prompt_tokens
prompt_tokens
¶
Shared prompt and sequence token-boundary assembly.
Classes:
| Name | Description |
|---|---|
EncodeOnlyTokenizer |
Tokenizer interface required for prompt construction. |
Functions:
| Name | Description |
|---|---|
encode_prompt_token_ids |
Encode a prompt with its configured BOS and EOS boundaries. |
wrap_sequence_token_ids |
Return sequence IDs with a leading BOS and an optional trailing EOS. |
EncodeOnlyTokenizer
¶
encode_prompt_token_ids(prompt, *, tokenizer, prompt_config)
¶
Encode a prompt with its configured BOS and EOS boundaries.
Source code in src/nemo_safe_synthesizer/data_processing/prompt_tokens.py
wrap_sequence_token_ids(token_ids, *, prompt_config, include_eos)
¶
Return sequence IDs with a leading BOS and an optional trailing EOS.