Create README.md
Browse files
README.md
ADDED
@@ -0,0 +1,36 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# PMC Helper Model Checkpoints
|
2 |
+
|
3 |
+
These checkpoints contain helper models used in creating the PMC-2M dataset.
|
4 |
+
|
5 |
+
## Download Weights
|
6 |
+
|
7 |
+
Make sure to run this script before using the pipeline.
|
8 |
+
|
9 |
+
```python
|
10 |
+
from huggingface_hub import snapshot_download
|
11 |
+
|
12 |
+
weights_dir = snapshot_download(
|
13 |
+
repo_id="vector-institute/pmc-helper-models",
|
14 |
+
local_dir="openpmcvl/granular/checkpoints",
|
15 |
+
allow_patterns=["*.pt", "*.pth"]
|
16 |
+
)
|
17 |
+
```
|
18 |
+
|
19 |
+
The weights will be downloaded to the specified `local_dir`. Your existing code can then load them from this location.
|
20 |
+
Change the `local_dir` to the path of the directory leading to `openpmcvl/granular/checkpoints` in your project.
|
21 |
+
|
22 |
+
## Citation
|
23 |
+
TBD
|
24 |
+
|
25 |
+
The original weights are trained and cited from PMC-CLIP paper. Please cite them if using the weights.
|
26 |
+
```
|
27 |
+
@misc{lin2023pmcclipcontrastivelanguageimagepretraining,
|
28 |
+
title={PMC-CLIP: Contrastive Language-Image Pre-training using Biomedical Documents},
|
29 |
+
author={Weixiong Lin and Ziheng Zhao and Xiaoman Zhang and Chaoyi Wu and Ya Zhang and Yanfeng Wang and Weidi Xie},
|
30 |
+
year={2023},
|
31 |
+
eprint={2303.07240},
|
32 |
+
archivePrefix={arXiv},
|
33 |
+
primaryClass={cs.CV},
|
34 |
+
url={https://arxiv.org/abs/2303.07240},
|
35 |
+
}
|
36 |
+
```
|