cometa21/tessy-ltx25

Tessy LoRA - LTX-2.5

LoRA trained on LTX-2.5 (dev transformer) to generate videos of Tessy, a woman with long dark hair, in various TV/podcast settings.

Training details

• Base model: Lightricks/LTX-2.5 (dev transformer)
• Training type: LoRA
• Rank / Alpha: 64 / 64
• Target modules: to_k, to_q, to_v, to_out.0
• Training steps: 500
• Batch size: 1 (grad accum 1)
• Learning rate: 1e-4
• Resolution: 832x480, 121 frames, 24 fps
• Dataset: 39 clips of Tessy
• Trigger word: tessy

Usage

uv run python -m ltx_pipelines.ti2vid_one_stage \
    --transformer-path /path/to/ltx-2.5-22b-dev-transformer-bf16.safetensors \
    --text-encoder-path /path/to/gemma4-12b-with-proj-ltx-2.5-bf16.safetensors \
    --video-vae-path /path/to/ltx-2.5-video-vae-bf16.safetensors \
    --audio-vae-path /path/to/ltx-2.5-audio-vae-bf16.safetensors \
    --lora lora_weights_step_00500.safetensors 1.0 \
    --num-frames 121 \
    --prompt "tessy A static medium close-up shot at eye level features a woman with long dark hair sitting at a podcast desk..." \
    --output-path test_tessy.mp4