Cseti/LTX2.3-22B_Arcane-Jinx_v1

LTX2.3-22B Arcane Jinx LoRA v1

Prompt

Nfj1nx A rain-soaked urban alley at night, illuminated by flickering neon signs casting red and blue reflections across wet pavement. A woman stands at the center of the frame, wearing a black leather jacket, her dark hair slicked back and glistening with rain. The atmosphere is tense, cinematic, with shallow depth of field and moody chiaroscuro lighting. Puddles reflect the neon glow, steam rises from a nearby grate, and distant thunder echoes through the scene. The shot opens in a medium close-up on the woman, static for a beat as rain drips from her hair. She tilts her head slowly, locking eyes with the camera. She exhales, then says with a calm, cutting voice: "You should've walked away when you had the chance." The camera pushes in slowly toward her face as thunder rumbles, holding on her cold, unwavering stare until the frame tightens into an extreme close-up.

Prompt

[VISUAL]: Nfj1nx. A continuous, high-tension cinematic long take set in a smoke-filled noir salon. The scene begins with the camera at waist level, executing a rapid, smooth arc shot around an elegant young woman in her early 20s. She is wearing a form-fitting black silk evening gown with a high slit, and her hair falls in polished waves over one shoulder. Initially standing still with a sleek pistol at her side, she transitions into action. Keeping her torso perfectly upright and her posture regally straight, she raises the weapon with cold precision. She extends her arm fully and locked, pointing the pistol straight forward toward the lens in a professional aiming stance. The camera follows this movement, transitioning from the rapid arc shot to a steady, frontal, upper-body framing (from the waist up). This stabilized final framing clearly shows her rigid upper body, her extended arm, and the pistol aimed directly forward, with no close-up on her face. The environment features grey smoke catching amber light and dramatic shadows on wooden décor and velvet furniture. [SPEECH]: (Her voice is a smooth, icy whisper that carries a deadly weight) "The music has stopped. It's time we settle the bill, don't you think?"

A character LoRA for LTX-Video 2.3 (22B) trained with the audio-video training mode, so the character's voice and speech delivery are baked into the adapter alongside her appearance and motion. Trained on 128 short Arcane clips (video + stereo audio).

Training details

• Base model: Lightricks/LTX-2.3 (22B)
• Training framework: ltx-trainer (Lightricks)
• Training strategy: text-to-video + audio branch (with_audio: true)
• Best checkpoint: step 21,000 (out of 22,000 total)
• LoRA rank / alpha: 128 / 128
• Target modules: to_k, to_q, to_v, to_out.0 (video + audio + cross-modal attention)
• Optimizer: Prodigy (auto-LR, lr=1.0 scale factor)
• Mixed precision: bf16
• Batch size: 1 (gradient checkpointing on)
• First-frame conditioning: 0.3 (the adapter also works in image-to-video mode)
• Resolution buckets: 960x544 @ 49 / 97 / 121 / 145 / 193 frames
• Dataset: 128 short clips

Inference

For inference I used ComfyUI.

Trigger word: Nfj1nx

Strength: 0.8-1.0.

Important Notes

This LoRA is created as part of a fan project for research purposes only and is not intended for commercial use. It is based on the TV series "Arcane" which is protected by copyright. Users utilize the model at their own risk. Users are obligated to comply with copyright laws and applicable regulations. The model has been developed for non-commercial purposes, and it is not my intention to infringe on any copyright. I assume no responsibility for any damages or legal consequences arising from the use of the model.

Acknowledgement

Special thanks to Lightricks for open-sourcing the LTX-2 trainer and releasing the 22B model with audio support.

Support

Fine-tuning models like this requires renting cloud GPUs, which gets expensive quickly. If you find this LoRA useful and would like me to keep contributing open-source models, your support is very much appreciated: