sxtforreal
commited on
Update README.md
Browse files
README.md
CHANGED
@@ -5,4 +5,28 @@ language:
|
|
5 |
tags:
|
6 |
- NLP
|
7 |
pipeline_tag: feature-extraction
|
8 |
-
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
5 |
tags:
|
6 |
- NLP
|
7 |
pipeline_tag: feature-extraction
|
8 |
+
---
|
9 |
+
|
10 |
+
# Usage
|
11 |
+
from transformers import AutoTokenizer
|
12 |
+
from model import (
|
13 |
+
BERTContrastiveLearning_simcse,
|
14 |
+
BERTContrastiveLearning_simcse_w,
|
15 |
+
BERTContrastiveLearning_samp,
|
16 |
+
BERTContrastiveLearning_samp_w,
|
17 |
+
)
|
18 |
+
|
19 |
+
str_list = data["string"].tolist() # Your list of strings here
|
20 |
+
tokenizer = AutoTokenizer.from_pretrained("emilyalsentzer/Bio_ClinicalBERT")
|
21 |
+
tokenized_inputs = tokenizer(
|
22 |
+
str_list, padding=True, max_length=50, truncation=True, return_tensors="pt"
|
23 |
+
)
|
24 |
+
input_ids = tokenized_inputs["input_ids"]
|
25 |
+
attention_mask = tokenized_inputs["attention_mask"]
|
26 |
+
|
27 |
+
model1 = BERTContrastiveLearning_simcse.load_from_checkpoint(ckpt1).eval()
|
28 |
+
model2 = BERTContrastiveLearning_simcse_w.load_from_checkpoint(ckpt2).eval()
|
29 |
+
model3 = BERTContrastiveLearning_samp.load_from_checkpoint(ckpt3).eval()
|
30 |
+
model4 = BERTContrastiveLearning_samp_w.load_from_checkpoint(ckpt4).eval()
|
31 |
+
|
32 |
+
cls, _ = model(input_ids, attention_mask) # embeddings
|