simiaoAA commited on
Commit
6d37c1d
·
verified ·
1 Parent(s): 42c2edf

Clean model card

Browse files
Files changed (1) hide show
  1. README.md +23 -40
README.md CHANGED
@@ -1,40 +1,23 @@
1
- ---
2
- license: cc-by-4.0
3
- library_name: keras
4
- tags:
5
- - keras
6
- - tensorflow
7
- - single-cell
8
- - aging
9
- - domain-adaptation
10
- - cross-species
11
- pretty_name: Cross-species immune ageing DANN models
12
- ---
13
-
14
- # Cross-species Immune Ageing DANN Models
15
-
16
- This model repository contains trained Keras/TensorFlow weights for cross-species ageing-state annotation in HSCs and CD8+ T cells.
17
-
18
- ## Model Families
19
-
20
- - HSC scRNA DANN classifier and embedder.
21
- - HSC scATAC DANN classifier and embedder.
22
- - CD8+ T-cell scRNA DANN classifier and embedder.
23
- - Exploratory HSC multimodal RNA+ATAC model artifacts.
24
-
25
- ## Expected Inputs
26
-
27
- Inputs must be aligned to the feature files distributed with the dataset repository. See the GitHub repository for scripts:
28
-
29
- - `scripts/evaluate_scrna_dann.py`
30
- - `scripts/search_hsc_dann_saved_models.py`
31
- - `scripts/train_scrna_dann_tf.py`
32
-
33
- ## Limitations
34
-
35
- Adversarial training is checkpoint-sensitive. Saved weights should be used with the exact feature lists, scaling choices and class order described in the GitHub documentation.
36
-
37
- ## Citation
38
-
39
- Please cite the manuscript and the archived model DOI when available.
40
-
 
1
+ ---
2
+ license: mit
3
+ library_name: keras
4
+ tags:
5
+ - keras
6
+ - tensorflow
7
+ - single-cell
8
+ - aging
9
+ - domain-adaptation
10
+ - cross-species
11
+ pretty_name: Cross-species aging DANN models
12
+ ---
13
+
14
+ # Cross-Species Aging DANN Models
15
+
16
+ This model repository stores trained Keras/TensorFlow model files required by
17
+ the GitHub reproducibility package:
18
+
19
+ https://github.com/SimiaoZhao/cross-species-aging-resource
20
+
21
+ Inputs must be aligned to the feature lists distributed with the dataset
22
+ repository. During peer review, access may be restricted to reviewers and
23
+ editors.