Automatic Speech Recognition
Transformers
Safetensors
Chinese
English
ming_v3_asr
text-generation
audio
speech-recognition
contextual-asr
custom_code
Instructions to use ArchiveStudio/Ming-3.0-asr-tiny with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use ArchiveStudio/Ming-3.0-asr-tiny with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("automatic-speech-recognition", model="ArchiveStudio/Ming-3.0-asr-tiny", trust_remote_code=True)# pip install -U transformers accelerate # Load model directly from transformers import AutoModelForCausalLM model = AutoModelForCausalLM.from_pretrained("ArchiveStudio/Ming-3.0-asr-tiny", trust_remote_code=True, device_map="auto") - Notebooks
- Google Colab
- Kaggle
Download config.json from ArchiveStudio/Ming-3.0-asr-tiny: direct link, hf CLI and curl.
- Browser
- Download file 2.35 kB
-
https://huggingface.co/ArchiveStudio/Ming-3.0-asr-tiny/resolve/main/config.json
- Command line
-
hf download hf://ArchiveStudio/Ming-3.0-asr-tiny/config.json
-
curl -L -o config.json https://huggingface.co/ArchiveStudio/Ming-3.0-asr-tiny/resolve/main/config.json
2.35 kB
| { | |
| "model_type": "ming_v3_asr", | |
| "architectures": [ | |
| "MingV3ASRForConditionalGeneration" | |
| ], | |
| "auto_map": { | |
| "AutoConfig": "configuration_ming_v3_asr.MingV3ASRConfig", | |
| "AutoModelForCausalLM": "modeling_ming_v3_asr.MingV3ASRForConditionalGeneration", | |
| "AutoProcessor": "processing_ming_v3_asr.MingV3ASRProcessor" | |
| }, | |
| "tie_word_embeddings": false, | |
| "audio_token_id": 157168, | |
| "text_config": { | |
| "vocab_size": 157184, | |
| "hidden_size": 1536, | |
| "intermediate_size": 4608, | |
| "num_hidden_layers": 24, | |
| "num_attention_heads": 16, | |
| "num_key_value_heads": 16, | |
| "hidden_act": "silu", | |
| "use_qkv_bias": false, | |
| "use_bias": false, | |
| "rms_norm_eps": 1e-06, | |
| "tie_word_embeddings": false, | |
| "embedding_dropout": 0.0, | |
| "attention_dropout": 0.0, | |
| "output_dropout": 0.0, | |
| "initializer_range": 0.02, | |
| "max_position_embeddings": 8192, | |
| "rope_theta": 10000, | |
| "use_cache": true, | |
| "max_window_layers": 20, | |
| "rope_scaling": null, | |
| "pad_token_id": 156892, | |
| "eos_token_id": 156892, | |
| "num_experts": 128, | |
| "num_shared_experts": 1, | |
| "num_experts_per_tok": 8, | |
| "n_group": 8, | |
| "topk_group": 4, | |
| "moe_intermediate_size": 512, | |
| "moe_shared_expert_intermediate_size": 512, | |
| "first_k_dense_replace": 1, | |
| "head_dim": 128, | |
| "output_router_logits": false, | |
| "use_qk_norm": true, | |
| "num_nextn_predict_layers": 1, | |
| "mtp_loss_scaling_factor": 0, | |
| "moe_router_enable_expert_bias": true, | |
| "routed_scaling_factor": 2.5, | |
| "layer_group_size": 4, | |
| "kv_lora_rank": 512, | |
| "q_lora_rank": 256, | |
| "qk_rope_head_dim": 64, | |
| "v_head_dim": 128, | |
| "qk_nope_head_dim": 128, | |
| "rope_interleave": true, | |
| "score_function": "sigmoid", | |
| "scoring_func": "sigmoid", | |
| "seq_aux": true, | |
| "topk_method": "noaux_tc", | |
| "router_dtype": "fp32", | |
| "gated_attention_proj_granularity_type": "head_wise", | |
| "no_kda_lora": true, | |
| "kda_safe_gate": true, | |
| "kda_lower_bound": -5.0, | |
| "short_conv_kernel_size": 4 | |
| }, | |
| "audio_config": { | |
| "n_mels": 128, | |
| "n_audio_ctx": 3000, | |
| "n_audio_state": 1280, | |
| "n_audio_head": 20, | |
| "n_audio_layer": 32, | |
| "downsampler_kernel_size_list": [ | |
| 3 | |
| ], | |
| "downsampler_stride_list": [ | |
| 2 | |
| ] | |
| }, | |
| "asr_output_format": "plain", | |
| "supported_modalities": [ | |
| "audio" | |
| ], | |
| "dtype": "bfloat16" | |
| } | |