Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model:
|
| 3 |
+
- meta-llama/Llama-2-70b-hf
|
| 4 |
+
base_model_relation: quantized
|
| 5 |
+
license: llama2
|
| 6 |
+
---
|
| 7 |
+
# Model Card
|
| 8 |
+
|
| 9 |
+
- Base model: `meta-llama/Llama-2-70b-hf`
|
| 10 |
+
- Quantization method: SqueezeLLM
|
| 11 |
+
- Target bit-width: 4
|
| 12 |
+
- Backend kernel: Any-Precision-LLM kernel (`ap-gemv`)
|
| 13 |
+
- Calibration data: RedPajama (1024 sentences / 4096 tokens)
|
| 14 |
+
- Calibration objective: Next-token prediction
|
| 15 |
+
|
| 16 |
+
# How to run
|
| 17 |
+
- Follow the instruction in https://github.com/snu-mllab/GuidedQuant.
|
| 18 |
+
|
| 19 |
+
# References
|
| 20 |
+
- [Model Paper](https://arxiv.org/abs/2505.07004)
|