MCPcopy Create free account

hub / github.com/MoonshotAI/Kimi-Audio / functions

Functions236 in github.com/MoonshotAI/Kimi-Audio

⨍ Functions 236◇ Types & classes 52↳ Endpoints 1
Methodforward
( self, input_ids: torch.LongTensor = None, text_input_ids: torch.LongTensor = None,
finetune_codes/model.py:80
Methodfreeze_encoder
Calling this function will disable the gradient computation for the Whisper encoder so that its parameters will not be updated during
kimia_infer/models/tokenizer/whisper_Lv3/modeling_whisper.py:1485
Methodfrom_pretrained
( cls, model_config, ckpt_path, device, max_prompt_chunk=2, ma
kimia_infer/models/detokenizer/semantic_fm_prefix_streaming.py:318
Methodget_decoder
(self)
kimia_infer/models/tokenizer/whisper_Lv3/modeling_whisper.py:1482
Methodget_decoder
(self)
finetune_codes/modeling_kimia.py:847
Methodget_encoder
(self)
kimia_infer/models/tokenizer/whisper_Lv3/modeling_whisper.py:1479
Methodget_input_embeddings
(self)
kimia_infer/models/tokenizer/whisper_Lv3/modeling_whisper.py:989
Methodget_input_embeddings
(self)
kimia_infer/models/tokenizer/whisper_Lv3/modeling_whisper.py:1185
Methodget_input_embeddings
(self)
kimia_infer/models/tokenizer/whisper_Lv3/modeling_whisper.py:1473
Methodget_input_embeddings
(self)
finetune_codes/modeling_kimia.py:575
Methodget_input_embeddings
(self)
finetune_codes/modeling_kimia.py:832
Methodget_output_embeddings
(self)
finetune_codes/modeling_kimia.py:838
Methodinfer_mel
semantic_tokens: [T_1], torch.LongTensor prompt: [T_2, 80], torch.Tensor, DO NOT normalize it outside prompt_semantic_tokens,
kimia_infer/models/detokenizer/semantic_fm_prefix_streaming.py:157
Functioninit_weights
(m, mean=0.0, std=0.01)
kimia_infer/models/detokenizer/vocoder/utils.py:103
Functionload
()
kimia_infer/models/detokenizer/vocoder/alias_free_activation/cuda/load.py:17
Functionload_bytesio_audio
(content, sr: int = SAMPLE_RATE)
kimia_infer/models/tokenizer/whisper_Lv3/whisper.py:20
Functionload_quantize_encoder
(model_path)
kimia_infer/models/tokenizer/glm4_utils.py:14
Methodmax_positions
Maximum number of supported positions.
kimia_infer/models/detokenizer/flow_matching/model.py:154
Functionmaybe_zero_3
(param)
finetune.py:61
Methodprefill
mel: [T, 80], torch.Tensor semantic_token: [T], torch.LongTensor chunk_size: int, default 150
kimia_infer/models/detokenizer/semantic_fm_prefix_streaming.py:241
Methodremove_weight_norm
(self)
kimia_infer/models/detokenizer/vocoder/bigvgan.py:139
Methodremove_weight_norm
(self)
kimia_infer/models/detokenizer/vocoder/bigvgan.py:230
Methodset_decoder
(self, decoder)
finetune_codes/modeling_kimia.py:844
Methodset_input_embeddings
(self, value: nn.Module)
kimia_infer/models/tokenizer/whisper_Lv3/modeling_whisper.py:992
Methodset_input_embeddings
(self, value)
kimia_infer/models/tokenizer/whisper_Lv3/modeling_whisper.py:1188
Methodset_input_embeddings
(self, value)
kimia_infer/models/tokenizer/whisper_Lv3/modeling_whisper.py:1476
Methodset_input_embeddings
(self, value)
finetune_codes/modeling_kimia.py:578
Methodset_input_embeddings
(self, value)
finetune_codes/modeling_kimia.py:835
Methodset_output_embeddings
(self, new_embeddings)
finetune_codes/modeling_kimia.py:841
Methodset_timesteps
(self, timesteps=15)
kimia_infer/models/detokenizer/flow_matching/scheduler.py:44
Functionshift_tokens_right
Shift input ids one token to the right.
kimia_infer/models/tokenizer/whisper_Lv3/modeling_whisper.py:79
Methodstate_dict
(self)
kimia_infer/models/detokenizer/semantic_fm_prefix_streaming.py:220
Methodstep
(self)
kimia_infer/models/detokenizer/flow_matching/scheduler.py:21
Methodstep
(self, xt, predicted_v)
kimia_infer/models/detokenizer/flow_matching/scheduler.py:47
Methodtext_prepend
(self, index: int, text_token_loss_mask: bool = False)
kimia_infer/utils/data.py:40
Methodtext_pretend
(self, ids: list[int], text_token_loss_mask: bool = False )
kimia_infer/utils/data.py:49
← previous201–236 of 236, ranked by callers