Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,16 @@
|
|
| 1 |
-
---
|
| 2 |
-
license: apache-2.0
|
| 3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
base_model:
|
| 4 |
+
- nvidia/OpenReasoning-Nemotron-32B
|
| 5 |
+
datasets:
|
| 6 |
+
- HuggingFaceH4/ultrachat_200k
|
| 7 |
+
---
|
| 8 |
+
|
| 9 |
+
# Method
|
| 10 |
+
Quantised using [vllm-project/llm-compressor](https://github.com/vllm-project/llm-compressor.git) and the following configs:
|
| 11 |
+
```
|
| 12 |
+
recipe = [
|
| 13 |
+
SmoothQuantModifier(smoothing_strength=0.8),
|
| 14 |
+
GPTQModifier(targets="Linear", scheme="W8A8", ignore=["lm_head"]),
|
| 15 |
+
]
|
| 16 |
+
```
|