Commit
·
d507fa8
1
Parent(s):
6700a58
up
Browse files
.DS_Store
ADDED
Binary file (6.15 kB). View file
|
|
README.md
CHANGED
@@ -2,6 +2,32 @@
|
|
2 |
license: mit
|
3 |
---
|
4 |
|
|
|
5 |
|
6 |
-
This is the LoRA
|
7 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2 |
license: mit
|
3 |
---
|
4 |
|
5 |
+
# GraphJudger
|
6 |
|
7 |
+
This is the LoRA weights for the paper [Can LLMs be Good Graph Judger for Knowledge Graph Construction?](https://arxiv.org/abs/2411.17388).
|
8 |
|
9 |
+
## Example
|
10 |
+
|
11 |
+
Loading:
|
12 |
+
|
13 |
+
```python
|
14 |
+
BASE_MODEL = "models/llama-2-7b-hf"
|
15 |
+
LORA_WEIGHTS = "models/llama2-7b-lora-genwiki/"
|
16 |
+
|
17 |
+
model = LlamaForCausalLM.from_pretrained(
|
18 |
+
BASE_MODEL,
|
19 |
+
load_in_8bit=False
|
20 |
+
).half().cuda()
|
21 |
+
pipeline = transformers.pipeline (
|
22 |
+
"text-generation",
|
23 |
+
model=model,
|
24 |
+
tokenizer=tokenizer,
|
25 |
+
torch_dtype=torch.float16,
|
26 |
+
device=device
|
27 |
+
)
|
28 |
+
pipeline.model = PeftModel.from_pretrained(
|
29 |
+
model,
|
30 |
+
LORA_WEIGHTS,
|
31 |
+
torch_dtype=torch.float16,
|
32 |
+
).half().cuda()
|
33 |
+
```
|