Upload quantize_config.json with huggingface_hub
Browse files- quantize_config.json +6 -1
quantize_config.json
CHANGED
|
@@ -1,7 +1,12 @@
|
|
| 1 |
{
|
| 2 |
"modes": [
|
| 3 |
"fp16",
|
| 4 |
-
"q8"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 5 |
],
|
| 6 |
"per_channel": true,
|
| 7 |
"reduce_range": true,
|
|
|
|
| 1 |
{
|
| 2 |
"modes": [
|
| 3 |
"fp16",
|
| 4 |
+
"q8",
|
| 5 |
+
"int8",
|
| 6 |
+
"uint8",
|
| 7 |
+
"q4",
|
| 8 |
+
"q4f16",
|
| 9 |
+
"bnb4"
|
| 10 |
],
|
| 11 |
"per_channel": true,
|
| 12 |
"reduce_range": true,
|