shaw/SillyTavern
1
0
Fork 0
Code Issues Pull Requests Actions 7 Packages Projects Releases Wiki Activity

Enhance text generation API to support model parameter for llama.cpp router mode. Update getStatusTextgen to include model in request body and improve context size handling. (#4914)

Browse Source
This commit is contained in:
my-alt
2025-12-24 22:30:45 +08:00
committed by GitHub
parent 432639e341
commit 29456b9bd2
2 changed files with 20 additions and 9 deletions
Download Patch File Download Diff File Expand all files Collapse all files
public/scripts/textgen-settings.js
+14 -7
View File
@@ -730,12 +730,17 @@ async function getStatusTextgen() {
const supportsChatTemplate = [textgen_types.KOBOLDCPP, textgen_types.LLAMACPP].includes(textgenerationwebui_settings.type);
if (supportsChatTemplate && (wantsInstructDerivation || wantsContextDerivation || wantsContextSize)) {
const model = textgenerationwebui_settings.type === textgen_types.LLAMACPP
? textgenerationwebui_settings.llamacpp_model
: undefined;
const response = await fetch('/api/backends/text-completions/props', {
method: 'POST',
headers: getRequestHeaders(),
body: JSON.stringify({
api_server: endpoint,
api_type: textgenerationwebui_settings.type,
model: model,
}),
});
@@ -747,13 +752,15 @@ async function getStatusTextgen() {
if (wantsContextSize && 'default_generation_settings' in data) {
const backend_max_context = data['default_generation_settings']['n_ctx'];
const old_value = max_context;
if (max_context !== backend_max_context) {
setGenerationParamsFromPreset({ max_length: backend_max_context });
}
if (old_value !== max_context) {
console.log(`Auto-switched max context from ${old_value} to ${max_context}`);
toastr.info(`${old_value} ⇒ ${max_context}`, 'Context Size Changed');
if (backend_max_context && typeof backend_max_context === 'number') {
const old_value = max_context;
if (max_context !== backend_max_context) {
setGenerationParamsFromPreset({ max_length: backend_max_context });
}
if (old_value !== max_context) {
console.log(`Auto-switched max context from ${old_value} to ${max_context}`);
toastr.info(`${old_value} ⇒ ${max_context}`, 'Context Size Changed');
}
}
}
console.log(`We have chat template ${chat_template.split('\n')[0]}...`);