Downloads · 30 days
9
30% of all-time downloads
recorrnet-v2-1ch-aishell1mix-ver1-epoch7
playwithmino/recorrnet-v2-1ch-aishell1mix-ver1-epoch7
recorrnet-v2-1ch-aishell1mix-ver1-epoch7 is a audio-to-audio model from playwithmino. Use it for the audio-to-audio task on the model card, and read the license before you ship it in a product. It is set up for sr-corrnet. The card lists the license as mit.
Single-channel RecorrNet-v2 (direct 9-tap filter head) fine-tuned on AISHELL-1 Mix ver1 (variable 1–5 speakers, mixclean + mixboth).
All-time downloads
30
Public
Repo size
56.7 MB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.pt56.7 MB · 100%
At a glance
- Task
- Audio-to-Audio
- Library
- sr-corrnet
- License
- mit
- Access
- Public
- Created
- Aug 29, 2026
- Updated
- Aug 29, 2026
- SHA
- 6cb13ca2
Model card
From the Hugging Face model README
RecorrNet-v2 — AISHELL-1 Mix ver1 epoch 7
Single-channel RecorrNet-v2 (direct 9-tap filter head) fine-tuned on AISHELL-1 Mix ver1 (variable 1–5 speakers, mix_clean + mix_both).
- Architecture: RecorrNet-v2 (
1ch_aishell1mix_var_2_5spk_recorr_v2) - Init: English WSJ0 variable 2–5spk SR-CorrNet
- Checkpoint:
epoch.0007.pth(end of epoch 7; Aishell-only, not AliMeeting continue) - Code: vu-duy-tung/byd-speech-separation (
recorrnetbranch)
This is the RecorrNet epoch-7 ver1 checkpoint used in the n100 SR vs Recorr vs later AliMeeting-continue comparisons.
Load
from sr_corrnet import SSInference
model = SSInference.from_pretrained(
"playwithmino/recorrnet-v2-1ch-aishell1mix-ver1-epoch7",
device="cuda:0",
)
result = model.process_file("mix.wav", output_dir="output/")
Files in this repo: model.pt (weights only) and config.yaml.
- Task
- Audio-to-Audio
- Library
- sr-corrnet
- License
- mit
- Created
- Aug 29, 2026
- Updated
- Aug 29, 2026