MCPcopy Create free account

hub / github.com/OpenMOSS/MOSS-TTS / functions

Functions1,298 in github.com/OpenMOSS/MOSS-TTS

⨍ Functions 1,298◇ Types & classes 178↳ Endpoints 16
↓ 1 callersMethod_validate_request
(self, request: StreamingRequest)
moss_tts_realtime/app.py:462
↓ 1 callersFunction_warmup_button_update
(snapshot: WarmupSnapshot)
moss_tts_realtime/app.py:843
↓ 1 callersFunction_warmup_status_update
(snapshot: WarmupSnapshot)
moss_tts_realtime/app.py:862
↓ 1 callersMethod_warmup_step_detail
(step_idx: int, total_steps: int)
moss_tts_realtime/app.py:720
↓ 1 callersFunction_warmup_timer_update
(snapshot: WarmupSnapshot)
moss_tts_realtime/app.py:866
↓ 1 callersMethodacquire
(self)
moss_soundeffect_v2/diffsynth/trainers/cache_shards.py:39
↓ 1 callersMethodadd_chunk
(self, chunk: np.ndarray)
moss_tts_realtime/app.py:185
↓ 1 callersMethodadd_chunk
(self, chunk: np.ndarray)
moss_tts_realtime/fast_api.py:96
↓ 1 callersMethodappend_audio
Append a single (N_VQ,) frame to the audio history buffer.
moss_tts_delay/llama_cpp/delay_state.py:74
↓ 1 callersFunctionapply_activation_checkpoint_wrapper_to_model
( model: torch.nn.Module, *, scope: str, use_reentrant: bool, )
moss_tts_local_v1.5/finetuning/sft.py:534
↓ 1 callersFunctionapply_de_delay_pattern
Remove delay pattern from generated codes.
moss_tts_delay/llama_cpp/delay_state.py:274
↓ 1 callersMethodapply_de_delay_pattern
(delay_codes: torch.Tensor)
moss_tts_delay/processing_moss_tts.py:548
↓ 1 callersFunctionapply_delay_pattern
Apply delay pattern to audio codes. Args: codes: (T, n_vq) — original audio codes Returns: delayed: (T + n_vq - 1, n_vq) — de
moss_tts_delay/llama_cpp/delay_state.py:259
↓ 1 callersMethodapply_delay_pattern
(codes: torch.Tensor, pad_code: int)
moss_tts_delay/processing_moss_tts.py:536
↓ 1 callersMethodapply_delay_pattern
( tokens: torch.Tensor, ch0_pad_id: int, pad_id: int )
moss_tts_delay/processing_moss_tts_delay_with_codec.py:217
↓ 1 callersFunctionapply_lora_to_language_backbone
( model: torch.nn.Module, args: argparse.Namespace, )
community/norwegian-lora/train_lora.py:138
↓ 1 callersFunctionapply_repetition_penalty
Apply repetition penalty independently per VQ head.
moss_tts_delay/llama_cpp/sampling.py:47
↓ 1 callersMethodapply_repetition_penalty
( self, scores: torch.Tensor, history_tokens: torch.Tensor, penalty: float =
moss_tts_realtime/inferencer.py:431
↓ 1 callersMethodapply_repetition_penalty
( self, scores: torch.Tensor, history_tokens: torch.Tensor, penalty: float = 1
moss_tts_realtime/mossttsrealtime/streaming_mossttsrealtime.py:413
↓ 1 callersFunctionapply_repetition_penalty_delay_pattern
logits: [B, H, V] or [N, V] prev_tokens: [B, T, H] or [N, T] or [B, H] Apply the repetition penalty independently for each H (VQ head).
moss_tts_local/inference_utils.py:51
↓ 1 callersFunctionapply_repetition_penalty_delay_pattern
logits: [B, H, V] or [N, V] prev_tokens: [B, T, H] or [N, T] or [B, H] Apply the repetition penalty independently for each H (VQ head).
moss_tts_delay/inference_utils.py:62
↓ 1 callersFunctionapply_rotary_pos_emb
(q, k, cos, sin, unsqueeze_dim=1)
moss_tts_realtime/mossttsrealtime/modeling_mossttsrealtime_local.py:79
↓ 1 callersFunctionapply_rotary_pos_emb
( query: torch.Tensor, key: torch.Tensor, cos: torch.Tensor, sin: torch.Tensor, )
moss_tts_local_v1.5/qwen3_decoder.py:86
↓ 1 callersFunctionapply_top_k
(logits, top_k)
moss_tts_local/inference_utils.py:8
↓ 1 callersFunctionapply_top_k
(logits, top_k)
moss_tts_delay/inference_utils.py:19
↓ 1 callersMethodapply_top_k
(self, logits, top_k, filter_value=float('-inf'), min_tokens_to_keep: int = 1)
moss_tts_realtime/inferencer.py:450
↓ 1 callersMethodapply_top_k
(self, logits, top_k, filter_value=float("-inf"), min_tokens_to_keep: int = 1)
moss_tts_realtime/mossttsrealtime/streaming_mossttsrealtime.py:451
↓ 1 callersMethodapply_top_p
(self, logits, top_p, filter_value=float('-inf'), min_tokens_to_keep: int = 1)
moss_tts_realtime/inferencer.py:459
↓ 1 callersMethodapply_top_p
(self, logits, top_p, filter_value=float("-inf"), min_tokens_to_keep: int = 1)
moss_tts_realtime/mossttsrealtime/streaming_mossttsrealtime.py:460
↓ 1 callersFunctionapply_top_p_optimized
(logits, top_p)
moss_tts_local/inference_utils.py:33
↓ 1 callersFunctionapply_top_p_optimized
(logits, top_p)
moss_tts_delay/inference_utils.py:44
↓ 1 callersFunctionattach_reference_audio_codes
( records: List[Dict[str, Any]], path_to_codes: Dict[str, List[List[int]]], )
moss_tts_local/finetuning/prepare_data.py:129
↓ 1 callersFunctionattach_reference_audio_codes
( records: List[Dict[str, Any]], path_to_codes: Dict[str, List[List[int]]], )
moss_tts_delay/finetuning/prepare_data.py:129
↓ 1 callersFunctionattach_reference_audio_codes
( records: List[Dict[str, Any]], path_to_codes: Dict[str, List[List[int]]], )
moss_tts_local_v1.5/finetuning/prepare_data.py:222
↓ 1 callersMethodaudio_all
Compute logits for all 32 audio heads (skips text head).
moss_tts_delay/llama_cpp/lm_heads.py:74
↓ 1 callersMethodaudio_history
Return (T, N_VQ) int64 view of audio history, or None if empty.
moss_tts_delay/llama_cpp/delay_state.py:68
↓ 1 callersFunctionbatch_encode
Encode a list of audio file paths to RVQ codes in batches.
moss_tts_realtime/finetuning/prepare_data.py:169
↓ 1 callersMethodblend_with_mask
(self, base, addition, mask)
moss_soundeffect_v2/diffsynth/utils/__init__.py:143
↓ 1 callersMethodbuffered_seconds
(self)
moss_tts_realtime/app.py:192
↓ 1 callersMethodbuffered_seconds
(self)
moss_tts_realtime/fast_api.py:103
↓ 1 callersMethodbuild_1d_mask
(self, length, left_bound, right_bound, border_width)
moss_soundeffect_v2/diffsynth/pipelines/wan_audio.py:635
↓ 1 callersFunctionbuild_arg_parser
()
moss_tts_realtime/example_multiturn_stream_to_tts.py:210
↓ 1 callersFunctionbuild_conversation
( text: str, reference_audio: str | None, mode_with_reference: str, expected_tokens: int | Non
clis/moss_tts_app.py:248
↓ 1 callersFunctionbuild_conversation
(text: str, instruction: str, processor)
clis/moss_voice_generator_app.py:135
↓ 1 callersFunctionbuild_conversation
( dialogue_text: str, reference_audio_codes: list[Optional[torch.Tensor]], prompt_audio: torch.Ten
clis/moss_ttsd_app.py:371
↓ 1 callersFunctionbuild_demo
(args: argparse.Namespace)
clis/moss_tts_app.py:414
↓ 1 callersFunctionbuild_demo
(args: argparse.Namespace)
clis/moss_voice_generator_app.py:210
↓ 1 callersFunctionbuild_demo
(args: argparse.Namespace)
clis/moss_ttsd_app.py:530
↓ 1 callersFunctionbuild_demo
(args: argparse.Namespace)
clis/moss_sound_effect_app.py:85
↓ 1 callersFunctionbuild_example_rows
()
clis/moss_tts_app.py:90
↓ 1 callersFunctionbuild_example_rows
()
clis/moss_voice_generator_app.py:36
↓ 1 callersFunctionbuild_example_rows
()
clis/moss_tts_local_v1.5_app.py:114
↓ 1 callersFunctionbuild_generation_prompt
Build the full multi-channel input_ids for generation. Returns: input_ids: (S, 33) int64
moss_tts_delay/llama_cpp/processor.py:70
↓ 1 callersFunctionbuild_lora_target_modules
(lm_heads_mode: str)
community/norwegian-lora/train_lora.py:127
↓ 1 callersFunctionbuild_lr_scheduler
( optimizer: torch.optim.Optimizer, warmup_steps: int, total_steps: int, )
community/norwegian-lora/train_lora.py:278
↓ 1 callersMethodbuild_mask
(self, data, is_bound, border_width)
moss_soundeffect_v2/diffsynth/pipelines/wan_audio.py:647
↓ 1 callersFunctionbuild_merged_config
(main_config: JSONDict, codec_config: JSONDict)
scripts/fuse_moss_tts_delay_with_codec.py:632
↓ 1 callersFunctionbuild_parser
()
moss_soundeffect_v2/finetuning/finetuning.py:187
↓ 1 callersFunctionbuild_prefixed_codec_keys
( codec_weight_map: dict[str, str], main_weight_map: dict[str, str], prefix: str, )
scripts/fuse_moss_tts_delay_with_codec.py:222
↓ 1 callersFunctionbuild_processor
Load tokenizer from model_path and wrap in MossTTSRealtimeProcessor.
moss_tts_realtime/finetuning/sft.py:233
↓ 1 callersFunctionbuild_processor
( model_path: str, codec_path: str, need_audio_tokenizer: bool, audio_tokenizer_device: Option
moss_tts_local/finetuning/sft.py:259
↓ 1 callersFunctionbuild_processor
( model_path: str, codec_path: str, need_audio_tokenizer: bool, audio_tokenizer_device: Option
moss_tts_delay/finetuning/sft.py:260
↓ 1 callersFunctionbuild_processor
( args: argparse.Namespace, *, need_audio_tokenizer: bool, default_audio_tokenizer_device: str
moss_tts_local_v1.5/finetuning/sft.py:400
↓ 1 callersFunctionbuild_processor
(args: argparse.Namespace, device: str)
moss_tts_local_v1.5/finetuning/prepare_data.py:116
↓ 1 callersFunctionbuild_qwen3_config
Build a Qwen3ForCausalLM config.json from the MOSS-TTS config.
moss_tts_delay/llama_cpp/conversion/extract_weights.py:73
↓ 1 callersFunctionbuild_qwen3_config
(moss_config: dict)
scripts/extract_weights_llama_cpp.py:65
↓ 1 callersMethodcalculate_shift
( self, image_seq_len, base_seq_len: int = 256, max_seq_len: int = 8192,
moss_soundeffect_v2/diffsynth/schedulers/flow_match.py:132
↓ 1 callersFunctioncanonicalize
(text, keep_punctuation_exact_string=None)
moss_soundeffect_v2/diffsynth/prompters/wan_prompter.py:22
↓ 1 callersMethodcheck
(self, dit: WanAudioModel, x, t_mod)
moss_soundeffect_v2/diffsynth/pipelines/wan_audio.py:599
↓ 1 callersMethodcheck_resize_num_channels_num_samples
(self, num_channels, num_samples)
moss_soundeffect_v2/diffsynth/pipelines/wan_audio.py:181
↓ 1 callersFunctioncleanup_partial_output_dir
(output_dir: Path)
scripts/fuse_moss_tts_delay_with_codec.py:627
↓ 1 callersMethodclear_kv
Clear the KV cache (for starting a new sequence).
moss_tts_delay/llama_cpp/backbone.py:226
↓ 1 callersMethodclose
(self, job_id: str)
clis/moss_tts_local_v1.5_app.py:366
↓ 1 callersMethodclose
(self)
moss_tts_delay/llama_cpp/pipeline.py:816
↓ 1 callersFunctionclose_session
(session_id: str)
moss_tts_realtime/tts_client.py:137
↓ 1 callersFunctioncollect_paths
(records: List[Dict[str, Any]], field_name: str)
moss_tts_realtime/finetuning/prepare_data.py:187
↓ 1 callersFunctioncollect_paths
(records: List[Dict[str, Any]], field_name: str)
moss_tts_local/finetuning/prepare_data.py:113
↓ 1 callersFunctioncollect_paths
(records: List[Dict[str, Any]], field_name: str)
moss_tts_delay/finetuning/prepare_data.py:113
↓ 1 callersFunctioncollect_paths
(records: List[Dict[str, Any]], field_name: str)
moss_tts_local_v1.5/finetuning/prepare_data.py:202
↓ 1 callersFunctioncollect_reference_paths
(records: List[Dict[str, Any]])
moss_tts_local/finetuning/prepare_data.py:122
↓ 1 callersFunctioncollect_reference_paths
(records: List[Dict[str, Any]])
moss_tts_delay/finetuning/prepare_data.py:122
↓ 1 callersFunctioncollect_reference_paths
(records: List[Dict[str, Any]])
moss_tts_local_v1.5/finetuning/prepare_data.py:215
↓ 1 callersMethodcompress
Processes an audio signal from a file or AudioSignal object into discrete codes. This function processes the signal in short windows,
moss_soundeffect_v2/diffsynth/models/dac_vae.py:129
↓ 1 callersFunctioncompute_supervised_loss
( model, *, input_ids: torch.LongTensor, attention_mask: torch.BoolTensor, labels: torch.L
moss_tts_local_v1.5/finetuning/sft.py:709
↓ 1 callersFunctioncompute_supervised_loss_from_hidden
( base_model, *, global_hidden_states: torch.FloatTensor, labels: torch.LongTensor, channe
moss_tts_local_v1.5/finetuning/sft.py:613
↓ 1 callersFunctionconfigure_audio_tokenizer
( audio_tokenizer, *, codec_compute_dtype: str, codec_attention_implementation: str, )
moss_tts_local_v1.5/finetuning/sft.py:483
↓ 1 callersFunctionconfigure_audio_tokenizer
( audio_tokenizer, *, codec_compute_dtype: str, codec_attention_implementation: str, )
moss_tts_local_v1.5/finetuning/prepare_data.py:175
↓ 1 callersFunctionconfigure_gradient_checkpointing
( model, *, scope: str, use_reentrant: bool, impl: str, )
moss_tts_local_v1.5/finetuning/sft.py:573
↓ 1 callersFunctionconfigure_torch_backends
()
moss_tts_realtime/finetuning/sft.py:133
↓ 1 callersFunctionconfigure_torch_backends
()
moss_tts_local/finetuning/sft.py:145
↓ 1 callersFunctionconfigure_torch_backends
()
moss_tts_delay/finetuning/sft.py:146
↓ 1 callersFunctionconfigure_torch_backends
()
moss_tts_local_v1.5/finetuning/sft.py:228
↓ 1 callersFunctionconvert_dit_keys
Custom WanAudioModel state_dict → diffusers state_dict.
moss_soundeffect_v2/hf_export.py:61
↓ 1 callersFunctionconvert_state_dict_keys_to_single_str
(state_dict, with_shape=True)
moss_soundeffect_v2/diffsynth/models/utils.py:148
↓ 1 callersFunctioncopy_inference_assets
Copy tokenizer files and write processor_config.json.
moss_tts_realtime/finetuning/sft.py:304
↓ 1 callersFunctioncopy_inference_assets
(model_path: str, codec_path: str, output_dir: Path)
moss_tts_local/finetuning/sft.py:344
↓ 1 callersFunctioncopy_inference_assets
(model_path: str, codec_path: str, output_dir: Path)
moss_tts_delay/finetuning/sft.py:345
↓ 1 callersFunctioncopy_inference_assets
(model_path: str, codec_path: str, output_dir: Path)
moss_tts_local_v1.5/finetuning/sft.py:758
↓ 1 callersFunctioncopy_processor_source
(output_dir: Path)
scripts/fuse_moss_tts_delay_with_codec.py:544
↓ 1 callersFunctioncopy_support_files
Copy the entire mossttsrealtime Python package into output_dir/mossttsrealtime/. This makes the checkpoint self-contained: users can load it by p
moss_tts_realtime/finetuning/sft.py:272
← previousnext →501–600 of 1,298, ranked by callers