SAWT V4, training state
These are the checkpoints needed to continue training SAWT V4. If you only want to restore audio, use Quran-Lab/sawt-v4 instead: it carries the trimmed inference weights and the client code.
| file | what it is |
|---|---|
generator_step150000.pt |
generator at step 150,000: live weights, EMA, optimizer state, standardisation statistics. The resume point. |
generator_step150000_noopt.pt |
the same step without optimizer state. This is the exact file used for every render reported in the technical report. |
anchor_final.pt |
the anchor as trained: w2v-BERT 2.0 LoRA adapters and all four heads, untrimmed. |
vae48_dacvae_full.pt |
the fine-tuned DAC-VAE decoder (D2) as trained, untrimmed. |
The released generator.pt, anchor.pt and vae48_dacvae_d2.pt in the model repository are
derived from these by stripping what inference does not need.
Technical report: https://comparison.quranlab.ai/sawt_v4.pdf
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support