rixty
/
SillyTavern
mirror of https://github.com/SillyTavern/SillyTavern.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Groq: Add new models and multimodal captions

This commit is contained in:
Cohee 2024-10-03 08:41:45 +03:00
parent 85b0d135f1
commit 6706cce10d
6 changed files with 47 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

35
public/index.html
View File

@ -2954,16 +2954,31 @@
</div>
<h4 data-i18n="Groq Model">Groq Model</h4>
<select id="model_groq_select">
<option value="llama-3.1-405b-reasoning">llama-3.1-405b-reasoning</option>
<option value="llama-3.1-70b-versatile">llama-3.1-70b-versatile</option>
<option value="llama-3.1-8b-instant">llama-3.1-8b-instant</option>
<option value="llama3-groq-70b-8192-tool-use-preview">llama3-groq-70b-8192-tool-use-preview</option>
<option value="llama3-groq-8b-8192-tool-use-preview">llama3-groq-8b-8192-tool-use-preview</option>
<option value="llama3-8b-8192">llama3-8b-8192</option>
<option value="llama3-70b-8192">llama3-70b-8192</option>
<option value="mixtral-8x7b-32768">mixtral-8x7b-32768</option>
<option value="gemma-7b-it">gemma-7b-it</option>
<option value="gemma2-9b-it">gemma2-9b-it</option>
<optgroup label="Llama 3.2">
<option value="llama-3.2-1b-preview">llama-3.2-1b-preview</option>
<option value="llama-3.2-3b-preview">llama-3.2-3b-preview</option>
<option value="llama-3.2-11b-vision-preview">llama-3.2-11b-vision-preview</option>
<option value="llama-3.2-90b-vision-preview">llama-3.2-90b-vision-preview</option>
</optgroup>
<optgroup label="Llama 3.1">
<option value="llama-3.1-8b-instant">llama-3.1-8b-instant</option>
<option value="llama-3.1-70b-versatile">llama-3.1-70b-versatile</option>
<option value="llama-3.1-405b-reasoning">llama-3.1-405b-reasoning</option>
</optgroup>
<optgroup label="Llama 3">
<option value="llama3-groq-8b-8192-tool-use-preview">llama3-groq-8b-8192-tool-use-preview</option>
<option value="llama3-groq-70b-8192-tool-use-preview">llama3-groq-70b-8192-tool-use-preview</option>
<option value="llama3-8b-8192">llama3-8b-8192</option>
<option value="llama3-70b-8192">llama3-70b-8192</option>
</optgroup>
<optgroup label="Gemma">
<option value="gemma-7b-it">gemma-7b-it</option>
<option value="gemma2-9b-it">gemma2-9b-it</option>
</optgroup>
<optgroup label="Other">
<option value="mixtral-8x7b-32768">mixtral-8x7b-32768</option>
<option value="llava-v1.5-7b-4096-preview">llava-v1.5-7b-4096-preview</option>
</optgroup>
</select>
</div>
<div id="perplexity_form" data-source="perplexity">

1
public/scripts/extensions/caption/index.js
View File

@ -403,6 +403,7 @@ jQuery(async function () {
(extension_settings.caption.source === 'multimodal' && extension_settings.caption.multimodal_api === 'openai' && (secret_state[SECRET_KEYS.OPENAI] || extension_settings.caption.allow_reverse_proxy)) ||
(extension_settings.caption.source === 'multimodal' && extension_settings.caption.multimodal_api === 'openrouter' && secret_state[SECRET_KEYS.OPENROUTER]) ||
(extension_settings.caption.source === 'multimodal' && extension_settings.caption.multimodal_api === 'zerooneai' && secret_state[SECRET_KEYS.ZEROONEAI]) ||
(extension_settings.caption.source === 'multimodal' && extension_settings.caption.multimodal_api === 'groq' && secret_state[SECRET_KEYS.GROQ]) ||
(extension_settings.caption.source === 'multimodal' && extension_settings.caption.multimodal_api === 'mistral' && (secret_state[SECRET_KEYS.MISTRALAI] || extension_settings.caption.allow_reverse_proxy)) ||
(extension_settings.caption.source === 'multimodal' && extension_settings.caption.multimodal_api === 'google' && (secret_state[SECRET_KEYS.MAKERSUITE] || extension_settings.caption.allow_reverse_proxy)) ||
(extension_settings.caption.source === 'multimodal' && extension_settings.caption.multimodal_api === 'anthropic' && (secret_state[SECRET_KEYS.CLAUDE] || extension_settings.caption.allow_reverse_proxy)) ||

4
public/scripts/extensions/caption/settings.html
View File

@ -21,6 +21,7 @@
<option value="anthropic">Anthropic</option>
<option value="custom" data-i18n="Custom (OpenAI-compatible)">Custom (OpenAI-compatible)</option>
<option value="google">Google AI Studio</option>
<option value="groq">Groq</option>
<option value="koboldcpp">KoboldCpp</option>
<option value="llamacpp">llama.cpp</option>
<option value="mistral">MistralAI</option>
@ -60,6 +61,9 @@
<option data-type="google" value="gemini-1.5-pro-exp-0801">gemini-1.5-pro-exp-0801</option>
<option data-type="google" value="gemini-1.5-pro-exp-0827">gemini-1.5-pro-exp-0827</option>
<option data-type="google" value="gemini-pro-vision">gemini-pro-vision</option>
<option data-type="groq" value="llama-3.2-11b-vision-preview">llama-3.2-11b-vision-preview</option>
<option data-type="groq" value="llama-3.2-90b-vision-preview">llama-3.2-90b-vision-preview</option>
<option data-type="groq" value="llava-v1.5-7b-4096-preview">llava-v1.5-7b-4096-preview</option>
<option data-type="openrouter" value="openai/gpt-4-vision-preview">openai/gpt-4-vision-preview</option>
<option data-type="openrouter" value="openai/gpt-4o">openai/gpt-4o</option>
<option data-type="openrouter" value="openai/gpt-4o-2024-05-13">openai/gpt-4o-2024-05-13</option>

6
public/scripts/extensions/shared.js
View File

@ -36,7 +36,7 @@ export async function getMultimodalCaption(base64Img, prompt) {
const isVllm = extension_settings.caption.multimodal_api === 'vllm';
const base64Bytes = base64Img.length * 0.75;
const compressionLimit = 2 * 1024 * 1024;
if ((['google', 'openrouter', 'mistral'].includes(extension_settings.caption.multimodal_api) && base64Bytes > compressionLimit) || isOoba || isKoboldCpp) {
if ((['google', 'openrouter', 'mistral', 'groq'].includes(extension_settings.caption.multimodal_api) && base64Bytes > compressionLimit) || isOoba || isKoboldCpp) {
const maxSide = 1024;
base64Img = await createThumbnail(base64Img, maxSide, maxSide, 'image/jpeg');
}
@ -135,6 +135,10 @@ function throwIfInvalidModel(useReverseProxy) {
throw new Error('01.AI API key is not set.');
}
if (extension_settings.caption.multimodal_api === 'groq' && !secret_state[SECRET_KEYS.GROQ]) {
throw new Error('Groq API key is not set.');
}
if (extension_settings.caption.multimodal_api === 'google' && !secret_state[SECRET_KEYS.MAKERSUITE] && !useReverseProxy) {
throw new Error('Google AI Studio API key is not set.');
}

5
public/scripts/openai.js
View File

@ -4196,7 +4196,10 @@ async function onModelChange() {
if (oai_settings.max_context_unlocked) {
$('#openai_max_context').attr('max', unlocked_max);
}
else if (oai_settings.groq_model.includes('llama-3.1')) {
else if (oai_settings.groq_model.includes('llama-3.2') && oai_settings.groq_model.includes('-preview')) {
$('#openai_max_context').attr('max', max_8k);
}
else if (oai_settings.groq_model.includes('llama-3.2') || oai_settings.groq_model.includes('llama-3.1')) {
$('#openai_max_context').attr('max', max_128k);
}
else if (oai_settings.groq_model.includes('llama3-groq')) {

8
src/endpoints/openai.js
View File

@ -55,6 +55,10 @@ router.post('/caption-image', jsonParser, async (request, response) => {
key = readSecret(request.user.directories, SECRET_KEYS.MISTRALAI);
}
if (request.body.api === 'groq') {
key = readSecret(request.user.directories, SECRET_KEYS.GROQ);
}
if (!key && !request.body.reverse_proxy && ['custom', 'ooba', 'koboldcpp', 'vllm'].includes(request.body.api) === false) {
console.log('No key found for API', request.body.api);
return response.sendStatus(400);
@ -111,6 +115,10 @@ router.post('/caption-image', jsonParser, async (request, response) => {
apiUrl = 'https://api.01.ai/v1/chat/completions';
}
if (request.body.api === 'groq') {
apiUrl = 'https://api.groq.com/openai/v1/chat/completions';
}
if (request.body.api === 'mistral') {
apiUrl = 'https://api.mistral.ai/v1/chat/completions';
}