Increase timeouts of OAI out of quota requests
This commit is contained in:
parent
b3ced2c4c5
commit
740f6548a2
|
@ -110,7 +110,7 @@ const max_4k = 4095;
|
||||||
const max_8k = 8191;
|
const max_8k = 8191;
|
||||||
const max_16k = 16383;
|
const max_16k = 16383;
|
||||||
const max_32k = 32767;
|
const max_32k = 32767;
|
||||||
const max_128k = 128 * 1024 - 1;
|
const max_128k = 128 * 1000;
|
||||||
const scale_max = 8191;
|
const scale_max = 8191;
|
||||||
const claude_max = 9000; // We have a proper tokenizer, so theoretically could be larger (up to 9k)
|
const claude_max = 9000; // We have a proper tokenizer, so theoretically could be larger (up to 9k)
|
||||||
const palm2_max = 7500; // The real context window is 8192, spare some for padding due to using turbo tokenizer
|
const palm2_max = 7500; // The real context window is 8192, spare some for padding due to using turbo tokenizer
|
||||||
|
|
|
@ -3222,7 +3222,8 @@ app.post("/generate_openai", jsonParser, function (request, response_generate_op
|
||||||
} else if (fetchResponse.status === 429 && retries > 0) {
|
} else if (fetchResponse.status === 429 && retries > 0) {
|
||||||
console.log(`Out of quota, retrying in ${Math.round(timeout / 1000)}s`);
|
console.log(`Out of quota, retrying in ${Math.round(timeout / 1000)}s`);
|
||||||
setTimeout(() => {
|
setTimeout(() => {
|
||||||
makeRequest(config, response_generate_openai, request, retries - 1);
|
timeout *= 2;
|
||||||
|
makeRequest(config, response_generate_openai, request, retries - 1, timeout);
|
||||||
}, timeout);
|
}, timeout);
|
||||||
} else {
|
} else {
|
||||||
await handleErrorResponse(fetchResponse);
|
await handleErrorResponse(fetchResponse);
|
||||||
|
|
Loading…
Reference in New Issue