model inference settings
This commit is contained in:
@@ -126,7 +126,7 @@ async function getRelevantEntities(userMessage) {
|
||||
}
|
||||
|
||||
async function chat(externalId, userMessage, options = {}) {
|
||||
const { recentEpisodeLimit, semanticLimit, scoreThreshold, temperature} =
|
||||
const { recentEpisodeLimit, semanticLimit, scoreThreshold, temperature, repeatPenalty, topP, topK} =
|
||||
appSettings.load();
|
||||
// 1. Resolve or create session
|
||||
let session = await memory.getSessionByExternalId(externalId);
|
||||
@@ -187,7 +187,7 @@ async function chat(externalId, userMessage, options = {}) {
|
||||
);
|
||||
|
||||
// 5. Run inference
|
||||
const result = await inference.complete(prompt, {...options, temperature});
|
||||
const result = await inference.complete(prompt, {...options, temperature, repeatPenalty, topP, topK});
|
||||
|
||||
// 6. Write episode back to memory
|
||||
memory
|
||||
@@ -217,7 +217,7 @@ async function chat(externalId, userMessage, options = {}) {
|
||||
|
||||
async function chatStream(externalId, userMessage, onChunk, options = {}) {
|
||||
try {
|
||||
const { recentEpisodeLimit, semanticLimit, scoreThreshold, temperature } = appSettings.load();
|
||||
const { recentEpisodeLimit, semanticLimit, scoreThreshold, temperature, repeatPenalty, topP, topK } = appSettings.load();
|
||||
let session = await memory.getSessionByExternalId(externalId);
|
||||
if (!session) session = await memory.createSession(externalId);
|
||||
|
||||
@@ -270,7 +270,7 @@ async function chatStream(externalId, userMessage, onChunk, options = {}) {
|
||||
entities,
|
||||
userMessage,
|
||||
);
|
||||
const res = await inference.completeStream(prompt, {...options, temperature});
|
||||
const res = await inference.completeStream(prompt, {...options, temperature, repeatPenalty, topP, topK});
|
||||
|
||||
let fullText = "";
|
||||
let model = "";
|
||||
|
||||
Reference in New Issue
Block a user