Add GitHub code link to metadata table

Remove Open in HF Spaces badge

Update banner link, add Discord/Spaces badges and Resemble logo

Update Dramabox.png banner

Add DramaBox logo at top of model card

Add DramaBox logo

Refresh sample outputs: ref + generated audio for each, add football commentator + Backstreet Boys

Fix base model: Lightricks/LTX-2.3 (not LTX-2)

Drop sample-audio dataset row from model card link table (Notion-only)

Drop misleading long-form override section (duration is auto-estimated)

Drop sample-outputs preamble

Drop auto-rescale + silence-patch internals from model card

Trim verbose preamble before Inference parameters table

Rewrite model card: proper inference docs, full parameter table, working audio embeds

Move LTX acknowledgement to top of model card; add base_model link in metadata block

Strip video_embeddings_connector + video_aggregate_embed (saves 4.77 GB)

Strip unused text_embedding_projection.video_aggregate_embed.* (saves 1.54 GB)

Revert license section to a short pointer; LICENSE file carries full terms

Model card: add proper LTX-2 attribution, license section, commercial-use notice

Add LTX-2 Community License (required by upstream license §3b)

Simplify language support to English only (multilingual release deferred)

Replace DiT with v15 best_step_09700 (LoRA merged, bf16, bit-identical to runtime-LoRA inference)

Upload dramabox-audio-components.safetensors with huggingface_hub

Initial release: Dramabox v1 - Expressive TTS with Voice Cloning