nvidia-cudnn-cu11==8.5.0.96
- nvidia-cufft-cu11==10.9.0.58
- nvidia-curand-cu11==10.2.10.91
- nvidia-cusolver-cu11==11.4.0.1
- nvidia-cusparse-cu11==11.7.4.91
- nvidia-nccl-cu11==2.14.3
- nvidia-nvtx-cu11==11.7.91
- packaging==23.2
- pandas==2.1.1
- protobuf==4.24.4
- psutil==5.9.5
- pyarrow==13.0.0
- pydantic==1.10.13
- python-dateutil==2.8.2
- python-dotenv==1.0.0
- pytz==2023.3.post1
- pyyaml==6.0.1
- ray==2.7.0
- referencing==0.30.2
- regex==2023.10.3
- requests==2.31.0
- rpds-py==0.10.4
- safetensors==0.4.0
- sentencepiece==0.1.99
- six==1.16.0
- sniffio==1.3.0
- starlette==0.27.0
- sympy==1.12
- tokenizers==0.14.1
- torch==2.0.1
- tqdm==4.66.1
- transformers==4.34.0
- triton==2.0.0
- typing-extensions==4.8.0
- tzdata==2023.3
- urllib3==2.0.6
- uvicorn==0.23.2
- uvloop==0.17.0
- vllm==0.2.0
- watchfiles==0.20.0
- websockets==11.0.3
- xformers==0.0.22
prefix: /opt/conda/envs/vllm