Update README.md
Browse files
README.md
CHANGED
@@ -22,11 +22,14 @@ Fine tunes a cross encoder on the Amazon ESCI dataset.
|
|
22 |
|
23 |
# Usage
|
24 |
|
|
|
|
|
25 |
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
|
26 |
|
27 |
|
28 |
```python
|
29 |
from transformers import AutoTokenizer, AutoModelForSequenceClassification
|
|
|
30 |
|
31 |
model_name = "lv12/esci-ms-marco-MiniLM-L-12-v2"
|
32 |
|
@@ -56,7 +59,38 @@ inputs = tokenizer(
|
|
56 |
truncation=True,
|
57 |
return_tensors="pt",
|
58 |
)
|
59 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
60 |
print(scores)
|
61 |
```
|
62 |
|
|
|
22 |
|
23 |
# Usage
|
24 |
|
25 |
+
## Transformers
|
26 |
+
|
27 |
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
|
28 |
|
29 |
|
30 |
```python
|
31 |
from transformers import AutoTokenizer, AutoModelForSequenceClassification
|
32 |
+
from torch import no_grad
|
33 |
|
34 |
model_name = "lv12/esci-ms-marco-MiniLM-L-12-v2"
|
35 |
|
|
|
59 |
truncation=True,
|
60 |
return_tensors="pt",
|
61 |
)
|
62 |
+
|
63 |
+
model.eval()
|
64 |
+
with no_grad():
|
65 |
+
scores = model(**inputs).logits.cpu().detach().numpy()
|
66 |
+
print(scores)
|
67 |
+
```
|
68 |
+
|
69 |
+
### Sentence Transformers
|
70 |
+
|
71 |
+
```python
|
72 |
+
from sentence_transformers import CrossEncoder
|
73 |
+
|
74 |
+
model_name = "lv12/esci-ms-marco-MiniLM-L-12-v2"
|
75 |
+
|
76 |
+
queries = [
|
77 |
+
"adidas shoes",
|
78 |
+
"adidas sambas",
|
79 |
+
"girls sandals",
|
80 |
+
"backpacks",
|
81 |
+
"shoes",
|
82 |
+
"mustard blouse"
|
83 |
+
]
|
84 |
+
documents = [
|
85 |
+
"Nike Air Max, with air cushion",
|
86 |
+
"Adidas Ultraboost, the best boost you can get",
|
87 |
+
"Women's sandals wide width 9",
|
88 |
+
"Girl's surf backpack",
|
89 |
+
"Fresh watermelon, all you can eat",
|
90 |
+
"Floral yellow dress with frills and lace"
|
91 |
+
]
|
92 |
+
model = CrossEncoder(model_name, max_length=512)
|
93 |
+
scores = model.predict([(q, d) for q, d in zip(queries, documents)])
|
94 |
print(scores)
|
95 |
```
|
96 |
|