rixty/SillyTavern
1
0
Fork 0
mirror of https://github.com/SillyTavern/SillyTavern.git synced 2025-06-05 21:59:27 +02:00
Code Issues Packages Projects Releases Wiki Activity

Fix Gemini captioning

Browse Source
This commit is contained in:
Cohee
2025-04-06 13:31:16 +03:00
parent 18cc17142f
commit df7c2226f6
3 changed files with 3 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
public/scripts/extensions/caption/settings.html
View File

@ -65,6 +65,7 @@
<option data-type="anthropic" value="claude-3-opus-20240229">claude-3-opus-20240229</option>
<option data-type="anthropic" value="claude-3-sonnet-20240229">claude-3-sonnet-20240229</option>
<option data-type="anthropic" value="claude-3-haiku-20240307">claude-3-haiku-20240307</option>
<option data-type="google" value="gemini-2.5-pro-preview-03-25">gemini-2.5-pro-preview-03-25</option>
<option data-type="google" value="gemini-2.5-pro-exp-03-25">gemini-2.5-pro-exp-03-25</option>
<option data-type="google" value="gemini-2.0-pro-exp">gemini-2.0-pro-exp</option>
<option data-type="google" value="gemini-2.0-pro-exp-02-05">gemini-2.0-pro-exp-02-05</option>

3
src/endpoints/google.js
View File

@ -16,7 +16,7 @@ router.post('/caption-image', async (request, response) => {
const base64Data = request.body.image.split(',')[1];
const apiKey = request.body.reverse_proxy ? request.body.proxy_password : readSecret(request.user.directories, SECRET_KEYS.MAKERSUITE);
const apiUrl = new URL(request.body.reverse_proxy || API_MAKERSUITE);
const model = request.body.model || 'gemini-pro-vision';
const model = request.body.model || 'gemini-2.0-flash';
const url = `${apiUrl.origin}/v1beta/models/${model}:generateContent?key=${apiKey}`;
const body = {
contents: [{
@ -30,7 +30,6 @@ router.post('/caption-image', async (request, response) => {
}],
}],
safetySettings: GEMINI_SAFETY,
generationConfig: { maxOutputTokens: 1000 },
};
console.debug('Multimodal captioning request', model, body);

1
src/prompt-converters.js
View File

@ -360,6 +360,7 @@ export function convertCohereMessages(messages, names) {
*/
export function convertGooglePrompt(messages, model, useSysPrompt, names) {
const visionSupportedModels = [
'gemini-2.5-pro-preview-03-25',
'gemini-2.5-pro-exp-03-25',
'gemini-2.0-pro-exp',
'gemini-2.0-pro-exp-02-05',