Abiray/LTX2.3-10Eros-1.4-Split-GGUF

LTX2.3 - 10Eros v1.4 (Unified Split & GGUF)

This repository serves as a unified repo for the 10Eros v1.4 model. We have combined the split architecture required for advanced workflows with a comprehensive selection of GGUF quantizations for efficient local execution.

• Original Creator: TenStrip
• Goal: Providing an all-in-one, easy-to-use directory for ComfyUI and local inference environments.

---

⚖️ Available Models & Sizes

Inside the model/ directory, you will find the base FP8 mixed model alongside the following GGUF options. Choose the one that best fits your VRAM limits:

Format	Quantization	Size
FP8 (Mixed)	fp8mixed_learned	Base Checkpoint
8-Bit	Q8_0	22.8 GB
6-Bit	Q6_K	17.8 GB
5-Bit	Q5_K_M	16.1 GB
5-Bit	Q5_K_S	15.0 GB
4-Bit	Q4_K_M	14.3 GB
4-Bit	Q4_K_S	13.2 GB
3-Bit	Q3_K_M	11.1 GB
3-Bit	Q3_K_S	10.3 GB

---

📂 Repository Structure

Everything is organized to be plug-and-play. The split files are in their respective directories, and all GGUF models are consolidated inside the model/ folder.

📦 LTX2.3-10Eros-1.4-Split-GGUF
 ┣ 📂 audio_vae     (10Eros_v1.4_audio_vae.safetensors)
 ┣ 📂 text_encoder  (10Eros_v1.4_text_encoder.safetensors)
 ┣ 📂 vae           (10Eros_v1.4_vae.safetensors)
 ┗ 📂 model
   ┣ 📜 10Eros_v1.4_fp8mixed_learned_model.safetensors
   ┣ 📜 10Eros_v1.4-Q3_K_M.gguf (11.1 GB)
   ┣ 📜 10Eros_v1.4-Q3_K_S.gguf (10.3 GB)
   ┣ 📜 10Eros_v1.4-Q4_K_M.gguf (14.3 GB)
   ┣ 📜 10Eros_v1.4-Q4_K_S.gguf (13.2 GB)
   ┣ 📜 10Eros_v1.4-Q5_K_M.gguf (16.1 GB)
   ┣ 📜 10Eros_v1.4-Q5_K_S.gguf (15.0 GB)
   ┣ 📜 10Eros_v1.4-Q6_K.gguf   (17.8 GB)
   ┗ 📜 10Eros_v1.4-Q8_0.gguf   (22.8 GB)

---

ℹ️ About 10Eros v1.4

Built off v1.3, bringing back explicit prompting and motion without the common issues of anatomy redraw. This version is designed to be a solid base for training and advanced LoRA stacking.

🛠 Essential Resources

• Recommended Workflows: 10Eros Workflows
• Custom Nodes: 10S-Comfy-nodes
• Distilled LoRAs: Use the cond_safe experiments to avoid harming the fine-tune.

---

💡 Prompting Guide

LTX is strict and directive. It has very little self-reasoning; you must command every motion, evolution, and audio cue.

For optimal results, use an Uncensored LLM or Grok to generate your prompts using this foreword:

"Generate a video scene script with a description based on the attached image for an LLM that has a tokenizer that uses interleaved attention to support long-context understanding that is fed into a multimodal video model. Strict specification, follow up to the word: No timestamps. No unnecessary embellishment. Output only plain English text and make it a copy box.

First, describe the image initial scene in concise natural language; subject(s), subject(s) appearance, subject(s) composition and pose, background, and context.

Next, formulate a naturally evolving scenario that would take place describing every moving body part, composition change, and manipulation from the uploaded initial frame that would be reflected in the video models post-latent evolution output. If the image is explicit or sexual in nature, use full anatomical terminology and spice it up slightly with visually representable erotic themes.

Center the prompt around this basic idea: [ concept ]

Interweave this dialogue or sound concept into the scene with descriptions of voice tone followed by the lines delivered in quotations, in a temporal sequence between or during motions. Dialogue should be concise and non-rambling as it will take away from video quality: [ dialogue ]

Inside that prompt describe only notable audio and audio queues, both normal and explicit; background noise as well as foley and natural sounds. In a temporal sequence paired with coinciding motions. In the case of absent dialogue or soundscapes and only if background music is fitting; describe a fitting genre and melodic tone with matching mood.

Output only text following above instruction. Follow-up suggestions should be on the topic of expanding or changing motion or dialogue from the output text."

---

Credits to TenStrip for the original 10Eros fine-tune and research.