mirror of
https://github.com/mudler/LocalAI.git
synced 2024-06-07 19:40:48 +00:00
fix: use bytes in gRPC proto instead of strings (#813)
Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
This commit is contained in:
parent
0af0df7423
commit
b96e30e66c
@ -67,9 +67,9 @@ func ModelInference(ctx context.Context, s string, loader *model.ModelLoader, c
|
|||||||
opts.Prompt = s
|
opts.Prompt = s
|
||||||
if tokenCallback != nil {
|
if tokenCallback != nil {
|
||||||
ss := ""
|
ss := ""
|
||||||
err := inferenceModel.PredictStream(ctx, opts, func(s string) {
|
err := inferenceModel.PredictStream(ctx, opts, func(s []byte) {
|
||||||
tokenCallback(s)
|
tokenCallback(string(s))
|
||||||
ss += s
|
ss += string(s)
|
||||||
})
|
})
|
||||||
return ss, err
|
return ss, err
|
||||||
} else {
|
} else {
|
||||||
@ -77,7 +77,7 @@ func ModelInference(ctx context.Context, s string, loader *model.ModelLoader, c
|
|||||||
if err != nil {
|
if err != nil {
|
||||||
return "", err
|
return "", err
|
||||||
}
|
}
|
||||||
return reply.Message, err
|
return string(reply.Message), err
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
|
|
||||||
|
@ -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(\t\"\xac\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\"*\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\"\xac\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\"*\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()
|
_globals = globals()
|
||||||
_builder.BuildMessageAndEnumDescriptors(DESCRIPTOR, _globals)
|
_builder.BuildMessageAndEnumDescriptors(DESCRIPTOR, _globals)
|
||||||
|
@ -15,7 +15,7 @@ _ONE_DAY_IN_SECONDS = 60 * 60 * 24
|
|||||||
# Implement the BackendServicer class with the service methods
|
# Implement the BackendServicer class with the service methods
|
||||||
class BackendServicer(backend_pb2_grpc.BackendServicer):
|
class BackendServicer(backend_pb2_grpc.BackendServicer):
|
||||||
def Health(self, request, context):
|
def Health(self, request, context):
|
||||||
return backend_pb2.Reply(message="OK")
|
return backend_pb2.Reply(message=bytes("OK", 'utf-8'))
|
||||||
def LoadModel(self, request, context):
|
def LoadModel(self, request, context):
|
||||||
model_name = request.Model
|
model_name = request.Model
|
||||||
model_name = os.path.basename(model_name)
|
model_name = os.path.basename(model_name)
|
||||||
|
@ -42,7 +42,7 @@ func (c *Client) HealthCheck(ctx context.Context) bool {
|
|||||||
return false
|
return false
|
||||||
}
|
}
|
||||||
|
|
||||||
if res.Message == "OK" {
|
if string(res.Message) == "OK" {
|
||||||
return true
|
return true
|
||||||
}
|
}
|
||||||
return false
|
return false
|
||||||
@ -80,7 +80,7 @@ func (c *Client) LoadModel(ctx context.Context, in *pb.ModelOptions, opts ...grp
|
|||||||
return client.LoadModel(ctx, in, opts...)
|
return client.LoadModel(ctx, in, opts...)
|
||||||
}
|
}
|
||||||
|
|
||||||
func (c *Client) PredictStream(ctx context.Context, in *pb.PredictOptions, f func(s string), opts ...grpc.CallOption) error {
|
func (c *Client) PredictStream(ctx context.Context, in *pb.PredictOptions, f func(s []byte), opts ...grpc.CallOption) error {
|
||||||
conn, err := grpc.Dial(c.address, grpc.WithTransportCredentials(insecure.NewCredentials()))
|
conn, err := grpc.Dial(c.address, grpc.WithTransportCredentials(insecure.NewCredentials()))
|
||||||
if err != nil {
|
if err != nil {
|
||||||
return err
|
return err
|
||||||
|
@ -14,3 +14,7 @@ type LLM interface {
|
|||||||
AudioTranscription(*pb.TranscriptRequest) (api.Result, error)
|
AudioTranscription(*pb.TranscriptRequest) (api.Result, error)
|
||||||
TTS(*pb.TTSRequest) error
|
TTS(*pb.TTSRequest) error
|
||||||
}
|
}
|
||||||
|
|
||||||
|
func newReply(s string) *pb.Reply {
|
||||||
|
return &pb.Reply{Message: []byte(s)}
|
||||||
|
}
|
||||||
|
@ -416,7 +416,7 @@ type Reply struct {
|
|||||||
sizeCache protoimpl.SizeCache
|
sizeCache protoimpl.SizeCache
|
||||||
unknownFields protoimpl.UnknownFields
|
unknownFields protoimpl.UnknownFields
|
||||||
|
|
||||||
Message string `protobuf:"bytes,1,opt,name=message,proto3" json:"message,omitempty"`
|
Message []byte `protobuf:"bytes,1,opt,name=message,proto3" json:"message,omitempty"`
|
||||||
}
|
}
|
||||||
|
|
||||||
func (x *Reply) Reset() {
|
func (x *Reply) Reset() {
|
||||||
@ -451,11 +451,11 @@ func (*Reply) Descriptor() ([]byte, []int) {
|
|||||||
return file_pkg_grpc_proto_backend_proto_rawDescGZIP(), []int{2}
|
return file_pkg_grpc_proto_backend_proto_rawDescGZIP(), []int{2}
|
||||||
}
|
}
|
||||||
|
|
||||||
func (x *Reply) GetMessage() string {
|
func (x *Reply) GetMessage() []byte {
|
||||||
if x != nil {
|
if x != nil {
|
||||||
return x.Message
|
return x.Message
|
||||||
}
|
}
|
||||||
return ""
|
return nil
|
||||||
}
|
}
|
||||||
|
|
||||||
type ModelOptions struct {
|
type ModelOptions struct {
|
||||||
@ -1174,7 +1174,7 @@ var file_pkg_grpc_proto_backend_proto_rawDesc = []byte{
|
|||||||
0x69, 0x76, 0x65, 0x50, 0x72, 0x6f, 0x6d, 0x70, 0x74, 0x18, 0x28, 0x20, 0x01, 0x28, 0x09, 0x52,
|
0x69, 0x76, 0x65, 0x50, 0x72, 0x6f, 0x6d, 0x70, 0x74, 0x18, 0x28, 0x20, 0x01, 0x28, 0x09, 0x52,
|
||||||
0x0e, 0x4e, 0x65, 0x67, 0x61, 0x74, 0x69, 0x76, 0x65, 0x50, 0x72, 0x6f, 0x6d, 0x70, 0x74, 0x22,
|
0x0e, 0x4e, 0x65, 0x67, 0x61, 0x74, 0x69, 0x76, 0x65, 0x50, 0x72, 0x6f, 0x6d, 0x70, 0x74, 0x22,
|
||||||
0x21, 0x0a, 0x05, 0x52, 0x65, 0x70, 0x6c, 0x79, 0x12, 0x18, 0x0a, 0x07, 0x6d, 0x65, 0x73, 0x73,
|
0x21, 0x0a, 0x05, 0x52, 0x65, 0x70, 0x6c, 0x79, 0x12, 0x18, 0x0a, 0x07, 0x6d, 0x65, 0x73, 0x73,
|
||||||
0x61, 0x67, 0x65, 0x18, 0x01, 0x20, 0x01, 0x28, 0x09, 0x52, 0x07, 0x6d, 0x65, 0x73, 0x73, 0x61,
|
0x61, 0x67, 0x65, 0x18, 0x01, 0x20, 0x01, 0x28, 0x0c, 0x52, 0x07, 0x6d, 0x65, 0x73, 0x73, 0x61,
|
||||||
0x67, 0x65, 0x22, 0xca, 0x03, 0x0a, 0x0c, 0x4d, 0x6f, 0x64, 0x65, 0x6c, 0x4f, 0x70, 0x74, 0x69,
|
0x67, 0x65, 0x22, 0xca, 0x03, 0x0a, 0x0c, 0x4d, 0x6f, 0x64, 0x65, 0x6c, 0x4f, 0x70, 0x74, 0x69,
|
||||||
0x6f, 0x6e, 0x73, 0x12, 0x14, 0x0a, 0x05, 0x4d, 0x6f, 0x64, 0x65, 0x6c, 0x18, 0x01, 0x20, 0x01,
|
0x6f, 0x6e, 0x73, 0x12, 0x14, 0x0a, 0x05, 0x4d, 0x6f, 0x64, 0x65, 0x6c, 0x18, 0x01, 0x20, 0x01,
|
||||||
0x28, 0x09, 0x52, 0x05, 0x4d, 0x6f, 0x64, 0x65, 0x6c, 0x12, 0x20, 0x0a, 0x0b, 0x43, 0x6f, 0x6e,
|
0x28, 0x09, 0x52, 0x05, 0x4d, 0x6f, 0x64, 0x65, 0x6c, 0x12, 0x20, 0x0a, 0x0b, 0x43, 0x6f, 0x6e,
|
||||||
|
@ -65,7 +65,7 @@ message PredictOptions {
|
|||||||
|
|
||||||
// The response message containing the result
|
// The response message containing the result
|
||||||
message Reply {
|
message Reply {
|
||||||
string message = 1;
|
bytes message = 1;
|
||||||
}
|
}
|
||||||
|
|
||||||
message ModelOptions {
|
message ModelOptions {
|
||||||
|
@ -26,7 +26,7 @@ type server struct {
|
|||||||
}
|
}
|
||||||
|
|
||||||
func (s *server) Health(ctx context.Context, in *pb.HealthMessage) (*pb.Reply, error) {
|
func (s *server) Health(ctx context.Context, in *pb.HealthMessage) (*pb.Reply, error) {
|
||||||
return &pb.Reply{Message: "OK"}, nil
|
return newReply("OK"), nil
|
||||||
}
|
}
|
||||||
|
|
||||||
func (s *server) Embedding(ctx context.Context, in *pb.PredictOptions) (*pb.EmbeddingResult, error) {
|
func (s *server) Embedding(ctx context.Context, in *pb.PredictOptions) (*pb.EmbeddingResult, error) {
|
||||||
@ -48,7 +48,7 @@ func (s *server) LoadModel(ctx context.Context, in *pb.ModelOptions) (*pb.Result
|
|||||||
|
|
||||||
func (s *server) Predict(ctx context.Context, in *pb.PredictOptions) (*pb.Reply, error) {
|
func (s *server) Predict(ctx context.Context, in *pb.PredictOptions) (*pb.Reply, error) {
|
||||||
result, err := s.llm.Predict(in)
|
result, err := s.llm.Predict(in)
|
||||||
return &pb.Reply{Message: result}, err
|
return newReply(result), err
|
||||||
}
|
}
|
||||||
|
|
||||||
func (s *server) GenerateImage(ctx context.Context, in *pb.GenerateImageRequest) (*pb.Result, error) {
|
func (s *server) GenerateImage(ctx context.Context, in *pb.GenerateImageRequest) (*pb.Result, error) {
|
||||||
@ -99,7 +99,7 @@ func (s *server) PredictStream(in *pb.PredictOptions, stream pb.Backend_PredictS
|
|||||||
done := make(chan bool)
|
done := make(chan bool)
|
||||||
go func() {
|
go func() {
|
||||||
for result := range resultChan {
|
for result := range resultChan {
|
||||||
stream.Send(&pb.Reply{Message: result})
|
stream.Send(newReply(result))
|
||||||
}
|
}
|
||||||
done <- true
|
done <- true
|
||||||
}()
|
}()
|
||||||
|
Loading…
Reference in New Issue
Block a user