Skip to content

DPDFNet: causal single-channel speech enhancement that boosts DeepFilterNet2 with dual-path RNN blocks for stronger long-range temporal and cross-band modeling. Repo includes PyTorch implementation + checkpoints, TFLite models with inference code, and a real-time demo.

active 2025-12-182026-06-20 (UTC)

Partial coverage20,279 / 26,018 hourly files (78%) · 2 absent upstream · 5,737 failed, retryable2023-08-242026-08-12 (UTC)— sampled evenly across the window, so rankings and trends hold; absolute counts scale up.
Events
74
Pushes
35
Pull requests
1
Issues
6
Stars
19
Forks
1

Activity over time

Daily event counts in the loaded window

Line chart, 185 days from 2025-12-18 to 2026-06-20. Pushes: 35 total, peak 14 in a day. Pull requests: 1 total, peak 1 in a day. Issues: 6 total, peak 1 in a day. Comments: 11 total, peak 2 in a day. Stars: 19 total, peak 2 in a day.

  • Pushes
  • Pull requests
  • Issues
  • Comments
  • Stars

Stars, PRs, issues and forks are under-captured in the later part of this window. GH Archive progressively stopped capturing non-push events during 2026 — −95% or worse by the end of the window. Every series here except Pushes fades for that reason, so a decline above reflects the archive, not this repository. Pushes stay reliable throughout, so read them, and the contributor counts derived from them, as the real signal. Data health has the measurements.

Top contributors

Pushes, PRs, issues, reviews and comments — stars and forks excluded, so this is contribution rather than popularity

ContributorContributionsPushesPRsComments
danielr-ceva473519
hlevring1000
openff1001
infometa1000
JIULIGUAN1000
he-jinlai1000
sakshibansalsb1001

Recent activity

Latest issues, pull requests and releases

  • Issue#12he-jinlai2026-04-22 01:21
    dpdfnet.onnx with NT8 quantization?
  • Issue comment#14danielr-ceva2026-04-18 17:45
    Does the dpdfnet model have any obvious drawbacks?
  • Issue comment#14sakshibansalsb2026-04-18 13:29
    Does the dpdfnet model have any obvious drawbacks?
  • Issue comment#14openff2026-04-17 04:06
    Does the dpdfnet model have any obvious drawbacks?
  • Issue comment#14danielr-ceva2026-04-16 06:47
    Does the dpdfnet model have any obvious drawbacks?
  • Releasedanielr-ceva2026-03-20 20:08
    v0.4.0
  • Issue comment#6danielr-ceva2026-03-06 10:22
    dpdfnet8 with 48khz support
  • Issue#10infometa2026-02-26 07:40
    Any plans to open-source the training code ?
  • Pull request#9danielr-ceva2026-02-23 11:04
  • Issue comment#6danielr-ceva2026-02-10 11:37
    dpdfnet8 with 48khz support
  • Issue comment#7danielr-ceva2026-02-07 12:54
    Curious about lookahead
  • Issue comment#6danielr-ceva2026-02-06 11:32
    dpdfnet8 with 48khz support
  • Issue#6hlevring2026-02-05 20:45
    dpdfnet8 with 48khz support
  • Issue comment#5danielr-ceva2026-02-05 07:03
    Training code
  • Issue#3danielr-ceva2026-01-17 08:50
    Speech Over-Suppression in TFLite Models + Request for Training Scripts
  • Issue comment#3danielr-ceva2025-12-29 15:31
    Speech Over-Suppression in TFLite Models + Request for Training Scripts
  • Issue#2JIULIGUAN2025-12-28 05:21
    Does the baseline model include OA Loss and Fine-Tuning?
  • Issue#1danielr-ceva2025-12-27 11:15
    This looks very cool congrats. Will there be a model trained on 48khz?
  • Issue comment#1danielr-ceva2025-12-27 11:15
    This looks very cool congrats. Will there be a model trained on 48khz?

Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 19 stars here means stars gained during the window, not the repo's star count.