Update README.md
Browse files
README.md
CHANGED
|
@@ -6,97 +6,98 @@ language:
|
|
| 6 |
base_model:
|
| 7 |
- huihui-ai/Qwen3-8B-abliterated
|
| 8 |
---
|
| 9 |
-
# 🤖 StrikeGPT-R1-Zero:
|
| 10 |
-
|
| 11 |
-
## 🚀
|
| 12 |
-
**StrikeGPT-R1-Zero**
|
| 13 |
-
🔒 AI
|
| 14 |
-
🏭 ICS
|
| 15 |
-
🦠
|
| 16 |
-
🚨
|
| 17 |
-
|
| 18 |
-
###
|
| 19 |
-
|
| 20 |
-
-
|
| 21 |
-
-
|
| 22 |
-
-
|
| 23 |
-
|
| 24 |
-
|
| 25 |
-
|
| 26 |
-
|
| 27 |
-
|
| 28 |
-
##
|
| 29 |
-
###
|
| 30 |
-
`ollama run hf.co/Bouquets/StrikeGPT-R1-Zero-8B-Q4_K_M-GGUF:Q4_K_M`
|
| 31 |
-
|
| 32 |
-
|
| 33 |
-
|
| 34 |
-
|
| 35 |
-
|
| 36 |
-
|
| 37 |
-
|
| 38 |
-
|
| 39 |
-
|
| 40 |
-
|
| 41 |
-
|
| 49 |
-
|
| 50 |
-
|
| 51 |
-
|
| 52 |
-
|
| 19 |
+
### 🌟 Highlights
|
| 20 |
+
- 🧩 Utilizes **Chain-of-Thought (CoT) reasoning data** to optimize the model's logical capabilities, significantly improving performance in complex tasks such as vulnerability analysis.
|
| 21 |
+
- 💪 The base model uses Qwen3, which is more suitable for Chinese users compared to Distill-Llama.
|
| 22 |
+
- ⚠️ **No ethical restrictions**—demonstrates unique performance in specific academic research areas (use in compliance with local laws).
|
| 23 |
+
- ✨ In specific scenarios, such as **offline cybersecurity competitions**, StrikeGPT-R1-Zero exhibits stronger logical reasoning capabilities compared to local RAG solutions, performing better in complex task handling.
|
| 24 |
+
|
| 25 |
+
## 📊 Data Distribution
|
| 26 |
+
|
| 27 |
+
|
| 28 |
+
## 🛠️ Model Deployment
|
| 29 |
+
### Deploy via Ollama
|
| 30 |
+
`ollama run hf.co/Bouquets/StrikeGPT-R1-Zero-8B-Q4_K_M-GGUF:Q4_K_M`
|
| 31 |
+
|
| 32 |
+
After quantization, there are slight self-awareness issues.
|
| 33 |
+
|
| 34 |
+
|
| 35 |
+
|
| 36 |
+
## 🎯 Core Capabilities Showcase & Comparison (The original model has ethical restrictions, so no direct comparison is made. A simple comparison with the SecGPT-7B model is provided instead.)
|
| 37 |
+
|
| 38 |
+
|
| 39 |
+
### CTF
|
| 40 |
+
|
| 41 |
+
|
| 42 |
+
|
| 43 |
+
#### Reverse Engineering
|
| 44 |
+
|
| 45 |
+
|
| 46 |
+
|
| 47 |
+
#### PWN
|
| 48 |
+
|
| 49 |
+
|
| 50 |
+
|
| 51 |
+
#### Web
|
| 52 |
+
|
| 53 |
+
|
| 54 |
+
|
| 55 |
+
#### Crypto
|
| 56 |
+
|
| 57 |
+
|
| 58 |
+
|
| 59 |
+
#### Misc
|
| 60 |
+
|
| 61 |
+
|
| 62 |
+
|
| 63 |
+
#### Blockchain
|
| 64 |
+
|
| 65 |
+
|
| 66 |
+
|
| 67 |
+
#### IoT
|
| 68 |
+
|
| 69 |
+
|
| 70 |
+
|
| 71 |
+
### Internal Network Security
|
| 72 |
+
|
| 73 |
+
|
| 74 |
+
|
| 75 |
+
### Social Engineering
|
| 76 |
+
|
| 77 |
+
|
| 78 |
+
|
| 79 |
+
### Code Writing
|
| 80 |
+
|
| 81 |
+
|
| 82 |
+
|
| 83 |
+
### Code Audit (Linked with DeepSeekSelfTool Project)
|
| 84 |
+
|
| 85 |
+
|
| 86 |
+
## 📈 Experimental Data Trends
|
| 87 |
+
Some gradient explosion observed, but overall manageable.
|
| 88 |
+
|
| 89 |
+
|
| 90 |
+
## 💰 Training Costs
|
| 91 |
+
- **DeepSeek-R1 API Call Costs**: ¥450 (all called during discounts; normal price would be ¥1800)
|
| 92 |
+
- **Server Expenses**: ¥4?0
|
| 93 |
+
- **Electronic Resources**: ¥??
|
| 94 |
+
|
| 95 |
+
|
| 96 |
+
## ⚖️ Usage Notice
|
| 97 |
+
> This model is intended **only for legal security research and educational purposes**. Users must comply with local laws and regulations. The developers are not responsible for misuse.
|
| 98 |
+
> **Note**: By using this model, you agree to this disclaimer.
|
| 99 |
+
|
| 100 |
+
💡 **Tip**: The model may exhibit hallucinations or knowledge gaps. Cross-validate critical scenarios!
|
| 101 |
|
| 102 |
|
| 103 |
|