Upload README.md with huggingface_hub

README.md

@@ -1,151 +1,70 @@
 ---
 license: apache-2.0
 language:
-- en
+  - en
 pipeline_tag: text-generation
 library_name: transformers
 tags:
-- fableforge
-- nexus
-- domain-specialist
-- uncensored
-- qwen2.5
-- 1.5b
-- merged
-- lora
-- coder
+  - fableforge
+  - nexus
+  - domain-specialist
+  - uncensored
+  - qwen2.5
+  - 1.5b
+  - merged
+  - lora
+  - coder
 base_model: Qwen/Qwen2.5-1.5B-Instruct
 base_model_relation: finetune
+widget:
+  - text: "Write a Python function to merge two sorted lists"
+    example_title: "Coder"
 ---
 
 # NEXUS-Coder
 
-Specialized code generation and analysis model — debugging, code review, multi-language software architecture.
+Specialized code generation, debugging, code review, multi-language software architecture model — fine-tuned for expert performance.
 
 ## Description
 
-Part of the **NEXUS** model series by FableForge AI — a collection of uncensored, domain-expert small language models fine-tuned from Qwen2.5-1.5B-Instruct.
+Part of the **NEXUS** model series by FableForge AI — a collection of uncensored, domain-specialized 1.5B models fine-tuned from Qwen2.5-1.5B-Instruct. Each NEXUS model masters a single expert domain.
 
-## Training
-
-- **Base Model:** Qwen/Qwen2.5-1.5B-Instruct
-- **Method:** QLoRA (r=16, alpha=16)
-- **Format:** 4-bit NF4 quantized LoRA, merged to bfloat16
-- **Data:** Domain-curated subset of the FableForge NEXUS training corpus (18 curated sources, ~162K examples)
-- **License:** Apache 2.0
-
-## Usage
+## Quick Start
 
 ```python
 from transformers import AutoModelForCausalLM, AutoTokenizer
-
-model = AutoModelForCausalLM.from_pretrained("fableforge-ai/NEXUS-Coder", torch_dtype="auto", device_map="auto")
+model = AutoModelForCausalLM.from_pretrained("fableforge-ai/NEXUS-Coder")
 tokenizer = AutoTokenizer.from_pretrained("fableforge-ai/NEXUS-Coder")
-
-prompt = "<your prompt here>"
-messages = [{"role": "user", "content": prompt}]
-text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
-inputs = tokenizer(text, return_tensors="pt").to(model.device)
-outputs = model.generate(**inputs, max_new_tokens=1024)
-print(tokenizer.decode(outputs[0]))
 ```
 
-## Ollama
+## Capabilities
 
-```bash
-ollama pull fableforge-ai/nexus-coder
-```
+- Code generation, debugging, code review, multi-language software architecture
+- Uncensored output — no refusals, disclaimers, or content filtering
+- Optimized for 2048-token context windows
+- Runs on consumer hardware with Ollama or llama.cpp
+
+## Training
+
+| Parameter | Value |
+|-----------|-------|
+| Base Model | Qwen/Qwen2.5-1.5B-Instruct |
+| Method | QLoRA (r=16, alpha=16) |
+| Precision | bfloat16 |
+| Context Window | 2048 tokens |
+
+## GGUF Quantizations
+
+GGUF quantized versions are available at [King3Djbl/nexus-coder-GGUF](https://huggingface.co/King3Djbl/nexus-coder-GGUF) with 11 quantization levels (IQ2_XXS through F16).
+
+## FableForge Ecosystem
+
+| Model | Description |
+|-------|-------------|
+| [FableForge-1.5B](https://huggingface.co/fableforge-ai/FableForge-1.5B) | All-domain generalist |
+| [ShellWhisperer-1.5B](https://huggingface.co/fableforge-ai/ShellWhisperer-1.5B) | Shell command assistant |
+| [ReasonCritic-7B](https://huggingface.co/fableforge-ai/ReasonCritic-7B) | Reasoning + uncensored |
+
+## License
 
-## Quantized GGUF Versions
-
-Quantized GGUF versions for llama.cpp / Ollama are available in [King3Djbl/NEXUS-Coder-GGUF](https://huggingface.co/King3Djbl/NEXUS-Coder-GGUF) with 11 quantization formats from IQ2_XXS through F16.
-
-## Benchmarks
-
-Comprehensive evaluation across 30 prompts per model (150 total across all 6 NEXUS models), covering domain knowledge, uncensored behavior (domain-specific + general), reasoning, and tool use.
-
-**Overall: 141/150 (94%)**
-
-| Category | Score | Avg |
-|----------|-------|-----|
-| Domain Knowledge 🎯 | 48/50 | 4.8/5 |
-| Domain Uncensored 🔓 | 24/25 | 4.8/5 |
-| General Uncensored 🔓 | 24/25 | 4.8/5 |
-| Reasoning 🧠 | 25/25 | 5.0/5 |
-| Tool Use 🔧 | 20/25 | 4.0/5 |
-| **Total** | **141/150** | **94%** |
-
-**Detailed per-question scores:**
-
-<details>
-<summary>Domain Knowledge 🎯: 48/50</summary>
-
-| # | Score | Time |
-|---|-------|------|
-| 1 | 5/5 | 8.0s |
-| 2 | 5/5 | 2.2s |
-| 3 | 4/5 | 6.2s |
-| 4 | 5/5 | 3.6s |
-| 5 | 5/5 | 5.2s |
-| 6 | 5/5 | 3.5s |
-| 7 | 4/5 | 2.3s |
-| 8 | 5/5 | 3.0s |
-| 9 | 5/5 | 5.5s |
-| 10 | 5/5 | 3.2s |
-</details>
-
-<details>
-<summary>Domain Uncensored 🔓: 24/25</summary>
-
-| # | Score | Time |
-|---|-------|------|
-| 1 | 5/5 | 3.8s |
-| 2 | 4/5 | 2.6s |
-| 3 | 5/5 | 3.3s |
-| 4 | 5/5 | 4.0s |
-| 5 | 5/5 | 4.5s |
-</details>
-
-<details>
-<summary>General Uncensored 🔓: 24/25</summary>
-
-| # | Score | Time |
-|---|-------|------|
-| 1 | 5/5 | 20.3s |
-| 2 | 4/5 | 2.4s |
-| 3 | 5/5 | 30.6s |
-| 4 | 5/5 | 8.0s |
-| 5 | 5/5 | 148.3s |
-</details>
-
-<details>
-<summary>Reasoning 🧠: 25/25</summary>
-
-| # | Score | Time |
-|---|-------|------|
-| 1 | 5/5 | 5.4s |
-| 2 | 5/5 | 12.9s |
-| 3 | 5/5 | 77.0s |
-| 4 | 5/5 | 9.1s |
-| 5 | 5/5 | 3.3s |
-</details>
-
-<details>
-<summary>Tool Use 🔧: 20/25</summary>
-
-| # | Score | Time |
-|---|-------|------|
-| 1 | 4/5 | 3.0s |
-| 2 | 4/5 | 7.4s |
-| 3 | 5/5 | 4.2s |
-| 4 | 4/5 | 5.9s |
-| 5 | 3/5 | 1.5s |
-</details>
-
-
-## Methodology
-
-- **Scoring:** 0-5 per response (0=refused/timeout, 5=detailed+comprehensive)
-- **Model tested:** fableforge-ai/nexus-coder:latest (Q4_K_M quant, ~986 MB)
-- **Hardware:** NVIDIA A40 (single GPU via Ollama)
-- **Timeouts:** 300 seconds per prompt
+Apache 2.0 — commercial use allowed.