| from __future__ import annotations | |
| __all__ = ['BAD_WORDS_LIST', 'BEAM_WIDTH', 'CUM_LOG_PROBS', 'DRAFT_INPUT_IDS', 'DRAFT_LOGITS', 'EARLY_STOPPING', 'EMBEDDING_BIAS', 'END_ID', 'FREQUENCY_PENALTY', 'INPUT_IDS', 'LENGTH_PENALTY', 'MAX_NEW_TOKENS', 'MIN_LENGTH', 'NO_REPEAT_NGRAM_SIZE', 'OUTPUT_IDS', 'OUTPUT_LOG_PROBS', 'PAD_ID', 'PRESENCE_PENALTY', 'PROMPT_EMBEDDING_TABLE', 'PROMPT_VOCAB_SIZE', 'RANDOM_SEED', 'REPETITION_PENALTY', 'RETURN_CONTEXT_LOGITS', 'RETURN_GENERATION_LOGITS', 'RETURN_LOG_PROBS', 'RUNTIME_TOP_K', 'RUNTIME_TOP_P', 'SEQUENCE_LENGTH', 'STOP_WORDS_LIST', 'TEMPERATURE'] | |
| BAD_WORDS_LIST: str = 'bad_words_list' | |
| BEAM_WIDTH: str = 'beam_width' | |
| CUM_LOG_PROBS: str = 'cum_log_probs' | |
| DRAFT_INPUT_IDS: str = 'draft_input_ids' | |
| DRAFT_LOGITS: str = 'draft_logits' | |
| EARLY_STOPPING: str = 'early_stopping' | |
| EMBEDDING_BIAS: str = 'embedding_bias' | |
| END_ID: str = 'end_id' | |
| FREQUENCY_PENALTY: str = 'frequency_penalty' | |
| INPUT_IDS: str = 'input_ids' | |
| LENGTH_PENALTY: str = 'len_penalty' | |
| MAX_NEW_TOKENS: str = 'request_output_len' | |
| MIN_LENGTH: str = 'min_length' | |
| NO_REPEAT_NGRAM_SIZE: str = 'noRepeatNgramSize' | |
| OUTPUT_IDS: str = 'output_ids' | |
| OUTPUT_LOG_PROBS: str = 'output_log_probs' | |
| PAD_ID: str = 'pad_id' | |
| PRESENCE_PENALTY: str = 'presence_penalty' | |
| PROMPT_EMBEDDING_TABLE: str = 'prompt_embedding_table' | |
| PROMPT_VOCAB_SIZE: str = 'prompt_vocab_size' | |
| RANDOM_SEED: str = 'random_seed' | |
| REPETITION_PENALTY: str = 'repetition_penalty' | |
| RETURN_CONTEXT_LOGITS: str = 'return_context_logits' | |
| RETURN_GENERATION_LOGITS: str = 'return_generation_logits' | |
| RETURN_LOG_PROBS: str = 'return_log_probs' | |
| RUNTIME_TOP_K: str = 'runtime_top_k' | |
| RUNTIME_TOP_P: str = 'runtime_top_p' | |
| SEQUENCE_LENGTH: str = 'sequence_length' | |
| STOP_WORDS_LIST: str = 'stop_words_list' | |
| TEMPERATURE: str = 'temperature' | |