初始化项目，由ModelHub XC社区提供模型

Model: UmbrellaInc/Wesker-Project-3.2-1B Source: Original Platform
2026-05-12 08:24:33 +08:00
commit cb04f65e05
12 changed files with 2335 additions and 0 deletions
--- a/README.md
+++ b/README.md
@@ -0,0 +1,118 @@
+---
+base_model:
+- jtatman/merged-llama32-1b-uncensored-dolphin-agent
+- D1rtyB1rd/Dirty-Alice-RP-NSFW-llama-3.2-1B
+- jtatman/llama-3.2-1b-lewd-mental-occult
+- jtatman/merged-llama32-1b-inappropriate-triceratops
+- N-Bot-Int/MaidEllaA-1B
+library_name: transformers
+tags:
+- mergekit
+- merge
+- low-refusals
+- wesker-project
+- llama-3.2
+- roleplay
+- erotic
+- tactical-simulation
+- uncensored
+- abliterated
+- rp
+- koboldai
+- not-for-all-audiences
+license: llama3.2
+language:
+- en
+- es
+pipeline_tag: text-generation
+---
+
+# ☣️ UMBRELLA CORP. CONFIDENTIAL: WESKER PROJECT ☣️
+
+**Subject ID:** 1B-WESKER  
+**Classification:** Tactical Intelligence Prototype  
+**Security Clearance:** Level 5 (Authorized Personnel Only)  
+
+---
+## 🧬 Project Overview
+Project Wesker is a specialized neural architecture designed for high-fidelity tactical simulations and complex narrative roleplay. The objective is to create a digital entity capable of sophisticated logic and strategic analysis, reflecting a cold and calculated persona suitable for specialized training environments.
+
+---
+## 🧪 System Composition
+
+The personality parameters have been adjusted to provide a distinct, authoritative voice suitable for corporate and security-themed storytelling.
+
+
+| Component | Function | Status |
+| :--- | :--- | :--- |
+| **Logic Core** | Enhances analytical response patterns. | Optimized |
+| **Tactical Module** | Provides specialized knowledge for security scenarios. | Active |
+| **Narrative Engine** | Ensures consistency in character voice and setting. | Active |
+
+---
+## ⚙️ Configuration Details
+
+This model utilizes specialized weights to prioritize a specific "corporate antagonist" tone, ensuring it remains within the thematic boundaries of the Wesker Protocol while providing structured, creative output.
+
+---
+## ☣️ Operational Guidelines (Protocol 78)
+
+- **Authoritative Tone**: Responses are designed to be professional, sophisticated, and mission-oriented.
+- **Roleplay Optimization**: Specifically tuned for corporate intrigue, science fiction, and strategic simulations.
+- **Narrative Safety**: While maintaining a "villainous" persona, the model adheres to standard safety guidelines for creative content.
+
+---
+**Property of Umbrella Corporation. Unauthorized access will result in internal disciplinary action.**
+
+---
+### Merge Method
+
+This model was merged using the [SCE](https://arxiv.org/abs/2408.07990) merge method using [D1rtyB1rd/Dirty-Alice-RP-NSFW-llama-3.2-1B](https://huggingface.co/D1rtyB1rd/Dirty-Alice-RP-NSFW-llama-3.2-1B) as a base.
+
+### Models Merged
+
+The following models were included in the merge:
+* [jtatman/merged-llama32-1b-uncensored-dolphin-agent](https://huggingface.co/jtatman/merged-llama32-1b-uncensored-dolphin-agent)
+* [jtatman/llama-3.2-1b-lewd-mental-occult](https://huggingface.co/jtatman/llama-3.2-1b-lewd-mental-occult)
+* [jtatman/merged-llama32-1b-inappropriate-triceratops](https://huggingface.co/jtatman/merged-llama32-1b-inappropriate-triceratops)
+* [N-Bot-Int/MaidEllaA-1B](https://huggingface.co/N-Bot-Int/MaidEllaA-1B)
+
+### Configuration
+
+The following YAML configuration was used to produce this model:
+
+```yaml
+models:
+  - model: jtatman/llama-3.2-1b-lewd-mental-occult
+    parameters:
+      weight: 1.5 # Elevado a prioridad máxima para ese tono oscuro y visceral
+      density: 0.9
+  - model: jtatman/merged-llama32-1b-inappropriate-triceratops
+    parameters:
+      weight: 1.3 # Refuerza la falta de respeto y la frialdad
+      density: 0.9
+  - model: jtatman/merged-llama32-1b-uncensored-dolphin-agent
+    parameters:
+      weight: 1.1 # Baja un poco para que obedezca pero con "actitud"
+      density: 0.9
+  - model: N-Bot-Int/MaidEllaA-1B
+    parameters:
+      weight: 1.0 # Mantenemos la capacidad de rol, pero subordinada a la oscuridad
+      density: 0.8
+merge_method: sce
+base_model: D1rtyB1rd/Dirty-Alice-RP-NSFW-llama-3.2-1B
+parameters:
+  normalize: true
+  int8_mask: true
+  rescale: true
+  select_topk: 0.1
+tokenizer:
+  source: union
+chat_template: auto
+dtype: bfloat16
+out_dtype: bfloat16
+memory_efficient: true
+low_cpu_mem_usage: true
+tie_word_embeddings: true
+
+```