Merge 4a362caae6 into cba82404ae

Co-authored-by: Star <iii9777@163.com>

.github/workflows/docker-image-amd64.yml

@@ -1,61 +0,0 @@
-name: Publish Docker image (amd64)
-
-on:
-  push:
-    tags:
-      - 'v*.*.*'
-  workflow_dispatch:
-    inputs:
-      name:
-        description: 'reason'
-        required: false
-jobs:
-  push_to_registries:
-    name: Push Docker image to multiple registries
-    runs-on: ubuntu-latest
-    permissions:
-      packages: write
-      contents: read
-    steps:
-      - name: Check out the repo
-        uses: actions/checkout@v3
-
-      - name: Check repository URL
-        run: |
-          REPO_URL=$(git config --get remote.origin.url)
-          if [[ $REPO_URL == *"pro" ]]; then
-            exit 1
-          fi        
-
-      - name: Save version info
-        run: |
-          git describe --tags > VERSION 
-
-      - name: Log in to Docker Hub
-        uses: docker/login-action@v2
-        with:
-          username: ${{ secrets.DOCKERHUB_USERNAME }}
-          password: ${{ secrets.DOCKERHUB_TOKEN }}
-
-      - name: Log in to the Container registry
-        uses: docker/login-action@v2
-        with:
-          registry: ghcr.io
-          username: ${{ github.actor }}
-          password: ${{ secrets.GITHUB_TOKEN }}
-
-      - name: Extract metadata (tags, labels) for Docker
-        id: meta
-        uses: docker/metadata-action@v4
-        with:
-          images: |
-            justsong/one-api
-            ghcr.io/${{ github.repository }}
-
-      - name: Build and push Docker images
-        uses: docker/build-push-action@v3
-        with:
-          context: .
-          push: true
-          tags: ${{ steps.meta.outputs.tags }}
-          labels: ${{ steps.meta.outputs.labels }}

.github/workflows/docker-image-en.yml

@@ -1,4 +1,4 @@
-name: Publish Docker image (amd64, English)
+name: Publish Docker image (English)
 
 on:
   push:
@@ -34,6 +34,13 @@ jobs:
       - name: Translate
         run: |
           python ./i18n/translate.py --repository_path . --json_file_path ./i18n/en.json
+
+      - name: Set up QEMU
+        uses: docker/setup-qemu-action@v2
+
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v2
+
       - name: Log in to Docker Hub
         uses: docker/login-action@v2
         with:
@@ -51,6 +58,7 @@ jobs:
         uses: docker/build-push-action@v3
         with:
           context: .
+          platforms: linux/amd64,linux/arm64
           push: true
           tags: ${{ steps.meta.outputs.tags }}
           labels: ${{ steps.meta.outputs.labels }}

.github/workflows/docker-image.yml

@@ -1,10 +1,9 @@
-name: Publish Docker image (arm64)
+name: Publish Docker image
 
 on:
   push:
     tags:
       - 'v*.*.*'
-      - '!*-alpha*'
   workflow_dispatch:
     inputs:
       name:

Dockerfile

@@ -1,4 +1,4 @@
-FROM node:16 as builder
+FROM --platform=$BUILDPLATFORM node:16 AS builder
 
 WORKDIR /web
 COPY ./VERSION .

README.md

@@ -89,6 +89,7 @@ _✨ 通过标准的 OpenAI API 格式访问所有的大模型，开箱即用
    + [x] [DeepL](https://www.deepl.com/)
    + [x] [together.ai](https://www.together.ai/)
    + [x] [novita.ai](https://www.novita.ai/)
+   + [x] [硅基流动 SiliconCloud](https://siliconflow.cn/siliconcloud)
 2. 支持配置镜像以及众多[第三方代理服务](https://iamazing.cn/page/openai-api-third-party-services)。
 3. 支持通过**负载均衡**的方式访问多个渠道。
 4. 支持 **stream 模式**，可以通过流式传输实现打字机效果。
@@ -251,9 +252,9 @@ docker run --name chatgpt-web -d -p 3002:3002 -e OPENAI_API_BASE_URL=https://ope
 #### QChatGPT - QQ机器人
 项目主页：https://github.com/RockChinQ/QChatGPT
 
-根据文档完成部署后，在`config.py`设置配置项`openai_config`的`reverse_proxy`为 One API 后端地址，设置`api_key`为 One API 生成的key，并在配置项`completion_api_params`的`model`参数设置为 One API 支持的模型名称。
+根据[文档](https://qchatgpt.rockchin.top)完成部署后，在 `data/provider.json`设置`requester.openai-chat-completions.base-url`为 One API 实例地址，并填写 API Key 到 `keys.openai` 组中，设置 `model` 为要使用的模型名称。
 
-可安装 [Switcher 插件](https://github.com/RockChinQ/Switcher)在运行时切换所使用的模型。
+运行期间可以通过`!model`命令查看、切换可用模型。
 
 ### 部署到第三方平台
 <details>

model/log.go

@@ -3,6 +3,7 @@ package model
 import (
 	"context"
 	"fmt"
+
 	"github.com/songquanpeng/one-api/common"
 	"github.com/songquanpeng/one-api/common/config"
 	"github.com/songquanpeng/one-api/common/helper"
@@ -152,7 +153,11 @@ func SearchUserLogs(userId int, keyword string) (logs []*Log, err error) {
 }
 
 func SumUsedQuota(logType int, startTimestamp int64, endTimestamp int64, modelName string, username string, tokenName string, channel int) (quota int64) {
-	tx := LOG_DB.Table("logs").Select("ifnull(sum(quota),0)")
+	ifnull := "ifnull"
+	if common.UsingPostgreSQL {
+		ifnull = "COALESCE"
+	}
+	tx := LOG_DB.Table("logs").Select(fmt.Sprintf("%s(sum(quota),0)", ifnull))
 	if username != "" {
 		tx = tx.Where("username = ?", username)
 	}
@@ -176,7 +181,11 @@ func SumUsedQuota(logType int, startTimestamp int64, endTimestamp int64, modelNa
 }
 
 func SumUsedToken(logType int, startTimestamp int64, endTimestamp int64, modelName string, username string, tokenName string) (token int) {
-	tx := LOG_DB.Table("logs").Select("ifnull(sum(prompt_tokens),0) + ifnull(sum(completion_tokens),0)")
+	ifnull := "ifnull"
+	if common.UsingPostgreSQL {
+		ifnull = "COALESCE"
+	}
+	tx := LOG_DB.Table("logs").Select(fmt.Sprintf("%s(sum(prompt_tokens),0) + %s(sum(completion_tokens),0)", ifnull, ifnull))
 	if username != "" {
 		tx = tx.Where("username = ?", username)
 	}

model/token.go

@@ -30,7 +30,7 @@ type Token struct {
 	RemainQuota    int64   `json:"remain_quota" gorm:"bigint;default:0"`
 	UnlimitedQuota bool    `json:"unlimited_quota" gorm:"default:false"`
 	UsedQuota      int64   `json:"used_quota" gorm:"bigint;default:0"` // used quota
-	Models         *string `json:"models" gorm:"default:''"`           // allowed models
+	Models         *string `json:"models" gorm:"type:text"`           // allowed models
 	Subnet         *string `json:"subnet" gorm:"default:''"`           // allowed subnet
 }
 

monitor/manage.go

@@ -1,10 +1,11 @@
 package monitor
 
 import (
-	"github.com/songquanpeng/one-api/common/config"
-	"github.com/songquanpeng/one-api/relay/model"
 	"net/http"
 	"strings"
+
+	"github.com/songquanpeng/one-api/common/config"
+	"github.com/songquanpeng/one-api/relay/model"
 )
 
 func ShouldDisableChannel(err *model.Error, statusCode int) bool {
@@ -18,31 +19,23 @@ func ShouldDisableChannel(err *model.Error, statusCode int) bool {
 		return true
 	}
 	switch err.Type {
-	case "insufficient_quota":
+	case "insufficient_quota", "authentication_error", "permission_error", "forbidden":
-		return true
-	// https://docs.anthropic.com/claude/reference/errors
-	case "authentication_error":
-		return true
-	case "permission_error":
-		return true
-	case "forbidden":
 		return true
 	}
 	if err.Code == "invalid_api_key" || err.Code == "account_deactivated" {
 		return true
 	}
-	if strings.HasPrefix(err.Message, "Your credit balance is too low") { // anthropic
+
-		return true
+	lowerMessage := strings.ToLower(err.Message)
-	} else if strings.HasPrefix(err.Message, "This organization has been disabled.") {
+	if strings.Contains(lowerMessage, "your access was terminated") ||
-		return true
+		strings.Contains(lowerMessage, "violation of our policies") ||
-	}
+		strings.Contains(lowerMessage, "your credit balance is too low") ||
-	//if strings.Contains(err.Message, "quota") {
+		strings.Contains(lowerMessage, "organization has been disabled") ||
-	//	return true
+		strings.Contains(lowerMessage, "credit") ||
-	//}
+		strings.Contains(lowerMessage, "balance") ||
-	if strings.Contains(err.Message, "credit") {
+		strings.Contains(lowerMessage, "permission denied") ||
-		return true
+  	strings.Contains(lowerMessage, "organization has been restricted") || // groq
-	}
+		strings.Contains(lowerMessage, "已欠费") {
-	if strings.Contains(err.Message, "balance") {
 		return true
 	}
 	return false

relay/adaptor/ali/main.go

@@ -3,6 +3,7 @@ package ali
 import (
 	"bufio"
 	"encoding/json"
+	"github.com/songquanpeng/one-api/common/ctxkey"
 	"github.com/songquanpeng/one-api/common/render"
 	"io"
 	"net/http"
@@ -59,7 +60,7 @@ func ConvertRequest(request model.GeneralOpenAIRequest) *ChatRequest {
 
 func ConvertEmbeddingRequest(request model.GeneralOpenAIRequest) *EmbeddingRequest {
 	return &EmbeddingRequest{
-		Model: "text-embedding-v1",
+		Model: request.Model,
 		Input: struct {
 			Texts []string `json:"texts"`
 		}{
@@ -102,8 +103,9 @@ func EmbeddingHandler(c *gin.Context, resp *http.Response) (*model.ErrorWithStat
 			StatusCode: resp.StatusCode,
 		}, nil
 	}
-
+	requestModel := c.GetString(ctxkey.RequestModel)
 	fullTextResponse := embeddingResponseAli2OpenAI(&aliResponse)
+	fullTextResponse.Model = requestModel
 	jsonResponse, err := json.Marshal(fullTextResponse)
 	if err != nil {
 		return openai.ErrorWrapper(err, "marshal_response_body_failed", http.StatusInternalServerError), nil

relay/adaptor/cloudflare/constant.go

@@ -1,6 +1,7 @@
 package cloudflare
 
 var ModelList = []string{
+	"@cf/meta/llama-3.1-8b-instruct",
 	"@cf/meta/llama-2-7b-chat-fp16",
 	"@cf/meta/llama-2-7b-chat-int8",
 	"@cf/mistral/mistral-7b-instruct-v0.1",

relay/adaptor/doubao/main.go

@@ -7,8 +7,12 @@ import (
 )
 
 func GetRequestURL(meta *meta.Meta) (string, error) {
-	if meta.Mode == relaymode.ChatCompletions {
+	switch meta.Mode {
+	case relaymode.ChatCompletions:
 		return fmt.Sprintf("%s/api/v3/chat/completions", meta.BaseURL), nil
+	case relaymode.Embeddings:
+		return fmt.Sprintf("%s/api/v3/embeddings", meta.BaseURL), nil
+	default:
 	}
 	return "", fmt.Errorf("unsupported relay mode %d for doubao", meta.Mode)
 }

relay/adaptor/gemini/constants.go

@@ -3,6 +3,5 @@ package gemini
 // https://ai.google.dev/models/gemini
 
 var ModelList = []string{
-	"gemini-pro", "gemini-1.0-pro-001", "gemini-1.5-pro",
+	"gemini-pro", "gemini-1.0-pro", "gemini-1.5-flash", "gemini-1.5-pro", "text-embedding-004", "aqa",
-	"gemini-pro-vision", "gemini-1.0-pro-vision-001", "embedding-001", "text-embedding-004",
 }

relay/adaptor/groq/constants.go

@@ -4,9 +4,14 @@ package groq
 
 var ModelList = []string{
 	"gemma-7b-it",
-	"llama2-7b-2048",
-	"llama2-70b-4096",
 	"mixtral-8x7b-32768",
 	"llama3-8b-8192",
 	"llama3-70b-8192",
+	"gemma2-9b-it",
+	"llama-3.1-405b-reasoning",
+	"llama-3.1-70b-versatile",
+	"llama-3.1-8b-instant",
+	"llama3-groq-70b-8192-tool-use-preview",
+	"llama3-groq-8b-8192-tool-use-preview",
+	"whisper-large-v3",
 }

relay/adaptor/ollama/adaptor.go

@@ -24,7 +24,7 @@ func (a *Adaptor) GetRequestURL(meta *meta.Meta) (string, error) {
 	// https://github.com/ollama/ollama/blob/main/docs/api.md
 	fullRequestURL := fmt.Sprintf("%s/api/chat", meta.BaseURL)
 	if meta.Mode == relaymode.Embeddings {
-		fullRequestURL = fmt.Sprintf("%s/api/embeddings", meta.BaseURL)
+		fullRequestURL = fmt.Sprintf("%s/api/embed", meta.BaseURL)
 	}
 	return fullRequestURL, nil
 }

relay/adaptor/ollama/main.go

@@ -31,6 +31,8 @@ func ConvertRequest(request model.GeneralOpenAIRequest) *ChatRequest {
 			TopP:             request.TopP,
 			FrequencyPenalty: request.FrequencyPenalty,
 			PresencePenalty:  request.PresencePenalty,
+			NumPredict:  	  request.MaxTokens,
+			NumCtx:  	  request.NumCtx,
 		},
 		Stream: request.Stream,
 	}
@@ -118,8 +120,10 @@ func StreamHandler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusC
 	common.SetEventStreamHeaders(c)
 
 	for scanner.Scan() {
-		data := strings.TrimPrefix(scanner.Text(), "}")
+		data := scanner.Text()
-		data = data + "}"
+		if strings.HasPrefix(data, "}") {
+		    data = strings.TrimPrefix(data, "}") + "}"
+		}
 
 		var ollamaResponse ChatResponse
 		err := json.Unmarshal([]byte(data), &ollamaResponse)
@@ -157,8 +161,15 @@ func StreamHandler(c *gin.Context, resp *http.Response) (*model.ErrorWithStatusC
 
 func ConvertEmbeddingRequest(request model.GeneralOpenAIRequest) *EmbeddingRequest {
 	return &EmbeddingRequest{
-		Model:  request.Model,
+		Model: request.Model,
-		Prompt: strings.Join(request.ParseInput(), " "),
+		Input: request.ParseInput(),
+		Options: &Options{
+			Seed:             int(request.Seed),
+			Temperature:      request.Temperature,
+			TopP:             request.TopP,
+			FrequencyPenalty: request.FrequencyPenalty,
+			PresencePenalty:  request.PresencePenalty,
+		},
 	}
 }
 
@@ -201,15 +212,17 @@ func embeddingResponseOllama2OpenAI(response *EmbeddingResponse) *openai.Embeddi
 	openAIEmbeddingResponse := openai.EmbeddingResponse{
 		Object: "list",
 		Data:   make([]openai.EmbeddingResponseItem, 0, 1),
-		Model:  "text-embedding-v1",
+		Model:  response.Model,
 		Usage:  model.Usage{TotalTokens: 0},
 	}
 
-	openAIEmbeddingResponse.Data = append(openAIEmbeddingResponse.Data, openai.EmbeddingResponseItem{
+	for i, embedding := range response.Embeddings {
-		Object:    `embedding`,
+		openAIEmbeddingResponse.Data = append(openAIEmbeddingResponse.Data, openai.EmbeddingResponseItem{
-		Index:     0,
+			Object:    `embedding`,
-		Embedding: response.Embedding,
+			Index:     i,
-	})
+			Embedding: embedding,
+		})
+	}
 	return &openAIEmbeddingResponse
 }
 

relay/adaptor/ollama/model.go

@@ -7,6 +7,8 @@ type Options struct {
 	TopP             float64 `json:"top_p,omitempty"`
 	FrequencyPenalty float64 `json:"frequency_penalty,omitempty"`
 	PresencePenalty  float64 `json:"presence_penalty,omitempty"`
+	NumPredict  	 int 	 `json:"num_predict,omitempty"`
+	NumCtx  	 int 	 `json:"num_ctx,omitempty"`
 }
 
 type Message struct {
@@ -37,11 +39,15 @@ type ChatResponse struct {
 }
 
 type EmbeddingRequest struct {
-	Model  string `json:"model"`
+	Model string   `json:"model"`
-	Prompt string `json:"prompt"`
+	Input []string `json:"input"`
+	// Truncate  bool     `json:"truncate,omitempty"`
+	Options *Options `json:"options,omitempty"`
+	// KeepAlive string   `json:"keep_alive,omitempty"`
 }
 
 type EmbeddingResponse struct {
-	Error     string    `json:"error,omitempty"`
+	Error      string      `json:"error,omitempty"`
-	Embedding []float64 `json:"embedding,omitempty"`
+	Model      string      `json:"model"`
+	Embeddings [][]float64 `json:"embeddings"`
 }

relay/adaptor/openai/compatible.go

@@ -13,6 +13,7 @@ import (
 	"github.com/songquanpeng/one-api/relay/adaptor/novita"
 	"github.com/songquanpeng/one-api/relay/adaptor/stepfun"
 	"github.com/songquanpeng/one-api/relay/adaptor/togetherai"
+	"github.com/songquanpeng/one-api/relay/adaptor/siliconflow"
 	"github.com/songquanpeng/one-api/relay/channeltype"
 )
 
@@ -30,6 +31,7 @@ var CompatibleChannels = []int{
 	channeltype.DeepSeek,
 	channeltype.TogetherAI,
 	channeltype.Novita,
+	channeltype.SiliconFlow,
 }
 
 func GetCompatibleChannelMeta(channelType int) (string, []string) {
@@ -60,6 +62,8 @@ func GetCompatibleChannelMeta(channelType int) (string, []string) {
 		return "doubao", doubao.ModelList
 	case channeltype.Novita:
 		return "novita", novita.ModelList
+	case channeltype.SiliconFlow:
+		return "siliconflow", siliconflow.ModelList
 	default:
 		return "openai", ModelList
 	}

relay/adaptor/openai/constants.go

@@ -8,6 +8,7 @@ var ModelList = []string{
 	"gpt-4-32k", "gpt-4-32k-0314", "gpt-4-32k-0613",
 	"gpt-4-turbo-preview", "gpt-4-turbo", "gpt-4-turbo-2024-04-09",
 	"gpt-4o", "gpt-4o-2024-05-13",
+	"gpt-4o-mini", "gpt-4o-mini-2024-07-18",
 	"gpt-4-vision-preview",
 	"text-embedding-ada-002", "text-embedding-3-small", "text-embedding-3-large",
 	"text-curie-001", "text-babbage-001", "text-ada-001", "text-davinci-002", "text-davinci-003",

relay/adaptor/openai/token.go

@@ -110,7 +110,7 @@ func CountTokenMessages(messages []model.Message, model string) int {
 						if imageUrl["detail"] != nil {
 							detail = imageUrl["detail"].(string)
 						}
-						imageTokens, err := countImageTokens(url, detail)
+						imageTokens, err := countImageTokens(url, detail, model)
 						if err != nil {
 							logger.SysError("error counting image tokens: " + err.Error())
 						} else {
@@ -134,11 +134,15 @@ const (
 	lowDetailCost         = 85
 	highDetailCostPerTile = 170
 	additionalCost        = 85
+	// gpt-4o-mini cost higher than other model
+	gpt4oMiniLowDetailCost  = 2833
+	gpt4oMiniHighDetailCost = 5667
+	gpt4oMiniAdditionalCost = 2833
 )
 
 // https://platform.openai.com/docs/guides/vision/calculating-costs
 // https://github.com/openai/openai-cookbook/blob/05e3f9be4c7a2ae7ecf029a7c32065b024730ebe/examples/How_to_count_tokens_with_tiktoken.ipynb
-func countImageTokens(url string, detail string) (_ int, err error) {
+func countImageTokens(url string, detail string, model string) (_ int, err error) {
 	var fetchSize = true
 	var width, height int
 	// Reference: https://platform.openai.com/docs/guides/vision/low-or-high-fidelity-image-understanding
@@ -172,6 +176,9 @@ func countImageTokens(url string, detail string) (_ int, err error) {
 	}
 	switch detail {
 	case "low":
+		if strings.HasPrefix(model, "gpt-4o-mini") {
+			return gpt4oMiniLowDetailCost, nil
+		}
 		return lowDetailCost, nil
 	case "high":
 		if fetchSize {
@@ -191,6 +198,9 @@ func countImageTokens(url string, detail string) (_ int, err error) {
 			height = int(float64(height) * ratio)
 		}
 		numSquares := int(math.Ceil(float64(width)/512) * math.Ceil(float64(height)/512))
+		if strings.HasPrefix(model, "gpt-4o-mini") {
+			return numSquares*gpt4oMiniHighDetailCost + gpt4oMiniAdditionalCost, nil
+		}
 		result := numSquares*highDetailCostPerTile + additionalCost
 		return result, nil
 	default:

relay/adaptor/siliconflow/constants.go

@@ -0,0 +1,36 @@
+package siliconflow
+
+// https://docs.siliconflow.cn/docs/getting-started
+
+var ModelList = []string{
+	"deepseek-ai/deepseek-llm-67b-chat",
+	"Qwen/Qwen1.5-14B-Chat",
+	"Qwen/Qwen1.5-7B-Chat",
+	"Qwen/Qwen1.5-110B-Chat",
+	"Qwen/Qwen1.5-32B-Chat",
+	"01-ai/Yi-1.5-6B-Chat",
+	"01-ai/Yi-1.5-9B-Chat-16K",
+	"01-ai/Yi-1.5-34B-Chat-16K",
+	"THUDM/chatglm3-6b",
+	"deepseek-ai/DeepSeek-V2-Chat",
+	"THUDM/glm-4-9b-chat",
+	"Qwen/Qwen2-72B-Instruct",
+	"Qwen/Qwen2-7B-Instruct",
+	"Qwen/Qwen2-57B-A14B-Instruct",
+	"deepseek-ai/DeepSeek-Coder-V2-Instruct",
+	"Qwen/Qwen2-1.5B-Instruct",
+	"internlm/internlm2_5-7b-chat",
+	"BAAI/bge-large-en-v1.5",
+	"BAAI/bge-large-zh-v1.5",
+	"Pro/Qwen/Qwen2-7B-Instruct",
+	"Pro/Qwen/Qwen2-1.5B-Instruct",
+	"Pro/Qwen/Qwen1.5-7B-Chat",
+	"Pro/THUDM/glm-4-9b-chat",
+	"Pro/THUDM/chatglm3-6b",
+	"Pro/01-ai/Yi-1.5-9B-Chat-16K",
+	"Pro/01-ai/Yi-1.5-6B-Chat",
+	"Pro/google/gemma-2-9b-it",
+	"Pro/internlm/internlm2_5-7b-chat",
+	"Pro/meta-llama/Meta-Llama-3-8B-Instruct",
+	"Pro/mistralai/Mistral-7B-Instruct-v0.2",
+}

relay/adaptor/stepfun/constants.go

@@ -1,7 +1,13 @@
 package stepfun
 
 var ModelList = []string{
+	"step-1-8k",
 	"step-1-32k",
+	"step-1-128k",
+	"step-1-256k",
+	"step-1-flash",
+	"step-2-16k",
+	"step-1v-8k",
 	"step-1v-32k",
-	"step-1-200k",
+	"step-1x-medium",
 }

relay/billing/ratio/image.go

@@ -30,6 +30,14 @@ var ImageSizeRatios = map[string]map[string]float64{
 		"720x1280":  1,
 		"1280x720":  1,
 	},
+	"step-1x-medium": {
+		"256x256":   1,
+		"512x512":   1,
+		"768x768":   1,
+		"1024x1024": 1,
+		"1280x800":  1,
+		"800x1280":  1,
+	},
 }
 
 var ImageGenerationAmounts = map[string][2]int{
@@ -39,6 +47,7 @@ var ImageGenerationAmounts = map[string][2]int{
 	"ali-stable-diffusion-v1.5": {1, 4}, // Ali
 	"wanx-v1":                   {1, 4}, // Ali
 	"cogview-3":                 {1, 1},
+	"step-1x-medium":            {1, 1},
 }
 
 var ImagePromptLengthLimitations = map[string]int{
@@ -48,6 +57,7 @@ var ImagePromptLengthLimitations = map[string]int{
 	"ali-stable-diffusion-v1.5": 4000,
 	"wanx-v1":                   4000,
 	"cogview-3":                 833,
+	"step-1x-medium":            4000,
 }
 
 var ImageOriginModelName = map[string]string{

relay/billing/ratio/model.go

@@ -28,15 +28,17 @@ var ModelRatio = map[string]float64{
 	"gpt-4-32k":               30,
 	"gpt-4-32k-0314":          30,
 	"gpt-4-32k-0613":          30,
-	"gpt-4-1106-preview":      5,    // $0.01 / 1K tokens
+	"gpt-4-1106-preview":      5,     // $0.01 / 1K tokens
-	"gpt-4-0125-preview":      5,    // $0.01 / 1K tokens
+	"gpt-4-0125-preview":      5,     // $0.01 / 1K tokens
-	"gpt-4-turbo-preview":     5,    // $0.01 / 1K tokens
+	"gpt-4-turbo-preview":     5,     // $0.01 / 1K tokens
-	"gpt-4-turbo":             5,    // $0.01 / 1K tokens
+	"gpt-4-turbo":             5,     // $0.01 / 1K tokens
-	"gpt-4-turbo-2024-04-09":  5,    // $0.01 / 1K tokens
+	"gpt-4-turbo-2024-04-09":  5,     // $0.01 / 1K tokens
-	"gpt-4o":                  2.5,  // $0.005 / 1K tokens
+	"gpt-4o":                  2.5,   // $0.005 / 1K tokens
-	"gpt-4o-2024-05-13":       2.5,  // $0.005 / 1K tokens
+	"gpt-4o-2024-05-13":       2.5,   // $0.005 / 1K tokens
-	"gpt-4-vision-preview":    5,    // $0.01 / 1K tokens
+	"gpt-4o-mini":             0.075, // $0.00015 / 1K tokens
-	"gpt-3.5-turbo":           0.25, // $0.0005 / 1K tokens
+	"gpt-4o-mini-2024-07-18":  0.075, // $0.00015 / 1K tokens
+	"gpt-4-vision-preview":    5,     // $0.01 / 1K tokens
+	"gpt-3.5-turbo":           0.25,  // $0.0005 / 1K tokens
 	"gpt-3.5-turbo-0301":      0.75,
 	"gpt-3.5-turbo-0613":      0.75,
 	"gpt-3.5-turbo-16k":       1.5, // $0.003 / 1K tokens
@@ -96,12 +98,11 @@ var ModelRatio = map[string]float64{
 	"bge-large-en":       0.002 * RMB,
 	"tao-8k":             0.002 * RMB,
 	// https://ai.google.dev/pricing
-	"PaLM-2":                    1,
+	"gemini-pro":       1, // $0.00025 / 1k characters -> $0.001 / 1k tokens
-	"gemini-pro":                1, // $0.00025 / 1k characters -> $0.001 / 1k tokens
+	"gemini-1.0-pro":   1,
-	"gemini-pro-vision":         1, // $0.00025 / 1k characters -> $0.001 / 1k tokens
+	"gemini-1.5-flash": 1,
-	"gemini-1.0-pro-vision-001": 1,
+	"gemini-1.5-pro":   1,
-	"gemini-1.0-pro-001":        1,
+	"aqa":              1,
-	"gemini-1.5-pro":            1,
 	// https://open.bigmodel.cn/pricing
 	"glm-4":         0.1 * RMB,
 	"glm-4v":        0.1 * RMB,
@@ -156,20 +157,29 @@ var ModelRatio = map[string]float64{
 	"mistral-large-latest":  8.0 / 1000 * USD,
 	"mistral-embed":         0.1 / 1000 * USD,
 	// https://wow.groq.com/#:~:text=inquiries%C2%A0here.-,Model,-Current%20Speed
-	"llama3-70b-8192":    0.59 / 1000 * USD,
+	"gemma-7b-it":                           0.07 / 1000000 * USD,
-	"mixtral-8x7b-32768": 0.27 / 1000 * USD,
+	"mixtral-8x7b-32768":                    0.24 / 1000000 * USD,
-	"llama3-8b-8192":     0.05 / 1000 * USD,
+	"llama3-8b-8192":                        0.05 / 1000000 * USD,
-	"gemma-7b-it":        0.1 / 1000 * USD,
+	"llama3-70b-8192":                       0.59 / 1000000 * USD,
-	"llama2-70b-4096":    0.64 / 1000 * USD,
+	"gemma2-9b-it":                          0.20 / 1000000 * USD,
-	"llama2-7b-2048":     0.1 / 1000 * USD,
+	"llama-3.1-405b-reasoning":              0.89 / 1000000 * USD,
+	"llama-3.1-70b-versatile":               0.59 / 1000000 * USD,
+	"llama-3.1-8b-instant":                  0.05 / 1000000 * USD,
+	"llama3-groq-70b-8192-tool-use-preview": 0.89 / 1000000 * USD,
+	"llama3-groq-8b-8192-tool-use-preview":  0.19 / 1000000 * USD,
 	// https://platform.lingyiwanwu.com/docs#-计费单元
 	"yi-34b-chat-0205": 2.5 / 1000 * RMB,
 	"yi-34b-chat-200k": 12.0 / 1000 * RMB,
 	"yi-vl-plus":       6.0 / 1000 * RMB,
-	// stepfun todo
+	// https://platform.stepfun.com/docs/pricing/details
-	"step-1v-32k": 0.024 * RMB,
+	"step-1-8k":    0.005 / 1000 * RMB,
-	"step-1-32k":  0.024 * RMB,
+	"step-1-32k":   0.015 / 1000 * RMB,
-	"step-1-200k": 0.15 * RMB,
+	"step-1-128k":  0.040 / 1000 * RMB,
+	"step-1-256k":  0.095 / 1000 * RMB,
+	"step-1-flash": 0.001 / 1000 * RMB,
+	"step-2-16k":   0.038 / 1000 * RMB,
+	"step-1v-8k":   0.005 / 1000 * RMB,
+	"step-1v-32k":  0.015 / 1000 * RMB,
 	// aws llama3 https://aws.amazon.com/cn/bedrock/pricing/
 	"llama3-8b-8192(33)":  0.0003 / 0.002,  // $0.0003 / 1K tokens
 	"llama3-70b-8192(33)": 0.00265 / 0.002, // $0.00265 / 1K tokens
@@ -195,8 +205,10 @@ var CompletionRatio = map[string]float64{
 	"llama3-70b-8192(33)": 0.0035 / 0.00265,
 }
 
-var DefaultModelRatio map[string]float64
+var (
-var DefaultCompletionRatio map[string]float64
+	DefaultModelRatio      map[string]float64
+	DefaultCompletionRatio map[string]float64
+)
 
 func init() {
 	DefaultModelRatio = make(map[string]float64)
@@ -308,6 +320,9 @@ func GetCompletionRatio(name string, channelType int) float64 {
 		return 4.0 / 3.0
 	}
 	if strings.HasPrefix(name, "gpt-4") {
+		if strings.HasPrefix(name, "gpt-4o-mini") {
+			return 4
+		}
 		if strings.HasPrefix(name, "gpt-4-turbo") ||
 			strings.HasPrefix(name, "gpt-4o") ||
 			strings.HasSuffix(name, "preview") {

relay/channeltype/define.go

@@ -45,5 +45,6 @@ const (
 	Novita
 	VertextAI
 	Proxy
+	SiliconFlow
 	Dummy
 )

relay/channeltype/url.go

@@ -45,6 +45,7 @@ var ChannelBaseURLs = []string{
 	"https://api.novita.ai/v3/openai",           // 41
 	"",                                          // 42
 	"",                                          // 43
+	"https://api.siliconflow.cn",                 // 44
 }
 
 func init() {

relay/controller/text.go

@@ -10,6 +10,7 @@ import (
 	"github.com/gin-gonic/gin"
 	"github.com/songquanpeng/one-api/common/logger"
 	"github.com/songquanpeng/one-api/relay"
+	"github.com/songquanpeng/one-api/relay/adaptor"
 	"github.com/songquanpeng/one-api/relay/adaptor/openai"
 	"github.com/songquanpeng/one-api/relay/apitype"
 	"github.com/songquanpeng/one-api/relay/billing"
@@ -31,9 +32,8 @@ func RelayTextHelper(c *gin.Context) *model.ErrorWithStatusCode {
 	meta.IsStream = textRequest.Stream
 
 	// map model name
-	var isModelMapped bool
 	meta.OriginModelName = textRequest.Model
-	textRequest.Model, isModelMapped = getMappedModelName(textRequest.Model, meta.ModelMapping)
+	textRequest.Model, _ = getMappedModelName(textRequest.Model, meta.ModelMapping)
 	meta.ActualModelName = textRequest.Model
 	// get model ratio & group ratio
 	modelRatio := billingratio.GetModelRatio(textRequest.Model, meta.ChannelType)
@@ -55,30 +55,9 @@ func RelayTextHelper(c *gin.Context) *model.ErrorWithStatusCode {
 	adaptor.Init(meta)
 
 	// get request body
-	var requestBody io.Reader
+	requestBody, err := getRequestBody(c, meta, textRequest, adaptor)
-	if meta.APIType == apitype.OpenAI {
+	if err != nil {
-		// no need to convert request for openai
+		return openai.ErrorWrapper(err, "convert_request_failed", http.StatusInternalServerError)
-		shouldResetRequestBody := isModelMapped || meta.ChannelType == channeltype.Baichuan // frequency_penalty 0 is not acceptable for baichuan
-		if shouldResetRequestBody {
-			jsonStr, err := json.Marshal(textRequest)
-			if err != nil {
-				return openai.ErrorWrapper(err, "json_marshal_failed", http.StatusInternalServerError)
-			}
-			requestBody = bytes.NewBuffer(jsonStr)
-		} else {
-			requestBody = c.Request.Body
-		}
-	} else {
-		convertedRequest, err := adaptor.ConvertRequest(c, meta.Mode, textRequest)
-		if err != nil {
-			return openai.ErrorWrapper(err, "convert_request_failed", http.StatusInternalServerError)
-		}
-		jsonData, err := json.Marshal(convertedRequest)
-		if err != nil {
-			return openai.ErrorWrapper(err, "json_marshal_failed", http.StatusInternalServerError)
-		}
-		logger.Debugf(ctx, "converted request: \n%s", string(jsonData))
-		requestBody = bytes.NewBuffer(jsonData)
 	}
 
 	// do request
@@ -103,3 +82,26 @@ func RelayTextHelper(c *gin.Context) *model.ErrorWithStatusCode {
 	go postConsumeQuota(ctx, usage, meta, textRequest, ratio, preConsumedQuota, modelRatio, groupRatio)
 	return nil
 }
+
+func getRequestBody(c *gin.Context, meta *meta.Meta, textRequest *model.GeneralOpenAIRequest, adaptor adaptor.Adaptor) (io.Reader, error) {
+	if meta.APIType == apitype.OpenAI && meta.OriginModelName == meta.ActualModelName && meta.ChannelType != channeltype.Baichuan {
+		// no need to convert request for openai
+		return c.Request.Body, nil
+	}
+
+	// get request body
+	var requestBody io.Reader
+	convertedRequest, err := adaptor.ConvertRequest(c, meta.Mode, textRequest)
+	if err != nil {
+		logger.Debugf(c.Request.Context(), "converted request failed: %s\n", err.Error())
+		return nil, err
+	}
+	jsonData, err := json.Marshal(convertedRequest)
+	if err != nil {
+		logger.Debugf(c.Request.Context(), "converted request json_marshal_failed: %s\n", err.Error())
+		return nil, err
+	}
+	logger.Debugf(c.Request.Context(), "converted request: \n%s", string(jsonData))
+	requestBody = bytes.NewBuffer(jsonData)
+	return requestBody, nil
+}

relay/model/general.go

@@ -1,7 +1,15 @@
 package model
 
 type ResponseFormat struct {
-	Type string `json:"type,omitempty"`
+	Type       string      `json:"type,omitempty"`
+	JsonSchema *JSONSchema `json:"json_schema,omitempty"`
+}
+
+type JSONSchema struct {
+	Description string                 `json:"description,omitempty"`
+	Name        string                 `json:"name"`
+	Schema      map[string]interface{} `json:"schema,omitempty"`
+	Strict      *bool                  `json:"strict,omitempty"`
 }
 
 type GeneralOpenAIRequest struct {
@@ -29,6 +37,7 @@ type GeneralOpenAIRequest struct {
 	Dimensions       int             `json:"dimensions,omitempty"`
 	Instruction      string          `json:"instruction,omitempty"`
 	Size             string          `json:"size,omitempty"`
+	NumCtx           int         	 `json:"num_ctx,omitempty"`
 }
 
 func (r GeneralOpenAIRequest) ParseInput() []string {

web/air/src/components/TokensTable.js

@@ -11,12 +11,14 @@ import EditToken from '../pages/Token/EditToken';
 const COPY_OPTIONS = [
   { key: 'next', text: 'ChatGPT Next Web', value: 'next' },
   { key: 'ama', text: 'ChatGPT Web & Midjourney', value: 'ama' },
-  { key: 'opencat', text: 'OpenCat', value: 'opencat' }
+  { key: 'opencat', text: 'OpenCat', value: 'opencat' },
+  { key: 'lobechat', text: 'LobeChat', value: 'lobechat' },
 ];
 
 const OPEN_LINK_OPTIONS = [
   { key: 'ama', text: 'ChatGPT Web & Midjourney', value: 'ama' },
-  { key: 'opencat', text: 'OpenCat', value: 'opencat' }
+  { key: 'opencat', text: 'OpenCat', value: 'opencat' },
+  { key: 'lobechat', text: 'LobeChat', value: 'lobechat' }
 ];
 
 function renderTimestamp(timestamp) {
@@ -60,7 +62,12 @@ const TokensTable = () => {
         onOpenLink('next-mj');
       }
     },
-    { node: 'item', key: 'opencat', name: 'OpenCat', value: 'opencat' }
+    { node: 'item', key: 'opencat', name: 'OpenCat', value: 'opencat' },
+    {
+      node: 'item', key: 'lobechat', name: 'LobeChat', onClick: () => {
+        onOpenLink('lobechat');
+      }
+    }
   ];
 
   const columns = [
@@ -177,6 +184,11 @@ const TokensTable = () => {
                   node: 'item', key: 'opencat', name: 'OpenCat', onClick: () => {
                     onOpenLink('opencat', record.key);
                   }
+                },
+                {
+                  node: 'item', key: 'lobechat', name: 'LobeChat', onClick: () => {
+                    onOpenLink('lobechat');
+                  }
                 }
               ]
             }
@@ -382,6 +394,9 @@ const TokensTable = () => {
       case 'next-mj':
         url = mjLink + `/#/?settings={"key":"sk-${key}","url":"${serverAddress}"}`;
         break;
+      case 'lobechat':
+        url = chatLink + `/?settings={"keyVaults":{"openai":{"apiKey":"sk-${key}","baseURL":"${serverAddress}"/v1"}}}`;
+        break;
       default:
         if (!chatLink) {
           showError('管理员未设置聊天链接');

web/air/src/constants/channel.constants.js

@@ -29,6 +29,7 @@ export const CHANNEL_OPTIONS = [
   { key: 39, text: 'together.ai', value: 39, color: 'blue' },
   { key: 42, text: 'VertexAI', value: 42, color: 'blue' },
   { key: 43, text: 'Proxy', value: 43, color: 'blue' },
+  { key: 44, text: 'SiliconFlow', value: 44, color: 'blue' },
   { key: 8, text: '自定义渠道', value: 8, color: 'pink' },
   { key: 22, text: '知识库：FastGPT', value: 22, color: 'blue' },
   { key: 21, text: '知识库：AI Proxy', value: 21, color: 'purple' },

web/berry/src/constants/ChannelConstants.js

@@ -173,6 +173,12 @@ export const CHANNEL_OPTIONS = {
     value: 43,
     color: 'primary'
   },
+  44: {
+    key: 44,
+    text: 'SiliconFlow',
+    value: 44,
+    color: 'primary'
+  },
   41: {
     key: 41,
     text: 'Novita',

web/berry/src/views/Token/component/TableRow.js

@@ -32,7 +32,8 @@ const COPY_OPTIONS = [
     encode: false
   },
   { key: 'ama', text: 'BotGem', url: 'ama://set-api-key?server={serverAddress}&key=sk-{key}', encode: true },
-  { key: 'opencat', text: 'OpenCat', url: 'opencat://team/join?domain={serverAddress}&token=sk-{key}', encode: true }
+  { key: 'opencat', text: 'OpenCat', url: 'opencat://team/join?domain={serverAddress}&token=sk-{key}', encode: true },
+  { key: 'lobechat', text: 'LobeChat', url: 'https://lobehub.com/?settings={"keyVaults":{"openai":{"apiKey":"user-key","baseURL":"https://your-proxy.com/v1"}}}', encode: true }
 ];
 
 function replacePlaceholders(text, key, serverAddress) {

web/default/src/components/TokensTable.js

@@ -10,12 +10,14 @@ const COPY_OPTIONS = [
   { key: 'next', text: 'ChatGPT Next Web', value: 'next' },
   { key: 'ama', text: 'BotGem', value: 'ama' },
   { key: 'opencat', text: 'OpenCat', value: 'opencat' },
+  { key: 'lobechat', text: 'LobeChat', value: 'lobechat' },
 ];
 
 const OPEN_LINK_OPTIONS = [
   { key: 'next', text: 'ChatGPT Next Web', value: 'next' },
   { key: 'ama', text: 'BotGem', value: 'ama' },
   { key: 'opencat', text: 'OpenCat', value: 'opencat' },
+  { key: 'lobechat', text: 'LobeChat', value: 'lobechat' },
 ];
 
 function renderTimestamp(timestamp) {
@@ -114,6 +116,9 @@ const TokensTable = () => {
       case 'next':
         url = nextUrl;
         break;
+      case 'lobechat':
+        url = nextLink + `/?settings={"keyVaults":{"openai":{"apiKey":"sk-${key}","baseURL":"${serverAddress}"/v1"}}}`;
+        break;
       default:
         url = `sk-${key}`;
     }
@@ -153,7 +158,11 @@ const TokensTable = () => {
       case 'opencat':
         url = `opencat://team/join?domain=${encodedServerAddress}&token=sk-${key}`;
         break;
-        
+
+      case 'lobechat':
+        url = chatLink + `/?settings={"keyVaults":{"openai":{"apiKey":"sk-${key}","baseURL":"${serverAddress}"/v1"}}}`;
+        break;
+
       default:
         url = defaultUrl;
     }

web/default/src/constants/channel.constants.js

@@ -29,6 +29,7 @@ export const CHANNEL_OPTIONS = [
     { key: 39, text: 'together.ai', value: 39, color: 'blue' },
     { key: 42, text: 'VertexAI', value: 42, color: 'blue' },
     { key: 43, text: 'Proxy', value: 43, color: 'blue' },
+    { key: 44, text: 'SiliconFlow', value: 44, color: 'blue' },
     { key: 8, text: '自定义渠道', value: 8, color: 'pink' },
     { key: 22, text: '知识库：FastGPT', value: 22, color: 'blue' },
     { key: 21, text: '知识库：AI Proxy', value: 21, color: 'purple' },