Update README.md
Browse files
README.md
CHANGED
|
@@ -10,4 +10,16 @@ base_model:
|
|
| 10 |
|
| 11 |
project original [source](https://huggingface.co/microsoft/Phi-3-mini-128k-instruct)
|
| 12 |
|
|
|
|
|
|
|
|
|
|
|
|
|
| 13 |
Q_4_K_M is recommanded (good for running with CPU as well)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 10 |
|
| 11 |
project original [source](https://huggingface.co/microsoft/Phi-3-mini-128k-instruct)
|
| 12 |
|
| 13 |
+
Q_2_K (not nice)
|
| 14 |
+
|
| 15 |
+
Q_3_K_M (acceptable)
|
| 16 |
+
|
| 17 |
Q_4_K_M is recommanded (good for running with CPU as well)
|
| 18 |
+
|
| 19 |
+
Q_5_K_M good in general
|
| 20 |
+
|
| 21 |
+
Q_6_K good also; if you want a better results; take this one instead of Q_5_K_M
|
| 22 |
+
|
| 23 |
+
Q_8_0 very good; need a reasonable size of RAM otherwise you might expect a long wait
|
| 24 |
+
|
| 25 |
+
16-bit and 32-bit are not provided here; since the size is similar to the original safetensors; once you have a good GPU, go ahead with the saftetensors
|