Automatic Speech Recognition
Transformers
qwen3-asr
latent-reasoning
test-time-compute
parameter-efficient
Instructions to use voidful/latentASR with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use voidful/latentASR with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("automatic-speech-recognition", model="voidful/latentASR")# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("voidful/latentASR", device_map="auto") - Notebooks
- Google Colab
- Kaggle
Download model_metadata.json from voidful/latentASR: direct link, hf CLI and curl.
- Browser
- Download file 675 Bytes
-
https://huggingface.co/voidful/latentASR/resolve/main/model_metadata.json
- Command line
-
hf download hf://voidful/latentASR/model_metadata.json
-
curl -L -o model_metadata.json https://huggingface.co/voidful/latentASR/resolve/main/model_metadata.json
675 Bytes
| { | |
| "project": "latentASR", | |
| "model_repo": "voidful/latentASR", | |
| "base_model": "Qwen/Qwen3-ASR-0.6B", | |
| "checkpoint_file": "checkpoints/latentASR_adapter.pth", | |
| "source_checkpoint": "activation_500_epoch10.pth", | |
| "checkpoint_sha256": "f0ce39fa5e6952fced6992508f3e2b32ea8467442b545678781a0f04e64f2430", | |
| "train_mode": "latent", | |
| "n_latent": 4, | |
| "train_max_samples": 500, | |
| "adapter_parameters": 5251077, | |
| "value_forced_neg_prob": 0.3, | |
| "latent_use_bounded_delta": true, | |
| "latent_use_injection_gate": true, | |
| "latent_use_embedding_anchor": true, | |
| "deployed_halt_threshold": 0.0, | |
| "paper": "Listen, Think, Transcribe: Continuous Latent Test-Time Scaling for ASR" | |
| } | |