Update README.md
Browse files
README.md
CHANGED
|
@@ -71,11 +71,13 @@ The Apache 2.0 license requires that attribution is included at the point of rel
|
|
| 71 |
<details>
|
| 72 |
<summary>Click to see details</summary>
|
| 73 |
The new methods available are:
|
| 74 |
-
|
| 75 |
-
|
| 76 |
-
|
| 77 |
-
|
| 78 |
-
|
|
|
|
|
|
|
| 79 |
Refer to the Provided Files table below to see what files use which methods, and how.
|
| 80 |
</details>
|
| 81 |
<!-- quantization_methods end -->
|
|
|
|
| 71 |
<details>
|
| 72 |
<summary>Click to see details</summary>
|
| 73 |
The new methods available are:
|
| 74 |
+
|
| 75 |
+
* GGML_TYPE_Q2_K - "type-1" 2-bit quantization in super-blocks containing 16 blocks, each block having 16 weight. Block scales and mins are quantized with 4 bits. This ends up effectively using 2.5625 bits per weight (bpw)
|
| 76 |
+
* GGML_TYPE_Q3_K - "type-0" 3-bit quantization in super-blocks containing 16 blocks, each block having 16 weights. Scales are quantized with 6 bits. This end up using 3.4375 bpw.
|
| 77 |
+
* GGML_TYPE_Q4_K - "type-1" 4-bit quantization in super-blocks containing 8 blocks, each block having 32 weights. Scales and mins are quantized with 6 bits. This ends up using 4.5 bpw.
|
| 78 |
+
* GGML_TYPE_Q5_K - "type-1" 5-bit quantization. Same super-block structure as GGML_TYPE_Q4_K resulting in 5.5 bpw
|
| 79 |
+
* GGML_TYPE_Q6_K - "type-0" 6-bit quantization. Super-blocks with 16 blocks, each block having 16 weights. Scales are quantized with 8 bits. This ends up using 6.5625 bpw
|
| 80 |
+
|
| 81 |
Refer to the Provided Files table below to see what files use which methods, and how.
|
| 82 |
</details>
|
| 83 |
<!-- quantization_methods end -->
|