dearwolf/LocalAI
1
0
Fork 0
mirror of https://github.com/mudler/LocalAI.git synced 2025-05-28 14:35:00 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 12

feat: add machine tag and inference timings (#4577)

Browse source 
* Add machine tag option, add extraUsage option, grpc-server -> proto -> endpoint extraUsage data is broken for now

Signed-off-by: mintyleaf <mintyleafdev@gmail.com>

* remove redurant timing fields, fix not working timings output

Signed-off-by: mintyleaf <mintyleafdev@gmail.com>

* use middleware for Machine-Tag only if tag is specified

Signed-off-by: mintyleaf <mintyleafdev@gmail.com>

---------

Signed-off-by: mintyleaf <mintyleafdev@gmail.com>
This commit is contained in:
mintyleaf 2025-01-17 20:05:58 +04:00 • committed by GitHub
parent 8027fdf1c7
commit 96f8ec0402
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
15 changed files with 137 additions and 48 deletions
Download patch file Download diff file Expand all files Collapse all files

59
core/http/endpoints/openai/chat.go
View file

@ -30,7 +30,7 @@ func ChatEndpoint(cl *config.BackendConfigLoader, ml *model.ModelLoader, evaluat
var id, textContentToReturn string
var created int
process := func(s string, req *schema.OpenAIRequest, config *config.BackendConfig, loader *model.ModelLoader, responses chan schema.OpenAIResponse) {
process := func(s string, req *schema.OpenAIRequest, config *config.BackendConfig, loader *model.ModelLoader, responses chan schema.OpenAIResponse, extraUsage bool) {
initialMessage := schema.OpenAIResponse{
ID: id,
Created: created,
@ -40,18 +40,24 @@ func ChatEndpoint(cl *config.BackendConfigLoader, ml *model.ModelLoader, evaluat
}
responses <- initialMessage
ComputeChoices(req, s, config, startupOptions, loader, func(s string, c *[]schema.Choice) {}, func(s string, usage backend.TokenUsage) bool {
ComputeChoices(req, s, config, startupOptions, loader, func(s string, c *[]schema.Choice) {}, func(s string, tokenUsage backend.TokenUsage) bool {
usage := schema.OpenAIUsage{
PromptTokens: tokenUsage.Prompt,
CompletionTokens: tokenUsage.Completion,
TotalTokens: tokenUsage.Prompt + tokenUsage.Completion,
}
if extraUsage {
usage.TimingTokenGeneration = tokenUsage.TimingTokenGeneration
usage.TimingPromptProcessing = tokenUsage.TimingPromptProcessing
}
resp := schema.OpenAIResponse{
ID: id,
Created: created,
Model: req.Model, // we have to return what the user sent here, due to OpenAI spec.
Choices: []schema.Choice{{Delta: &schema.Message{Content: &s}, Index: 0}},
Object: "chat.completion.chunk",
Usage: schema.OpenAIUsage{
PromptTokens: usage.Prompt,
CompletionTokens: usage.Completion,
TotalTokens: usage.Prompt + usage.Completion,
},
Usage: usage,
}
responses <- resp
@ -59,7 +65,7 @@ func ChatEndpoint(cl *config.BackendConfigLoader, ml *model.ModelLoader, evaluat
})
close(responses)
}
processTools := func(noAction string, prompt string, req *schema.OpenAIRequest, config *config.BackendConfig, loader *model.ModelLoader, responses chan schema.OpenAIResponse) {
processTools := func(noAction string, prompt string, req *schema.OpenAIRequest, config *config.BackendConfig, loader *model.ModelLoader, responses chan schema.OpenAIResponse, extraUsage bool) {
result := ""
_, tokenUsage, _ := ComputeChoices(req, prompt, config, startupOptions, loader, func(s string, c *[]schema.Choice) {}, func(s string, usage backend.TokenUsage) bool {
result += s
@ -90,6 +96,15 @@ func ChatEndpoint(cl *config.BackendConfigLoader, ml *model.ModelLoader, evaluat
log.Error().Err(err).Msg("error handling question")
return
}
usage := schema.OpenAIUsage{
PromptTokens: tokenUsage.Prompt,
CompletionTokens: tokenUsage.Completion,
TotalTokens: tokenUsage.Prompt + tokenUsage.Completion,
}
if extraUsage {
usage.TimingTokenGeneration = tokenUsage.TimingTokenGeneration
usage.TimingPromptProcessing = tokenUsage.TimingPromptProcessing
}
resp := schema.OpenAIResponse{
ID: id,
@ -97,11 +112,7 @@ func ChatEndpoint(cl *config.BackendConfigLoader, ml *model.ModelLoader, evaluat
Model: req.Model, // we have to return what the user sent here, due to OpenAI spec.
Choices: []schema.Choice{{Delta: &schema.Message{Content: &result}, Index: 0}},
Object: "chat.completion.chunk",
Usage: schema.OpenAIUsage{
PromptTokens: tokenUsage.Prompt,
CompletionTokens: tokenUsage.Completion,
TotalTokens: tokenUsage.Prompt + tokenUsage.Completion,
},
Usage: usage,
}
responses <- resp
@ -170,6 +181,9 @@ func ChatEndpoint(cl *config.BackendConfigLoader, ml *model.ModelLoader, evaluat
}
c.Set("X-Correlation-ID", correlationID)
// Opt-in extra usage flag
extraUsage := c.Get("LocalAI-Extra-Usage", "") != ""
modelFile, input, err := readRequest(c, cl, ml, startupOptions, true)
if err != nil {
return fmt.Errorf("failed reading parameters from request:%w", err)
@ -319,9 +333,9 @@ func ChatEndpoint(cl *config.BackendConfigLoader, ml *model.ModelLoader, evaluat
responses := make(chan schema.OpenAIResponse)
if !shouldUseFn {
go process(predInput, input, config, ml, responses)
go process(predInput, input, config, ml, responses, extraUsage)
} else {
go processTools(noActionName, predInput, input, config, ml, responses)
go processTools(noActionName, predInput, input, config, ml, responses, extraUsage)
}
c.Context().SetBodyStreamWriter(fasthttp.StreamWriter(func(w *bufio.Writer) {
@ -449,6 +463,15 @@ func ChatEndpoint(cl *config.BackendConfigLoader, ml *model.ModelLoader, evaluat
if err != nil {
return err
}
usage := schema.OpenAIUsage{
PromptTokens: tokenUsage.Prompt,
CompletionTokens: tokenUsage.Completion,
TotalTokens: tokenUsage.Prompt + tokenUsage.Completion,
}
if extraUsage {
usage.TimingTokenGeneration = tokenUsage.TimingTokenGeneration
usage.TimingPromptProcessing = tokenUsage.TimingPromptProcessing
}
resp := &schema.OpenAIResponse{
ID: id,
@ -456,11 +479,7 @@ func ChatEndpoint(cl *config.BackendConfigLoader, ml *model.ModelLoader, evaluat
Model: input.Model, // we have to return what the user sent here, due to OpenAI spec.
Choices: result,
Object: "chat.completion",
Usage: schema.OpenAIUsage{
PromptTokens: tokenUsage.Prompt,
CompletionTokens: tokenUsage.Completion,
TotalTokens: tokenUsage.Prompt + tokenUsage.Completion,
},
Usage: usage,
}
respData, _ := json.Marshal(resp)
log.Debug().Msgf("Response: %s", respData)