Lal Claude Opus 4.6 commited on
Commit ·
9f2aaba
1
Parent(s): a492bc0
Add performance metrics, training details, fix loading code
Browse files- Fix malformed YAML frontmatter
- Add performance metrics (MSE, Pearson for val/test)
- Add training hyperparameters
- Add parameter count (6.3M)
- Add weights_only=False to loading code
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
README.md
CHANGED
|
@@ -1,5 +1,4 @@
|
|
| 1 |
---
|
| 2 |
-
# 1. Metadata Block
|
| 3 |
license: mit
|
| 4 |
library_name: pytorch-lightning
|
| 5 |
pipeline_tag: tabular-regression
|
|
@@ -16,8 +15,31 @@ datasets:
|
|
| 16 |
This model is a single-task regression model trained to take in 2114 bp genomic intervals and predict the total GM12878 DNase-seq coverage in the central 1000 bp. It is described in Lal et al. 2025 (https://www.nature.com/articles/s41592-025-02868-z).
|
| 17 |
|
| 18 |
- **Architecture:** DilatedConvModel (gReLU)
|
| 19 |
-
- **Input:**
|
| 20 |
-
- **Output:** Total DNase-seq coverage in the central 1000 bp
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 21 |
|
| 22 |
## Repository Content
|
| 23 |
1. `model.ckpt`: The trained model weights and hyperparameters (PyTorch Lightning checkpoint).
|
|
@@ -33,10 +55,10 @@ from grelu.lightning import LightningModel
|
|
| 33 |
from huggingface_hub import hf_hub_download
|
| 34 |
|
| 35 |
ckpt_path = hf_hub_download(
|
| 36 |
-
repo_id="Genentech/GM12878_dnase-model",
|
| 37 |
filename="model.ckpt"
|
| 38 |
)
|
| 39 |
|
| 40 |
-
model = LightningModel.load_from_checkpoint(ckpt_path)
|
| 41 |
model.eval()
|
| 42 |
-
```
|
|
|
|
| 1 |
---
|
|
|
|
| 2 |
license: mit
|
| 3 |
library_name: pytorch-lightning
|
| 4 |
pipeline_tag: tabular-regression
|
|
|
|
| 15 |
This model is a single-task regression model trained to take in 2114 bp genomic intervals and predict the total GM12878 DNase-seq coverage in the central 1000 bp. It is described in Lal et al. 2025 (https://www.nature.com/articles/s41592-025-02868-z).
|
| 16 |
|
| 17 |
- **Architecture:** DilatedConvModel (gReLU)
|
| 18 |
+
- **Input:** 2,114 bp genomic sequences (hg38)
|
| 19 |
+
- **Output:** Total DNase-seq coverage in the central 1000 bp
|
| 20 |
+
- **Parameters:** 6.3M
|
| 21 |
+
|
| 22 |
+
## Performance
|
| 23 |
+
|
| 24 |
+
| Split | MSE | Pearson |
|
| 25 |
+
|-------|-----|---------|
|
| 26 |
+
| Validation | 0.4458 | 0.7524 |
|
| 27 |
+
| Test | 0.4113 | 0.8056 |
|
| 28 |
+
|
| 29 |
+
## Training Details
|
| 30 |
+
|
| 31 |
+
| Parameter | Value |
|
| 32 |
+
|-----------|-------|
|
| 33 |
+
| Task | Regression |
|
| 34 |
+
| Loss | MSE |
|
| 35 |
+
| Optimizer | Adam |
|
| 36 |
+
| Learning rate | 0.0001 |
|
| 37 |
+
| Batch size | 512 |
|
| 38 |
+
| Max epochs | 15 |
|
| 39 |
+
| Channels | 512 |
|
| 40 |
+
| n_conv | 9 |
|
| 41 |
+
| crop_len | 557 |
|
| 42 |
+
| grelu version | 1.0.4.post1.dev39 |
|
| 43 |
|
| 44 |
## Repository Content
|
| 45 |
1. `model.ckpt`: The trained model weights and hyperparameters (PyTorch Lightning checkpoint).
|
|
|
|
| 55 |
from huggingface_hub import hf_hub_download
|
| 56 |
|
| 57 |
ckpt_path = hf_hub_download(
|
| 58 |
+
repo_id="Genentech/GM12878_dnase-model",
|
| 59 |
filename="model.ckpt"
|
| 60 |
)
|
| 61 |
|
| 62 |
+
model = LightningModel.load_from_checkpoint(ckpt_path, weights_only=False)
|
| 63 |
model.eval()
|
| 64 |
+
```
|