Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
## Usage
|
| 2 |
|
| 3 |
```python
|
|
@@ -28,6 +32,16 @@ Getting into a good university is a complex process that involves factors such a
|
|
| 28 |
|
| 29 |
## Recipe
|
| 30 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 31 |
```
|
| 32 |
test_stage:
|
| 33 |
obcq_modifiers:
|
|
|
|
| 1 |
+
---
|
| 2 |
+
tags:
|
| 3 |
+
- deepsparse
|
| 4 |
+
---
|
| 5 |
## Usage
|
| 6 |
|
| 7 |
```python
|
|
|
|
| 32 |
|
| 33 |
## Recipe
|
| 34 |
|
| 35 |
+
```
|
| 36 |
+
git clone https://github.com/neuralmagic/sparseml
|
| 37 |
+
pip install -e "sparseml[transformers]" "torch<2"
|
| 38 |
+
python sparseml/src/sparseml/transformers/sparsification/obcq/obcq.py PY007/TinyLlama-1.1B-Chat-v0.3 open_platypus --recipe recipe.yaml --save True
|
| 39 |
+
python sparseml/src/sparseml/transformers/sparsification/obcq/export.py --task text-generation --model_path obcq_deployment --sequence_length 512
|
| 40 |
+
cp deployment/model.onnx deployment/model-orig.onnx
|
| 41 |
+
python onnx_kv_inject.py --input-file deployment/model-orig.onnx --output-file deployment/model.onnx
|
| 42 |
+
```
|
| 43 |
+
|
| 44 |
+
`recipe.yaml`
|
| 45 |
```
|
| 46 |
test_stage:
|
| 47 |
obcq_modifiers:
|