Views
No views yet
| File | Size | Format |
|---|---|---|
Saif-Kernel-7B-v0.5-Q4_K_M.gguf | ~4.36 GB | Q4_K_M (llama.cpp compatible) |
1# Download and create
2curl -L -o Saif-Kernel-7B-Q4.gguf "https://huggingface.co/BinSaqban/Saif-Kernel-7B-GGUF/resolve/main/Saif-Kernel-7B-v0.5-Q4_K_M.gguf"
3
4# Create Modelfile
5cat > Modelfile << 'EOF'
6FROM Saif-Kernel-7B-Q4.gguf
7PARAMETER temperature 0.3
8PARAMETER num_ctx 4096
9SYSTEM "You are Saif, a cybersecurity analysis assistant."
10EOF
11
12ollama create saif-kernel -f Modelfile
13ollama run saif-kernel| RAM | Performance |
|---|---|
| 8 GB | ✅ Works with num_ctx 2048, may use swap |
| 16 GB | ✅ Full speed, num_ctx 8192 |
| 32 GB+ | ✅ Max context |
| Model | Purpose |
|---|---|
| Saif-Kernel-7B | 🧠 Core reasoning & cybersecurity analysis |
| Saif-Browser-7B | 🌐 Browser automation & web interaction |
| Saif-Router-7B | 🔀 Query classification & routing |