huggingface/text-generation-inference
5
0
Fork 0
mirror of https://github.com/huggingface/text-generation-inference.git synced 2025-06-19 15:52:08 +00:00
Code Issues Packages Projects Releases Wiki Activity

Change const back to u32

Browse Source
This commit is contained in:
Nick Hill 2022-12-31 09:00:50 -08:00
parent b2acd1b15e
commit 894957daff
1 changed files with 3 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
router/src/validation.rs
View File

@ -6,7 +6,7 @@ use thiserror::Error;
use tokenizers::tokenizer::Tokenizer;
use tokio::sync::{mpsc, oneshot};
const MAX_MAX_NEW_TOKENS: usize = 512;
const MAX_MAX_NEW_TOKENS: u32 = 512;
const MAX_STOP_SEQUENCES: usize = 4;
/// Validation
@ -112,7 +112,7 @@ fn validate(
if request.parameters.top_k < 0 {
return Err(ValidationError::TopK);
}
if request.parameters.max_new_tokens as usize > MAX_MAX_NEW_TOKENS {
if request.parameters.max_new_tokens > MAX_MAX_NEW_TOKENS {
return Err(ValidationError::MaxNewTokens(MAX_MAX_NEW_TOKENS));
}
if request.parameters.stop.len() > MAX_STOP_SEQUENCES {
@ -153,7 +153,7 @@ pub enum ValidationError {
#[error("top_k must be strictly positive")]
TopK,
#[error("max_new_tokens must be <= {0}")]
MaxNewTokens(usize),
MaxNewTokens(u32),
#[error("inputs must have less than {1} tokens. Given: {0}")]
InputLength(usize, usize),
#[error("stop supports up to {0} stop sequences. Given: {1}")]