vllm_omni.model_executor.models.ming_tts.constants ¶

AGGREGATOR_HIDDEN_SIZE = 1024

AUDIO_DUMMY_TOKEN_ID = 151705

AUDIO_END_TOKEN_ID = 151707

AUDIO_EOS_TOKEN_ID = 151704

AUDIO_FRAME_HOP = 882

AUDIO_START_TOKEN_ID = 151706

DEFAULT_CFG = 2.0

DEFAULT_SIGMA = 0.25

DEFAULT_TEMPERATURE = 0.0

HISTORY_PATCH_SIZE = 32

KEY_CFG = 'ming_cfg'

KEY_CHUNK_ID = 'ming_chunk_id'

KEY_DECODE_STEP = 'ming_decode_step'

KEY_LAST_STOP_PROB = 'ming_last_stop_prob'

KEY_LATENT_HISTORY = 'ming_latent_history'

KEY_MAX_DECODE_STEPS = 'ming_max_decode_steps'

KEY_MIN_DECODE_STEPS = 'ming_min_decode_steps'

KEY_NEXT_EMBEDS = 'ming_next_embeds'

KEY_PROMPT_LATENTS = 'ming_prompt_latents'

KEY_PROMPT_LATENT_TAIL = 'ming_prompt_latent_tail'

KEY_REQUEST_ID = 'ming_request_id'

KEY_SIGMA = 'ming_sigma'

KEY_SPEAKER_EMBEDDING = 'ming_speaker_embedding'

KEY_TEMPERATURE = 'ming_temperature'

KEY_TEXT_MODE = 'ming_text_mode'

LATENT_CHUNK_SIZE = 25

LATENT_DIM = 64

LATENT_LEFT_CONTEXT = 0

LLM_HIDDEN_SIZE = 896

LLM_VOCAB_SIZE = 151936

MAX_DECODE_STEPS = 200

PATCH_SIZE = 4

SAMPLE_RATE = 44100

STOP_HEAD_MIN_STEPS = 3

STOP_HEAD_THRESHOLD = 0.5

TEXT_EOS_TOKEN_ID = 151669

VAE_PATCH_SIZE = 4

VISION_START_TOKEN_ID = 151652