Chat model by kaitchup · 3.82B parameters
Phi 3 mini 4k instruct gptq 4bit fits comfortably in your 24 GB Mac using Q8_0 quantization, using 20% of your RAM.
~42 tok/s
Q8_0 · 4.8 GB
| Quantization | Memory | Speed | Fits? |
|---|---|---|---|
| Q8_0 Recommended | 4.8 GB | ~42 tok/s | ✓ |
| Q6_K | 3.8 GB | ~55 tok/s | ✓ |
| Q5_K_M | 3.3 GB | ~65 tok/s | ✓ |
| Q4_K_M | 3.0 GB | ~76 tok/s | ✓ |
| Q3_K_M | 2.6 GB | ~92 tok/s | ✓ |
| Q2_K | 2.2 GB | ~119 tok/s | ✓ |
ToolPiper downloads, manages, and runs models with one click. Apple Silicon optimized.
Get ToolPiper — Free