Upload README.md with huggingface_hub
Browse files
README.md
ADDED
|
@@ -0,0 +1,46 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
title: Autoencoder TRACERx-focused 64D
|
| 3 |
+
emoji: 🧬
|
| 4 |
+
colorFrom: blue
|
| 5 |
+
colorTo: green
|
| 6 |
+
sdk: pytorch
|
| 7 |
+
tags:
|
| 8 |
+
- transcriptomics
|
| 9 |
+
- dimensionality-reduction
|
| 10 |
+
- ae
|
| 11 |
+
- tracerx
|
| 12 |
+
license: mit
|
| 13 |
+
---
|
| 14 |
+
|
| 15 |
+
# Autoencoder (TRACERx-focused, 64D)
|
| 16 |
+
|
| 17 |
+
This model is part of the TRACERx Datathon 2025 transcriptomics analysis pipeline.
|
| 18 |
+
|
| 19 |
+
## Model Details
|
| 20 |
+
|
| 21 |
+
- **Model Type**: Autoencoder
|
| 22 |
+
- **Dataset**: TRACERx-focused
|
| 23 |
+
- **Latent Dimensions**: 64
|
| 24 |
+
- **Compression Mode**: transcriptome
|
| 25 |
+
- **Framework**: PyTorch
|
| 26 |
+
|
| 27 |
+
## Usage
|
| 28 |
+
|
| 29 |
+
This model is designed to be used with the TRACERx Datathon 2025 analysis pipeline.
|
| 30 |
+
It will be automatically downloaded and cached when needed.
|
| 31 |
+
|
| 32 |
+
## Model Architecture
|
| 33 |
+
|
| 34 |
+
- Input: Gene expression data
|
| 35 |
+
- Hidden layers: [input_size, 512, 256, 128, 64]
|
| 36 |
+
- Output: 64-dimensional latent representation
|
| 37 |
+
- Activation: ELU with batch normalization
|
| 38 |
+
|
| 39 |
+
## Training Data
|
| 40 |
+
|
| 41 |
+
Trained exclusively on TRACERx open dataset
|
| 42 |
+
|
| 43 |
+
## Files
|
| 44 |
+
|
| 45 |
+
- `autoencoder_64_latent_dims_oos_mode.pt`: Main model weights
|
| 46 |
+
- `latent_df.csv`: Example latent representations (if available)
|