avantikalal commited on
Commit
6913420
·
1 Parent(s): c5f5d2c

Upload folder using huggingface_hub (#1)

Browse files

- Upload folder using huggingface_hub (2b92e9863c406e077cc0b751fc6ad7514f78b865)

Files changed (9) hide show
  1. README.md +52 -3
  2. rep0.ckpt +3 -0
  3. rep0.safetensors +3 -0
  4. rep1.ckpt +3 -0
  5. rep1.safetensors +3 -0
  6. rep2.ckpt +3 -0
  7. rep2.safetensors +3 -0
  8. rep3.ckpt +3 -0
  9. rep3.safetensors +3 -0
README.md CHANGED
@@ -1,3 +1,52 @@
1
- ---
2
- license: mit
3
- ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: mit
3
+ library_name: pytorch-lightning
4
+ pipeline_tag: tabular-regression
5
+ tags:
6
+ - biology
7
+ - genomics
8
+ datasets:
9
+ - Genentech/decima-data
10
+ ---
11
+
12
+ # Decima
13
+
14
+ ## Model Description
15
+ Decima is a multi-task regression model designed to predict gene expression from genomic DNA sequences. This model was developed by fine-tuning the **Borzoi** architecture. It maps the genomic DNA sequence to quantitative expression levels across diverse cell types and conditions.
16
+
17
+ For more details, please refer to the original paper: https://www.biorxiv.org/content/10.1101/2024.10.09.617507v3.
18
+
19
+ - **Architecture:** Fine-tuned Borzoi
20
+ - **Task:** Multi-task Regression
21
+ - **Input:** Genomic sequences (hg38)
22
+ - **Output:** Predicted expression values (log(CPM) + 1) for 8,856 pseudobulks.
23
+
24
+ ## Repository Content
25
+ This repository contains four model replicates (`rep0` through `rep3`). Each replicate is provided in two formats:
26
+ 1. **`.ckpt`**: PyTorch Lightning checkpoints containing model weights, optimizer states, and hyperparameters.
27
+ 2. **`.safetensors`**: A lightweight, secure format for weights only.
28
+
29
+ **Files:**
30
+ * `rep0.ckpt`, `rep1.ckpt`, `rep2.ckpt`, `rep3.ckpt`
31
+ * `rep0.safetensors`, `rep1.safetensors`, `rep2.safetensors`, `rep3.safetensors`
32
+
33
+ ## How to Use
34
+ You can load any of the model replicates for inference or further fine-tuning using the `decima` package (https://github.com/Genentech/decima).
35
+
36
+ ### Loading via PyTorch Lightning Checkpoint
37
+ ```python
38
+ from decima.model.lightning import LightningModel
39
+ from huggingface_hub import hf_hub_download
40
+
41
+ # Download a specific replicate (e.g., rep0)
42
+ ckpt_path = hf_hub_download(
43
+ repo_id="Genentech/decima-model",
44
+ filename="rep0.ckpt"
45
+ )
46
+
47
+ # Load the model
48
+ model = LightningModel.load_from_checkpoint(ckpt_path)
49
+ model.eval()
50
+
51
+ # For a safetensor file, use LightningModel.load_safetensor(path)
52
+ ```
rep0.ckpt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9b4efc2967d09d05c34ced1877744ad1d499d3899863463d5107d072046bfb31
3
+ size 2260600894
rep0.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7b4a4ddd20f4c4ee8792459ee2f006576e255adc633c9347b6d3d8378f9173e0
3
+ size 755007760
rep1.ckpt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7f2322061c92f238af320b1fdcd949c345859b51e31788e18d28521a51fca408
3
+ size 2260600894
rep1.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:245bf6f4b4357f39ee15d8e38e32f79e6976ef64936a276d37bd03dbc266a7e3
3
+ size 755007760
rep2.ckpt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:da95328d7a7690edaba49943b26722064a0f4159141c792d1221dc8c9100a9b4
3
+ size 2260600894
rep2.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9200c7640fe6774b3a3c033c50009a4b002f439f6be10515c66bd727ad4328a0
3
+ size 755007760
rep3.ckpt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:31127cc544b048db99bf84b3c859ef764b60be6f844d28195da26ace24cb69c5
3
+ size 2260600894
rep3.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9a390d8be850879d8a533e6826dafa0329f6a4d26c612f280244ce7c19ee1992
3
+ size 755007760