Create README.md

README.md

@@ -0,0 +1,80 @@
+---
+tags:
+- lora
+- nsfw
+- text-to-video
+- motion-enhancement
+base_model: NSFW-API/NSFW_Wan_1.3b
+---
+# NSFW Wan 1.3b Motion Helper LoRA
+
+## LoRA Description
+
+The **NSFW Wan 1.3b Motion Helper LoRA** is a Low-Rank Adaptation specifically designed to significantly enhance video quality, motion consistency, and overall visual style when used with the **[NSFW Wan 1.3B T2V text-to-video model](https://huggingface.co/NSFW-API/NSFW_Wan_1.3b)**.
+
+It is particularly optimized for and trained with the `wan_1.3B_e10.safetensors` checkpoint of the base model.
+
+## LoRA Details
+
+*   **Type:** LoRA
+*   **Base Model Compatibility:** [NSFW Wan 1.3B T2V](https://huggingface.co/NSFW-API/NSFW_Wan_1.3b)
+*   **Recommended Base Checkpoint:** `wan_1.3B_e10.safetensors`
+*   **Purpose:** To improve video quality, temporal coherence, motion rendering, and stylistic elements in generations from the base NSFW Wan 1.3B T2V model.
+
+## Why This LoRA Exists
+
+As detailed in the main [NSFW Wan 1.3B T2V model card](https://huggingface.co/NSFW-API/NSFW_Wan_1.3b), the base checkpoint was primarily fine-tuned on images. While capable, this can sometimes lead to:
+*   Deterioration in video quality.
+*   Inconsistencies in motion or temporal coherence.
+*   Less refined stylistic output in video format.
+
+This LoRA was trained specifically to address these limitations, acting as a "motion and quality" enhancement layer on top of the base model's understanding of NSFW content.
+
+## How to Use
+
+1.  **Download the Base Model:** Ensure you have the [NSFW Wan 1.3B T2V model](https://huggingface.co/NSFW-API/NSFW_Wan_1.3b). The `wan_1.3B_e10.safetensors` checkpoint is highly recommended for use with this LoRA.
+2.  **Download this LoRA:** Get the `NSFW_Wan_1.3b_motion_helper.safetensors` file.
+3.  **Load in Your Pipeline:** In your preferred text-to-video generation environment, load the base model checkpoint (e.g., `wan_1.3B_e10.safetensors`) and then apply this LoRA.
+4.  **Prompt as Usual:** Craft your text prompts as you normally would for the base model. Refer to the `prompting-guide.json` in the [base model repository](https://huggingface.co/NSFW-API/NSFW_Wan_1.3b) for guidance.
+5.  **Generate:** Enjoy improved video outputs!
+
+## Expected Improvements
+
+Using this LoRA with the recommended base checkpoint should result in:
+*   Smoother and more natural motion.
+*   Better temporal consistency across frames.
+*   Enhanced visual fidelity and detail in video sequences.
+*   Improved overall style and aesthetic coherence for NSFW video content.
+
+## Base Model Repository
+
+This LoRA is intended to be used with the main NSFW Wan 1.3B T2V model. You can find the base model, its various epoch checkpoints, and the `prompting-guide.json` here:
+
+[**NSFW Wan 1.3B T2V on Hugging Face**](https://huggingface.co/NSFW-API/NSFW_Wan_1.3b)
+
+## Limitations
+
+*   This LoRA enhances the output of the base model but is still bound by its inherent capabilities and limitations (e.g., understanding of highly complex or abstract prompts).
+*   The focus remains on NSFW content; performance on SFW prompts will likely be poor.
+*   Visual artifacts or inaccuracies may still occur, though they should be reduced compared to using the base model for video without this LoRA.
+*   All limitations and biases mentioned in the [base model card](https://huggingface.co/NSFW-API/NSFW_Wan_1.3b) apply.
+
+## Ethical Considerations & Responsible AI
+
+This LoRA is an add-on for a model designed to generate NSFW content. **All ethical considerations, responsible use guidelines, and limitations outlined in the main [NSFW Wan 1.3B T2V model card](https://huggingface.co/NSFW-API/NSFW_Wan_1.3b) apply fully when using this LoRA.**
+
+*   **Intended for adult users (18+/21+ depending on local regulations) only.**
+*   **Must not** be used to create non-consensual depictions of real individuals, to impersonate, defame, harass, or generate content that could cause harm.
+*   Users are solely responsible for ensuring that their use of this LoRA and the content they generate complies with all applicable laws and regulations.
+*   Exercise extreme caution and responsibility if distributing content generated.
+*   The creators do not endorse or condone the creation or distribution of illegal, unethical, or harmful content.
+
+Please refer to the [NSFW Wan 1.3B T2V model card](https://huggingface.co/NSFW-API/NSFW_Wan_1.3b) for a more comprehensive discussion on ethical considerations.
+
+## License
+
+Steal this LoRA!
+
+## Disclaimer
+
+The outputs generated using this LoRA in conjunction with the base model are entirely synthetic and computer-generated. They do not depict real people or events unless explicitly prompted to do so with user-provided data. The developers of this LoRA and the base model are not responsible for the outputs created by users.