Skip to content

Enhanced ChatGPT Clone: Features Agents, MCP, Skills, DeepSeek, Anthropic, AWS, OpenAI, Responses API, Azure, Groq, o1, GPT-5, Mistral, OpenRouter, Vertex AI, Gemini, Artifacts, AI model switching, message search, Code Interpreter, langchain, DALL-E-3, OpenAPI Actions, Functions, Secure Multi-User Auth, Presets, open-source for self-hosting. Active

📋 feat: Log Custom Config File and Add Known Model Limits to Custom Endpoint (#1657)

Danny Avilaf7f7f929a030
diff --git a/api/app/clients/OpenAIClient.js b/api/app/clients/OpenAIClient.js
index ca0c8d8424..d1b908efbb 100644
--- a/api/app/clients/OpenAIClient.js
+++ b/api/app/clients/OpenAIClient.js
@@ -131,7 +131,8 @@ class OpenAIClient extends BaseClient {
const { isChatGptModel } = this;
this.isUnofficialChatGptModel =
model.startsWith('text-chat') || model.startsWith('text-davinci-002-render');
- this.maxContextTokens = getModelMaxTokens(model) ?? 4095; // 1 less than maximum
+ this.maxContextTokens =
+ getModelMaxTokens(model, this.options.endpointType ?? this.options.endpoint) ?? 4095; // 1 less than maximum
if (this.shouldSummarize) {
this.maxContextTokens = Math.floor(this.maxContextTokens / 2);
diff --git a/api/server/services/Config/loadCustomConfig.js b/api/server/services/Config/loadCustomConfig.js
index c17d3283b4..769e4fb1e7 100644
--- a/api/server/services/Config/loadCustomConfig.js
+++ b/api/server/services/Config/loadCustomConfig.js
@@ -25,7 +25,8 @@ async function loadCustomConfig() {
logger.error(`Invalid custom config file at ${configPath}`, result.error);
return null;
} else {
- logger.info('Loaded custom config file');
+ logger.info('Loaded custom config file:');
+ logger.info(JSON.stringify(customConfig, null, 2));
}
if (customConfig.cache) {
diff --git a/api/utils/tokens.js b/api/utils/tokens.js
index 3c95cd96a2..fb6e363d8b 100644
--- a/api/utils/tokens.js
+++ b/api/utils/tokens.js
@@ -57,28 +57,32 @@ const openAIModels = {
'mistral-': 31990, // -10 from max
};
+const googleModels = {
+ /* Max I/O is combined so we subtract the amount from max response tokens for actual total */
+ gemini: 32750, // -10 from max
+ 'text-bison-32k': 32758, // -10 from max
+ 'chat-bison-32k': 32758, // -10 from max
+ 'code-bison-32k': 32758, // -10 from max
+ 'codechat-bison-32k': 32758,
+ /* Codey, -5 from max: 6144 */
+ 'code-': 6139,
+ 'codechat-': 6139,
+ /* PaLM2, -5 from max: 8192 */
+ 'text-': 8187,
+ 'chat-': 8187,
+};
+
+const anthropicModels = {
+ 'claude-2.1': 200000,
+ 'claude-': 100000,
+};
+
// Order is important here: by model series and context size (gpt-4 then gpt-3, ascending)
const maxTokensMap = {
[EModelEndpoint.openAI]: openAIModels,
- [EModelEndpoint.custom]: openAIModels,
- [EModelEndpoint.google]: {
- /* Max I/O is combined so we subtract the amount from max response tokens for actual total */
- gemini: 32750, // -10 from max
- 'text-bison-32k': 32758, // -10 from max
- 'chat-bison-32k': 32758, // -10 from max
- 'code-bison-32k': 32758, // -10 from max
- 'codechat-bison-32k': 32758,
- /* Codey, -5 from max: 6144 */
- 'code-': 6139,
- 'codechat-': 6139,
- /* PaLM2, -5 from max: 8192 */
- 'text-': 8187,
- 'chat-': 8187,
- },
- [EModelEndpoint.anthropic]: {
- 'claude-2.1': 200000,
- 'claude-': 100000,
- },
+ [EModelEndpoint.custom]: { ...openAIModels, ...googleModels, ...anthropicModels },
+ [EModelEndpoint.google]: googleModels,
+ [EModelEndpoint.anthropic]: anthropicModels,
};
/**