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

* refactor(custom): add all recognized models to maxTokensMap for custom endpoint

* feat(librechat.yaml): log the custom config file on initial load

* fix(OpenAIClient): pass endpointType/endpoint to `getModelMaxTokens` call
This commit is contained in:
Danny Avila 2024-01-27 08:59:04 -05:00 committed by GitHub
parent c470147ea2
commit f7f7f929a0
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
3 changed files with 27 additions and 21 deletions

View file

@ -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,
};
/**