mirror of
https://github.com/mudler/LocalAI.git
synced 2025-05-20 02:24:59 +00:00
feat: Use ubuntu as base for container images, drop deprecated ggml-transformers backends (#1689)
* cleanup backends * switch image to ubuntu 22.04 * adapt commands for ubuntu * transformers cleanup * no contrib on ubuntu * Change test model to gguf * ci: disable bark tests (too cpu-intensive) Signed-off-by: Ettore Di Giacinto <mudler@localai.io> * cleanup * refinements * use intel base image * Makefile: Add docker targets * Change test model --------- Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
This commit is contained in:
Ettore Di Giacinto
GitHub
parent
d0a6a35b55
commit
ddd21f1644
20 changed files with 161 additions and 466 deletions
|
|
@ -1,44 +0,0 @@
|
|||
package transformers
|
||||
|
||||
// This is a wrapper to statisfy the GRPC service interface
|
||||
// It is meant to be used by the main executable that is the server for the specific backend type (falcon, gpt3, etc)
|
||||
import (
|
||||
"fmt"
|
||||
|
||||
"github.com/go-skynet/LocalAI/pkg/grpc/base"
|
||||
pb "github.com/go-skynet/LocalAI/pkg/grpc/proto"
|
||||
|
||||
transformers "github.com/go-skynet/go-ggml-transformers.cpp"
|
||||
)
|
||||
|
||||
type Dolly struct {
|
||||
base.SingleThread
|
||||
|
||||
dolly *transformers.Dolly
|
||||
}
|
||||
|
||||
func (llm *Dolly) Load(opts *pb.ModelOptions) error {
|
||||
model, err := transformers.NewDolly(opts.ModelFile)
|
||||
llm.dolly = model
|
||||
return err
|
||||
}
|
||||
|
||||
func (llm *Dolly) Predict(opts *pb.PredictOptions) (string, error) {
|
||||
return llm.dolly.Predict(opts.Prompt, buildPredictOptions(opts)...)
|
||||
}
|
||||
|
||||
// fallback to Predict
|
||||
func (llm *Dolly) PredictStream(opts *pb.PredictOptions, results chan string) error {
|
||||
|
||||
go func() {
|
||||
res, err := llm.dolly.Predict(opts.Prompt, buildPredictOptions(opts)...)
|
||||
|
||||
if err != nil {
|
||||
fmt.Println("err: ", err)
|
||||
}
|
||||
results <- res
|
||||
close(results)
|
||||
}()
|
||||
|
||||
return nil
|
||||
}
|
||||
|
|
@ -1,42 +0,0 @@
|
|||
package transformers
|
||||
|
||||
// This is a wrapper to statisfy the GRPC service interface
|
||||
// It is meant to be used by the main executable that is the server for the specific backend type (falcon, gpt3, etc)
|
||||
import (
|
||||
"fmt"
|
||||
|
||||
"github.com/go-skynet/LocalAI/pkg/grpc/base"
|
||||
pb "github.com/go-skynet/LocalAI/pkg/grpc/proto"
|
||||
|
||||
transformers "github.com/go-skynet/go-ggml-transformers.cpp"
|
||||
)
|
||||
|
||||
type GPT2 struct {
|
||||
base.SingleThread
|
||||
|
||||
gpt2 *transformers.GPT2
|
||||
}
|
||||
|
||||
func (llm *GPT2) Load(opts *pb.ModelOptions) error {
|
||||
model, err := transformers.New(opts.ModelFile)
|
||||
llm.gpt2 = model
|
||||
return err
|
||||
}
|
||||
|
||||
func (llm *GPT2) Predict(opts *pb.PredictOptions) (string, error) {
|
||||
return llm.gpt2.Predict(opts.Prompt, buildPredictOptions(opts)...)
|
||||
}
|
||||
|
||||
// fallback to Predict
|
||||
func (llm *GPT2) PredictStream(opts *pb.PredictOptions, results chan string) error {
|
||||
go func() {
|
||||
res, err := llm.gpt2.Predict(opts.Prompt, buildPredictOptions(opts)...)
|
||||
|
||||
if err != nil {
|
||||
fmt.Println("err: ", err)
|
||||
}
|
||||
results <- res
|
||||
close(results)
|
||||
}()
|
||||
return nil
|
||||
}
|
||||
|
|
@ -1,42 +0,0 @@
|
|||
package transformers
|
||||
|
||||
// This is a wrapper to statisfy the GRPC service interface
|
||||
// It is meant to be used by the main executable that is the server for the specific backend type (falcon, gpt3, etc)
|
||||
import (
|
||||
"fmt"
|
||||
|
||||
"github.com/go-skynet/LocalAI/pkg/grpc/base"
|
||||
pb "github.com/go-skynet/LocalAI/pkg/grpc/proto"
|
||||
|
||||
transformers "github.com/go-skynet/go-ggml-transformers.cpp"
|
||||
)
|
||||
|
||||
type GPTJ struct {
|
||||
base.SingleThread
|
||||
|
||||
gptj *transformers.GPTJ
|
||||
}
|
||||
|
||||
func (llm *GPTJ) Load(opts *pb.ModelOptions) error {
|
||||
model, err := transformers.NewGPTJ(opts.ModelFile)
|
||||
llm.gptj = model
|
||||
return err
|
||||
}
|
||||
|
||||
func (llm *GPTJ) Predict(opts *pb.PredictOptions) (string, error) {
|
||||
return llm.gptj.Predict(opts.Prompt, buildPredictOptions(opts)...)
|
||||
}
|
||||
|
||||
// fallback to Predict
|
||||
func (llm *GPTJ) PredictStream(opts *pb.PredictOptions, results chan string) error {
|
||||
go func() {
|
||||
res, err := llm.gptj.Predict(opts.Prompt, buildPredictOptions(opts)...)
|
||||
|
||||
if err != nil {
|
||||
fmt.Println("err: ", err)
|
||||
}
|
||||
results <- res
|
||||
close(results)
|
||||
}()
|
||||
return nil
|
||||
}
|
||||
|
|
@ -1,42 +0,0 @@
|
|||
package transformers
|
||||
|
||||
// This is a wrapper to statisfy the GRPC service interface
|
||||
// It is meant to be used by the main executable that is the server for the specific backend type (falcon, gpt3, etc)
|
||||
import (
|
||||
"fmt"
|
||||
|
||||
"github.com/go-skynet/LocalAI/pkg/grpc/base"
|
||||
pb "github.com/go-skynet/LocalAI/pkg/grpc/proto"
|
||||
|
||||
transformers "github.com/go-skynet/go-ggml-transformers.cpp"
|
||||
)
|
||||
|
||||
type GPTNeoX struct {
|
||||
base.SingleThread
|
||||
|
||||
gptneox *transformers.GPTNeoX
|
||||
}
|
||||
|
||||
func (llm *GPTNeoX) Load(opts *pb.ModelOptions) error {
|
||||
model, err := transformers.NewGPTNeoX(opts.ModelFile)
|
||||
llm.gptneox = model
|
||||
return err
|
||||
}
|
||||
|
||||
func (llm *GPTNeoX) Predict(opts *pb.PredictOptions) (string, error) {
|
||||
return llm.gptneox.Predict(opts.Prompt, buildPredictOptions(opts)...)
|
||||
}
|
||||
|
||||
// fallback to Predict
|
||||
func (llm *GPTNeoX) PredictStream(opts *pb.PredictOptions, results chan string) error {
|
||||
go func() {
|
||||
res, err := llm.gptneox.Predict(opts.Prompt, buildPredictOptions(opts)...)
|
||||
|
||||
if err != nil {
|
||||
fmt.Println("err: ", err)
|
||||
}
|
||||
results <- res
|
||||
close(results)
|
||||
}()
|
||||
return nil
|
||||
}
|
||||
|
|
@ -1,42 +0,0 @@
|
|||
package transformers
|
||||
|
||||
// This is a wrapper to statisfy the GRPC service interface
|
||||
// It is meant to be used by the main executable that is the server for the specific backend type (falcon, gpt3, etc)
|
||||
import (
|
||||
"fmt"
|
||||
|
||||
"github.com/go-skynet/LocalAI/pkg/grpc/base"
|
||||
pb "github.com/go-skynet/LocalAI/pkg/grpc/proto"
|
||||
|
||||
transformers "github.com/go-skynet/go-ggml-transformers.cpp"
|
||||
)
|
||||
|
||||
type MPT struct {
|
||||
base.SingleThread
|
||||
|
||||
mpt *transformers.MPT
|
||||
}
|
||||
|
||||
func (llm *MPT) Load(opts *pb.ModelOptions) error {
|
||||
model, err := transformers.NewMPT(opts.ModelFile)
|
||||
llm.mpt = model
|
||||
return err
|
||||
}
|
||||
|
||||
func (llm *MPT) Predict(opts *pb.PredictOptions) (string, error) {
|
||||
return llm.mpt.Predict(opts.Prompt, buildPredictOptions(opts)...)
|
||||
}
|
||||
|
||||
// fallback to Predict
|
||||
func (llm *MPT) PredictStream(opts *pb.PredictOptions, results chan string) error {
|
||||
go func() {
|
||||
res, err := llm.mpt.Predict(opts.Prompt, buildPredictOptions(opts)...)
|
||||
|
||||
if err != nil {
|
||||
fmt.Println("err: ", err)
|
||||
}
|
||||
results <- res
|
||||
close(results)
|
||||
}()
|
||||
return nil
|
||||
}
|
||||
|
|
@ -1,26 +0,0 @@
|
|||
package transformers
|
||||
|
||||
import (
|
||||
pb "github.com/go-skynet/LocalAI/pkg/grpc/proto"
|
||||
transformers "github.com/go-skynet/go-ggml-transformers.cpp"
|
||||
)
|
||||
|
||||
func buildPredictOptions(opts *pb.PredictOptions) []transformers.PredictOption {
|
||||
predictOptions := []transformers.PredictOption{
|
||||
transformers.SetTemperature(float64(opts.Temperature)),
|
||||
transformers.SetTopP(float64(opts.TopP)),
|
||||
transformers.SetTopK(int(opts.TopK)),
|
||||
transformers.SetTokens(int(opts.Tokens)),
|
||||
transformers.SetThreads(int(opts.Threads)),
|
||||
}
|
||||
|
||||
if opts.Batch != 0 {
|
||||
predictOptions = append(predictOptions, transformers.SetBatch(int(opts.Batch)))
|
||||
}
|
||||
|
||||
if opts.Seed != 0 {
|
||||
predictOptions = append(predictOptions, transformers.SetSeed(int(opts.Seed)))
|
||||
}
|
||||
|
||||
return predictOptions
|
||||
}
|
||||
|
|
@ -1,42 +0,0 @@
|
|||
package transformers
|
||||
|
||||
// This is a wrapper to statisfy the GRPC service interface
|
||||
// It is meant to be used by the main executable that is the server for the specific backend type (falcon, gpt3, etc)
|
||||
import (
|
||||
"fmt"
|
||||
|
||||
"github.com/go-skynet/LocalAI/pkg/grpc/base"
|
||||
pb "github.com/go-skynet/LocalAI/pkg/grpc/proto"
|
||||
|
||||
transformers "github.com/go-skynet/go-ggml-transformers.cpp"
|
||||
)
|
||||
|
||||
type Replit struct {
|
||||
base.SingleThread
|
||||
|
||||
replit *transformers.Replit
|
||||
}
|
||||
|
||||
func (llm *Replit) Load(opts *pb.ModelOptions) error {
|
||||
model, err := transformers.NewReplit(opts.ModelFile)
|
||||
llm.replit = model
|
||||
return err
|
||||
}
|
||||
|
||||
func (llm *Replit) Predict(opts *pb.PredictOptions) (string, error) {
|
||||
return llm.replit.Predict(opts.Prompt, buildPredictOptions(opts)...)
|
||||
}
|
||||
|
||||
// fallback to Predict
|
||||
func (llm *Replit) PredictStream(opts *pb.PredictOptions, results chan string) error {
|
||||
go func() {
|
||||
res, err := llm.replit.Predict(opts.Prompt, buildPredictOptions(opts)...)
|
||||
|
||||
if err != nil {
|
||||
fmt.Println("err: ", err)
|
||||
}
|
||||
results <- res
|
||||
close(results)
|
||||
}()
|
||||
return nil
|
||||
}
|
||||
|
|
@ -1,43 +0,0 @@
|
|||
package transformers
|
||||
|
||||
// This is a wrapper to statisfy the GRPC service interface
|
||||
// It is meant to be used by the main executable that is the server for the specific backend type (falcon, gpt3, etc)
|
||||
import (
|
||||
"fmt"
|
||||
|
||||
"github.com/go-skynet/LocalAI/pkg/grpc/base"
|
||||
pb "github.com/go-skynet/LocalAI/pkg/grpc/proto"
|
||||
|
||||
transformers "github.com/go-skynet/go-ggml-transformers.cpp"
|
||||
)
|
||||
|
||||
type Starcoder struct {
|
||||
base.SingleThread
|
||||
|
||||
starcoder *transformers.Starcoder
|
||||
}
|
||||
|
||||
func (llm *Starcoder) Load(opts *pb.ModelOptions) error {
|
||||
model, err := transformers.NewStarcoder(opts.ModelFile)
|
||||
llm.starcoder = model
|
||||
return err
|
||||
}
|
||||
|
||||
func (llm *Starcoder) Predict(opts *pb.PredictOptions) (string, error) {
|
||||
return llm.starcoder.Predict(opts.Prompt, buildPredictOptions(opts)...)
|
||||
}
|
||||
|
||||
// fallback to Predict
|
||||
func (llm *Starcoder) PredictStream(opts *pb.PredictOptions, results chan string) error {
|
||||
go func() {
|
||||
res, err := llm.starcoder.Predict(opts.Prompt, buildPredictOptions(opts)...)
|
||||
|
||||
if err != nil {
|
||||
fmt.Println("err: ", err)
|
||||
}
|
||||
results <- res
|
||||
close(results)
|
||||
}()
|
||||
|
||||
return nil
|
||||
}
|
||||
Loading…
Add table
Add a link
Reference in a new issue