Upload README.md with huggingface_hub
Browse files
README.md
ADDED
|
@@ -0,0 +1,23 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: cell-fm
|
| 3 |
+
tags:
|
| 4 |
+
- biology
|
| 5 |
+
- microscopy
|
| 6 |
+
- protein
|
| 7 |
+
- condensate
|
| 8 |
+
- flow-matching
|
| 9 |
+
---
|
| 10 |
+
|
| 11 |
+
# CELL-FM CondenSeq weights
|
| 12 |
+
|
| 13 |
+
Checkpoints behind the [CELL-FM CondenSeq demo](https://huggingface.co/spaces/dihan-zheng/CELL-FM-condenseq-demo).
|
| 14 |
+
|
| 15 |
+
| File | Model | Source checkpoint |
|
| 16 |
+
|---|---|---|
|
| 17 |
+
| `cellfm_seq2img.bin` | CELL-FM CS sequence-to-image generator (includes the ESM-C 600M encoder) | `pretrain_condenseq/cellfm_seq2img/checkpoint-50000` |
|
| 18 |
+
| `vae.bin` | Image VAE, 160x160, 3 down blocks, 4 latent channels | `pretrain_condenseq/vae/checkpoint-50000` |
|
| 19 |
+
| `vit_cls.bin` | ViT condensed/diffuse classifier, 2-channel 160x160 input | `PT_CondenSeq_img_ViT_cls_R1/checkpoint-10000` |
|
| 20 |
+
|
| 21 |
+
Hyperparameters are set in `pipeline.py` in the Space and mirror
|
| 22 |
+
`scripts/cell_fm_cs/evaluate_seq2img.sh` and
|
| 23 |
+
`scripts/vit_cls_condenseq_img/pretrain.sh` in the CELL-FM repository.
|