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

Revert #1963 (#2056)

Browse source 
* Revert "fix(fncall): fix regression introduced in #1963 (#2048)"

This reverts commit 6b06d4e0af.

* Revert "fix: action-tmate back to upstream, dead code removal (#2038)"

This reverts commit fdec8a9d00.

* Revert "feat(grpc): return consumed token count and update response accordingly (#2035)"

This reverts commit e843d7df0e.

* Revert "refactor: backend/service split, channel-based llm flow (#1963)"

This reverts commit eed5706994.

* feat(grpc): return consumed token count and update response accordingly

Fixes: #1920

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 2024-04-17 23:33:49 +02:00 • committed by GitHub
parent af8c705ecd
commit af9e5a2d05
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
52 changed files with 2295 additions and 3065 deletions
Download patch file Download diff file Expand all files Collapse all files

90
core/backend/embeddings.go
View file

@ -2,100 +2,14 @@ package backend
import (
"fmt"
"time"
"github.com/go-skynet/LocalAI/core/config"
"github.com/go-skynet/LocalAI/core/schema"
"github.com/google/uuid"
"github.com/go-skynet/LocalAI/pkg/concurrency"
"github.com/go-skynet/LocalAI/pkg/grpc"
"github.com/go-skynet/LocalAI/pkg/model"
model "github.com/go-skynet/LocalAI/pkg/model"
)
type EmbeddingsBackendService struct {
ml *model.ModelLoader
bcl *config.BackendConfigLoader
appConfig *config.ApplicationConfig
}
func NewEmbeddingsBackendService(ml *model.ModelLoader, bcl *config.BackendConfigLoader, appConfig *config.ApplicationConfig) *EmbeddingsBackendService {
return &EmbeddingsBackendService{
ml: ml,
bcl: bcl,
appConfig: appConfig,
}
}
func (ebs *EmbeddingsBackendService) Embeddings(request *schema.OpenAIRequest) <-chan concurrency.ErrorOr[*schema.OpenAIResponse] {
resultChannel := make(chan concurrency.ErrorOr[*schema.OpenAIResponse])
go func(request *schema.OpenAIRequest) {
if request.Model == "" {
request.Model = model.StableDiffusionBackend
}
bc, request, err := ebs.bcl.LoadBackendConfigForModelAndOpenAIRequest(request.Model, request, ebs.appConfig)
if err != nil {
resultChannel <- concurrency.ErrorOr[*schema.OpenAIResponse]{Error: err}
close(resultChannel)
return
}
items := []schema.Item{}
for i, s := range bc.InputToken {
// get the model function to call for the result
embedFn, err := modelEmbedding("", s, ebs.ml, bc, ebs.appConfig)
if err != nil {
resultChannel <- concurrency.ErrorOr[*schema.OpenAIResponse]{Error: err}
close(resultChannel)
return
}
embeddings, err := embedFn()
if err != nil {
resultChannel <- concurrency.ErrorOr[*schema.OpenAIResponse]{Error: err}
close(resultChannel)
return
}
items = append(items, schema.Item{Embedding: embeddings, Index: i, Object: "embedding"})
}
for i, s := range bc.InputStrings {
// get the model function to call for the result
embedFn, err := modelEmbedding(s, []int{}, ebs.ml, bc, ebs.appConfig)
if err != nil {
resultChannel <- concurrency.ErrorOr[*schema.OpenAIResponse]{Error: err}
close(resultChannel)
return
}
embeddings, err := embedFn()
if err != nil {
resultChannel <- concurrency.ErrorOr[*schema.OpenAIResponse]{Error: err}
close(resultChannel)
return
}
items = append(items, schema.Item{Embedding: embeddings, Index: i, Object: "embedding"})
}
id := uuid.New().String()
created := int(time.Now().Unix())
resp := &schema.OpenAIResponse{
ID: id,
Created: created,
Model: request.Model, // we have to return what the user sent here, due to OpenAI spec.
Data: items,
Object: "list",
}
resultChannel <- concurrency.ErrorOr[*schema.OpenAIResponse]{Value: resp}
close(resultChannel)
}(request)
return resultChannel
}
func modelEmbedding(s string, tokens []int, loader *model.ModelLoader, backendConfig *config.BackendConfig, appConfig *config.ApplicationConfig) (func() ([]float32, error), error) {
func ModelEmbedding(s string, tokens []int, loader *model.ModelLoader, backendConfig config.BackendConfig, appConfig *config.ApplicationConfig) (func() ([]float32, error), error) {
modelFile := backendConfig.Model
grpcOpts := gRPCModelOpts(backendConfig)