huggingface/candle
1
0
Fork 0
mirror of https://github.com/huggingface/candle.git synced 2025-06-15 10:26:33 +00:00
Code Issues Packages Projects Releases Wiki Activity

Fix token generation in bilingual models (non-English outputs) (#1668)

Browse Source 
Co-authored-by: Guoqing Bao <guoqing.bao@enflame-tech.com>
This commit is contained in:
Guoqing Bao
2024-02-06 19:03:53 +08:00
committed by GitHub
parent b545f54a19
commit 678f64dd27
2 changed files with 2 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
candle-examples/examples/yi/main.rs
View File

@ -104,6 +104,7 @@ impl TextGeneration {
break; break;
} }
if let Some(t) = self.tokenizer.next_token(next_token)? { if let Some(t) = self.tokenizer.next_token(next_token)? {
let t = t.replace("<|im_end|>", "\n");
print!("{t}"); print!("{t}");
std::io::stdout().flush()?; std::io::stdout().flush()?;
} }

2
candle-examples/src/token_output_stream.rs
View File

@ -40,7 +40,7 @@ impl TokenOutputStream {
}; };
self.tokens.push(token); self.tokens.push(token);
let text = self.decode(&self.tokens[self.prev_index..])?; let text = self.decode(&self.tokens[self.prev_index..])?;
if text.len() > prev_text.len() && text.chars().last().unwrap().is_ascii() { if text.len() > prev_text.len() && text.chars().last().unwrap().is_alphabetic() {
let text = text.split_at(prev_text.len()); let text = text.split_at(prev_text.len());
self.prev_index = self.current_index; self.prev_index = self.current_index;
self.current_index = self.tokens.len(); self.current_index = self.tokens.len();