Download tiny/conversion_manifest.json from AEmotionStudio/ming-tts-models: direct link, hf CLI and curl.
- Browser
- Download file 8.48 kB
-
https://huggingface.co/AEmotionStudio/ming-tts-models/resolve/main/tiny/conversion_manifest.json
- Command line
-
hf download hf://AEmotionStudio/ming-tts-models/tiny/conversion_manifest.json
-
curl -L -o conversion_manifest.json https://huggingface.co/AEmotionStudio/ming-tts-models/resolve/main/tiny/conversion_manifest.json
8.48 kB
| { | |
| "source_checkpoint": "/pcache-mnt/ro/checkpoint/854014/duleilei.dll/ckpts/tiny_old_sft2_plus_v6_0914_64gpu_muon_8task_nowall_oldemo_pt1_36k_alias100_20b_seq4096_all_12p5hz/iter_0039000", | |
| "source_common_metadata": "/pcache-mnt/ro/checkpoint/854014/duleilei.dll/ckpts/tiny_old_sft2_plus_v6_0914_64gpu_muon_8task_nowall_oldemo_pt1_36k_alias100_20b_seq4096_all_12p5hz/iter_0039000/for_infer/pure_common.pt", | |
| "source_metadata_sha256": "71b4d63075f127b8d1fd73c3a814dbab9427b6752470ecd94bfce83b201cc64c", | |
| "source_iteration": 39000, | |
| "conversion": "Megatron DCP global model tensors to LingV3 TTS HF safetensors; optimizer excluded; MTP=0", | |
| "schema": { | |
| "checkpoint_path": "/pcache-mnt/ro/checkpoint/854014/duleilei.dll/ckpts/tiny_old_sft2_plus_v6_0914_64gpu_muon_8task_nowall_oldemo_pt1_36k_alias100_20b_seq4096_all_12p5hz/iter_0039000", | |
| "iteration": 39000, | |
| "source_tensor_count": 1475, | |
| "source_component_counts": { | |
| "audio": 643, | |
| "flowloss": 236, | |
| "linear_proj_audio": 119, | |
| "model": 473, | |
| "spk_head": 2, | |
| "stop_head": 2 | |
| }, | |
| "excluded_tensor_count": 2421, | |
| "excluded_component_counts": { | |
| "chained_0": 2283, | |
| "chained_1": 138 | |
| }, | |
| "source_dtype_counts": { | |
| "bfloat16": 1416, | |
| "float32": 59 | |
| }, | |
| "target_tensor_count": 10285, | |
| "target_component_counts": { | |
| "audio": 643, | |
| "flowloss": 236, | |
| "linear_proj_audio": 119, | |
| "model": 9283, | |
| "spk_head": 2, | |
| "stop_head": 2 | |
| }, | |
| "target_dtype_counts": { | |
| "bfloat16": 10226, | |
| "float32": 59 | |
| }, | |
| "target_parameter_count": 8981004468, | |
| "target_total_bytes": 17962089128, | |
| "target_schema_sha256": "627a6b3a872b210e38d5c30778305155b9a0f6710924a33558226e85aaf58fa3", | |
| "groups": [ | |
| { | |
| "name": "model-base", | |
| "source_tensor_count": 3, | |
| "target_tensor_count": 3 | |
| }, | |
| { | |
| "name": "model-layer-00", | |
| "source_tensor_count": 17, | |
| "target_tensor_count": 18 | |
| }, | |
| { | |
| "name": "model-layer-01", | |
| "source_tensor_count": 21, | |
| "target_tensor_count": 404 | |
| }, | |
| { | |
| "name": "model-layer-02", | |
| "source_tensor_count": 21, | |
| "target_tensor_count": 404 | |
| }, | |
| { | |
| "name": "model-layer-03", | |
| "source_tensor_count": 16, | |
| "target_tensor_count": 399 | |
| }, | |
| { | |
| "name": "model-layer-04", | |
| "source_tensor_count": 21, | |
| "target_tensor_count": 404 | |
| }, | |
| { | |
| "name": "model-layer-05", | |
| "source_tensor_count": 21, | |
| "target_tensor_count": 404 | |
| }, | |
| { | |
| "name": "model-layer-06", | |
| "source_tensor_count": 21, | |
| "target_tensor_count": 404 | |
| }, | |
| { | |
| "name": "model-layer-07", | |
| "source_tensor_count": 16, | |
| "target_tensor_count": 399 | |
| }, | |
| { | |
| "name": "model-layer-08", | |
| "source_tensor_count": 21, | |
| "target_tensor_count": 404 | |
| }, | |
| { | |
| "name": "model-layer-09", | |
| "source_tensor_count": 21, | |
| "target_tensor_count": 404 | |
| }, | |
| { | |
| "name": "model-layer-10", | |
| "source_tensor_count": 21, | |
| "target_tensor_count": 404 | |
| }, | |
| { | |
| "name": "model-layer-11", | |
| "source_tensor_count": 16, | |
| "target_tensor_count": 399 | |
| }, | |
| { | |
| "name": "model-layer-12", | |
| "source_tensor_count": 21, | |
| "target_tensor_count": 404 | |
| }, | |
| { | |
| "name": "model-layer-13", | |
| "source_tensor_count": 21, | |
| "target_tensor_count": 404 | |
| }, | |
| { | |
| "name": "model-layer-14", | |
| "source_tensor_count": 21, | |
| "target_tensor_count": 404 | |
| }, | |
| { | |
| "name": "model-layer-15", | |
| "source_tensor_count": 16, | |
| "target_tensor_count": 399 | |
| }, | |
| { | |
| "name": "model-layer-16", | |
| "source_tensor_count": 21, | |
| "target_tensor_count": 404 | |
| }, | |
| { | |
| "name": "model-layer-17", | |
| "source_tensor_count": 21, | |
| "target_tensor_count": 404 | |
| }, | |
| { | |
| "name": "model-layer-18", | |
| "source_tensor_count": 21, | |
| "target_tensor_count": 404 | |
| }, | |
| { | |
| "name": "model-layer-19", | |
| "source_tensor_count": 16, | |
| "target_tensor_count": 399 | |
| }, | |
| { | |
| "name": "model-layer-20", | |
| "source_tensor_count": 21, | |
| "target_tensor_count": 404 | |
| }, | |
| { | |
| "name": "model-layer-21", | |
| "source_tensor_count": 21, | |
| "target_tensor_count": 404 | |
| }, | |
| { | |
| "name": "model-layer-22", | |
| "source_tensor_count": 21, | |
| "target_tensor_count": 404 | |
| }, | |
| { | |
| "name": "model-layer-23", | |
| "source_tensor_count": 16, | |
| "target_tensor_count": 399 | |
| }, | |
| { | |
| "name": "audio", | |
| "source_tensor_count": 643, | |
| "target_tensor_count": 643 | |
| }, | |
| { | |
| "name": "flowloss", | |
| "source_tensor_count": 236, | |
| "target_tensor_count": 236 | |
| }, | |
| { | |
| "name": "linear-proj-audio", | |
| "source_tensor_count": 119, | |
| "target_tensor_count": 119 | |
| }, | |
| { | |
| "name": "spk-head", | |
| "source_tensor_count": 2, | |
| "target_tensor_count": 2 | |
| }, | |
| { | |
| "name": "stop-head", | |
| "source_tensor_count": 2, | |
| "target_tensor_count": 2 | |
| } | |
| ], | |
| "independent_target_schema": "/workspace/bin/moe-model-convert/schemas/lingv3_moe_tts_8b.json", | |
| "independent_target_schema_validated": true | |
| }, | |
| "audio_config_source": "embedded-inference-architecture", | |
| "campplus_source": "/workspace/bin/moe-model-convert/assets/campplus.onnx", | |
| "campplus_sha256": "a6ac6a63997761ae2997373e2ee1c47040854b4b759ea41ec48e4e42df0f4d73", | |
| "tts_config": { | |
| "protocol_version": "thinking_v1", | |
| "prompt_format": "legacy_antmmf", | |
| "tokenizer_contract": "legacy_antmmf", | |
| "sampling_contract": "native_fp32", | |
| "enable_tts_target_language": true, | |
| "infer_tts_language_from_text": true, | |
| "conditioning_mode": "speaker_only", | |
| "sampling_temperature": 1.5, | |
| "prompt_eos_token_id": 156895, | |
| "sample_rate": 44100, | |
| "speaker_embedding_dim": 192, | |
| "latent_dim": 64, | |
| "patch_size": 1, | |
| "history_size": 1, | |
| "audio_patch_token_id": 157164, | |
| "audio_start_token_id": 157165, | |
| "audio_end_token_id": 157166, | |
| "speaker_start_token_id": 157175, | |
| "speaker_end_token_id": 157176, | |
| "stop_threshold": 0.5, | |
| "minimum_decode_steps": 13, | |
| "maximum_decode_steps": 250, | |
| "decode_alignment": "old_shift", | |
| "task_contract": "lingv3_sft_multitask_v1", | |
| "task_decode_alignment": { | |
| "podcast": "old_shift" | |
| }, | |
| "with_rml": true | |
| }, | |
| "tokenizer": { | |
| "source": "/pcache-mnt/ro/checkpoint/854014/duleilei.dll/ckpts/tiny_old_sft2_plus_v6_0914_64gpu_muon_8task_nowall_oldemo_pt1_36k_alias100_20b_seq4096_all_12p5hz/iter_0039000/for_infer", | |
| "contract": "legacy_antmmf", | |
| "prompt_eos_token_id": 156895, | |
| "tts_token_ids": { | |
| "<audioPatch>": 157164, | |
| "<audio>": 157165, | |
| "</audio>": 157166, | |
| "<spk>": 157175, | |
| "</spk>": 157176 | |
| }, | |
| "source_files_sha256": { | |
| "tokenizer.json": "7e831c9b79b2cc4af0d3c7c7878a073371a5cbbaf1f4cb33f6fbce20417722f2", | |
| "tokenizer_config.json": "7368e7cb06ef86c782f959b9796e21622aae4aace075c2cd5356604277ee2dee", | |
| "special_tokens_map.json": "97cb54787ce5ed537fe2b0211768bfce371defabe8241d39459493de9f23b1fc" | |
| }, | |
| "portable_wrapper": "PreTrainedTokenizerFast", | |
| "multimodal_token_count": 24 | |
| }, | |
| "inference_contract": { | |
| "training_recipe": "rc1.6", | |
| "runtime_protocol": "thinking_v1", | |
| "prompt_protocol": "legacy_antmmf", | |
| "tokenizer_contract": "legacy_antmmf", | |
| "sampling_contract": "native_fp32", | |
| "decode_alignment": "old_shift", | |
| "task_decode_alignment": { | |
| "podcast": "old_shift" | |
| }, | |
| "decode_policy": { | |
| "minimum_decode_steps": 13, | |
| "maximum_decode_steps": 250, | |
| "stop_threshold": 0.5, | |
| "basis": "1--20 second training filter at 12.5 latent Hz" | |
| }, | |
| "llm_hidden_source": "24-layer main decoder final hidden (MTP=0)", | |
| "autoregressive_feedback": "Flow latent -> Aggregator -> LLM KV cache", | |
| "flow": { | |
| "latent_dim": 64, | |
| "patch_size": 1, | |
| "history_size": 1, | |
| "depth": 16 | |
| } | |
| }, | |
| "topology": { | |
| "world_size": 64, | |
| "tensor_parallel": 1, | |
| "pipeline_parallel": 1, | |
| "expert_parallel": 8, | |
| "context_parallel": 1 | |
| } | |
| } | |