corrected --matrix to --imatrix
Browse files
README.md
CHANGED
@@ -119,7 +119,7 @@ cd <llama.cpp directory>
|
|
119 |
```
|
120 |
4. Use the generated matrix file to quantise the model (see further down for some pre-computed matrix files)
|
121 |
```
|
122 |
-
./quantize --
|
123 |
```
|
124 |
Note: normal quantisation also benefits from using a matrix file. It also seem that a bigger input matrix is
|
125 |
better for higher quantisation.
|
|
|
119 |
```
|
120 |
4. Use the generated matrix file to quantise the model (see further down for some pre-computed matrix files)
|
121 |
```
|
122 |
+
./quantize --imatrix <output.matrix> <model_path>/ggml-model-f16.gguf <quantisation_level, eg:IQ4_XS>
|
123 |
```
|
124 |
Note: normal quantisation also benefits from using a matrix file. It also seem that a bigger input matrix is
|
125 |
better for higher quantisation.
|