Update README.md
Browse files
README.md
CHANGED
@@ -78,13 +78,27 @@ inputs = processor(text=prompt, images=[image], return_tensors="pt")
|
|
78 |
inputs = inputs.to(DEVICE)
|
79 |
|
80 |
# Generate outputs
|
81 |
-
generated_ids = model.generate(**inputs, max_new_tokens=
|
82 |
-
|
83 |
-
|
84 |
-
|
85 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
86 |
|
87 |
-
print(generated_texts[0])
|
88 |
```
|
89 |
</details>
|
90 |
|
|
|
78 |
inputs = inputs.to(DEVICE)
|
79 |
|
80 |
# Generate outputs
|
81 |
+
generated_ids = model.generate(**inputs, max_new_tokens=8192)
|
82 |
+
prompt_length = inputs.input_ids.shape[1]
|
83 |
+
trimmed_generated_ids = generated_ids[:, prompt_length:]
|
84 |
+
doctags = processor.batch_decode(
|
85 |
+
trimmed_generated_ids,
|
86 |
+
skip_special_tokens=False,
|
87 |
+
)[0].lstrip()
|
88 |
+
|
89 |
+
# create a docling document
|
90 |
+
doc = DoclingDocument(name="Document")
|
91 |
+
|
92 |
+
# populate it
|
93 |
+
doc.load_from_document_tokens([doctags], [image])
|
94 |
+
|
95 |
+
# export as any format
|
96 |
+
# HTML
|
97 |
+
# print(doc.export_to_html())
|
98 |
+
|
99 |
+
# MD
|
100 |
+
# print(doc.export_to_markdown())
|
101 |
|
|
|
102 |
```
|
103 |
</details>
|
104 |
|