mirror of
https://github.com/QuantumNous/new-api.git
synced 2026-03-29 23:10:35 +00:00
f5b409d74f
- Introduced new OpenAI text models in `common/model.go`. - Added `IsOpenAITextModel` function to check for OpenAI text models. - Refactored token estimation methods across various channels to use estimated prompt tokens instead of direct prompt token counts. - Updated related functions and structures to accommodate the new token estimation approach, enhancing overall token management.
34 lines
1.2 KiB
Go
34 lines
1.2 KiB
Go
package service
|
|
|
|
import (
|
|
"github.com/QuantumNous/new-api/common"
|
|
"github.com/QuantumNous/new-api/constant"
|
|
"github.com/QuantumNous/new-api/dto"
|
|
"github.com/gin-gonic/gin"
|
|
)
|
|
|
|
//func GetPromptTokens(textRequest dto.GeneralOpenAIRequest, relayMode int) (int, error) {
|
|
// switch relayMode {
|
|
// case constant.RelayModeChatCompletions:
|
|
// return CountTokenMessages(textRequest.Messages, textRequest.Model)
|
|
// case constant.RelayModeCompletions:
|
|
// return CountTokenInput(textRequest.Prompt, textRequest.Model), nil
|
|
// case constant.RelayModeModerations:
|
|
// return CountTokenInput(textRequest.Input, textRequest.Model), nil
|
|
// }
|
|
// return 0, errors.New("unknown relay mode")
|
|
//}
|
|
|
|
func ResponseText2Usage(c *gin.Context, responseText string, modeName string, promptTokens int) *dto.Usage {
|
|
common.SetContextKey(c, constant.ContextKeyLocalCountTokens, true)
|
|
usage := &dto.Usage{}
|
|
usage.PromptTokens = promptTokens
|
|
usage.CompletionTokens = EstimateTokenByModel(modeName, responseText)
|
|
usage.TotalTokens = usage.PromptTokens + usage.CompletionTokens
|
|
return usage
|
|
}
|
|
|
|
func ValidUsage(usage *dto.Usage) bool {
|
|
return usage != nil && (usage.PromptTokens != 0 || usage.CompletionTokens != 0)
|
|
}
|