Dataset Viewer
Auto-converted to Parquet Duplicate
Search is not available for this dataset
video
video

Enochiatron Showcase

Generated outputs from the Enochiatron Audio+Video LoRA, trained on LTX-2.3 22B DiT.

What's Here

Batch Generation (batch/)

Standard batch-generated clips using the P2 (audio+visual) LoRA checkpoint.

  • 2 clips at different seeds (42, 123)
  • index.html — gallery viewer (download and open locally)
  • batch_manifest.json — generation metadata

P1 vs P2 Comparison (comparison/)

Side-by-side comparison of Phase 1 (visual-only LoRA) vs Phase 2 (audio+visual LoRA):

  • p1/ — Visual-only clips (no audio track)
  • p2/ — Audio+visual clips (Enochian chanting audio)
  • sidebyside/ — ffmpeg hstack composites with P1 left, P2 right
  • index.html — comparison gallery viewer (download and open locally)
  • comparison_manifest.json — full generation metadata

Generation Parameters

Parameter Value
Steps 30
CFG Scale 3.0
STG Scale 1.0
Modality Scale 3.0
LoRA Scale 1.0
Resolution 768×352×81 frames
FPS 25
GPU A100 80GB PCIe
P1 gen time ~104s/clip
P2 gen time ~118s/clip (+15% for audio)

Related Repositories

Repo Description
enochiatron-lora Trained LoRA checkpoints (P1 visual + P2 audio+visual)
enochiatron-muxed Muxed training dataset (1027 Tron clips + 20 Enochian WAV)
enochiatron-encoded Pre-encoded latents (VAE + Gemma cached .pt files)
timm156/enochiatron Source code — full training & inference pipeline

Gallery Viewing

The HTML gallery files use inline CSS with zero external dependencies. To view:

  1. Download the batch or comparison directory
  2. Open index.html in any browser
  3. Videos play directly in the browser

About Enochiatron

Enochiatron trains a simultaneous audio+video LoRA on LTX-2.3 22B DiT, fusing:

  • Video: 1027 clips from Tron (digital world / "inside the computer" aesthetic)
  • Audio: 20 WAV recordings of the complete Enochian language (Angelic Keys / Calls of the Aethyrs)

The result is a generative model that produces video+audio in the aesthetic intersection: digital beings speaking Enochian in neon-lit geometric spaces.

Built with Flimmer (Alvdansen Labs) training toolkit.

Downloads last month
39