Expand description
Token-greedy helper for C1: prefill the prompt, then decode one token at a time.
Functionsยง
- greedy_
decode - Greedy decode. Returns
prompt || new_tokens(lengthprompt.len() + max_new).
Token-greedy helper for C1: prefill the prompt, then decode one token at a time.
prompt || new_tokens (length prompt.len() + max_new).