Skip to content

Commit 4d275ec

Browse files
MingzeDongMingzeDong
committed
update configs
1 parent 36ebbe7 commit 4d275ec

21 files changed

Lines changed: 6940 additions & 303 deletions

‎.ipynb_checkpoints/LICENSE-checkpoint‎

Lines changed: 437 additions & 0 deletions
Large diffs are not rendered by default.

‎.ipynb_checkpoints/pyproject-checkpoint.toml‎

Lines changed: 46 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,46 @@
1+
[build-system]
2+
requires = ["setuptools>=68", "wheel"]
3+
build-backend = "setuptools.build_meta"
4+
5+
[project]
6+
name = "arc-stack"
7+
version = "0.1.1"
8+
description = "Stack is a single-cell foundation model that enables in-context learning at inference time."
9+
readme = "README.md"
10+
license = { file = "LICENSE" }
11+
authors = [
12+
{ name = "Mingze Dong", email = "mingze.dong@arcinstitute.org" },
13+
{ name = "Yusuf Roohani", email = "yusuf.roohani@arcinstitute.org" },
14+
]
15+
requires-python = ">=3.9"
16+
dependencies = [
17+
"anndata>=0.9",
18+
"h5py>=3.8",
19+
"geomloss>=0.2.6",
20+
"numpy>=1.22",
21+
"pandas>=1.5",
22+
"psutil>=5.9",
23+
"pytorch-lightning>=2.1",
24+
"PyYAML>=6.0",
25+
"scipy>=1.8",
26+
"scvi-tools>=1.0",
27+
"torch>=2.0",
28+
"wandb>=0.15",
29+
]
30+
31+
[project.urls]
32+
Homepage = "https://github.com/ArcInstitute/stack"
33+
Source = "https://github.com/ArcInstitute/stack"
34+
35+
[project.scripts]
36+
stack-train = "stack.cli.launch_training:main"
37+
stack-finetune = "stack.cli.launch_finetuning:main"
38+
stack-embedding = "stack.cli.embedding:main"
39+
stack-generation = "stack.cli.generation:main"
40+
41+
[tool.setuptools]
42+
package-dir = {"" = "src"}
43+
44+
[tool.setuptools.packages.find]
45+
where = ["src"]
46+
include = ["stack*"]

‎.ipynb_checkpoints/setup-checkpoint.cfg‎

Lines changed: 41 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,41 @@
1+
[metadata]
2+
name = arc-stack
3+
version = 0.1.1
4+
description = Training and fine-tuning utilities for the Stack model
5+
long_description = file: README.md
6+
long_description_content_type = text/markdown
7+
license = Apache-2.0
8+
license_files = LICENSE
9+
10+
[options]
11+
packages = find:
12+
package_dir =
13+
= src
14+
python_requires = >=3.9
15+
include_package_data = True
16+
install_requires =
17+
anndata>=0.9
18+
h5py>=3.8
19+
geomloss>=0.2.6
20+
numpy>=1.22
21+
pandas>=1.5
22+
psutil>=5.9
23+
pytorch-lightning>=2.1
24+
PyYAML>=6.0
25+
scipy>=1.8
26+
scvi-tools>=1.0
27+
torch>=2.0
28+
wandb>=0.15
29+
30+
[options.entry_points]
31+
console_scripts =
32+
stack-train = stack.cli.launch_training:main
33+
stack-finetune = stack.cli.launch_finetuning:main
34+
stack-embedding = stack.cli.embedding:main
35+
stack-generation = stack.cli.generation:main
36+
37+
[options.packages.find]
38+
where =
39+
src
40+
include =
41+
stack*

‎.ipynb_checkpoints/setup-checkpoint.py‎

Lines changed: 4 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,4 @@
1+
from setuptools import setup
2+
3+
if __name__ == "__main__":
4+
setup()

‎.ipynb_checkpoints/uv-checkpoint.lock‎

Lines changed: 2854 additions & 0 deletions
Large diffs are not rendered by default.

‎configs/finetuning/.ipynb_checkpoints/ft_parsecg-checkpoint.yaml‎

Lines changed: 20 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,20 @@
1+
checkpoint_path: "Stack-Large/bc_large.ckpt"
2+
dataset_configs:
3+
- "human:parse_bio:sample:cell_type:false"
4+
- "human:Stack-CellxGene45M:donor_id:author_cell_type:true:feature_name"
5+
genelist_path: "Stack-Large-Aligned/basecount_1000per_15000max.pkl"
6+
save_dir: "stack_checkpoints/bc_large_aligned"
7+
max_epochs: 8
8+
sample_size: 512
9+
batch_size: 8
10+
accumulate_grad_batches: 4
11+
num_workers: 4
12+
learning_rate: 0.00002
13+
weight_decay: 0.003
14+
scheduler_eta_min: 0.000005
15+
replacement_ratio: 0.75
16+
scheduler: cosine
17+
scheduler_T_max: 8
18+
scheduler_warmup_epochs: 1
19+
early_stopping_patience: 5
20+
cache_file: "cache/parse_cgft.pkl"

‎configs/finetuning/ft_parsecg.yaml‎

Lines changed: 2 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -1,7 +1,7 @@
11
checkpoint_path: "Stack-Large/bc_large.ckpt"
22
dataset_configs:
33
- "human:parse_bio:sample:cell_type:false"
4-
- "human:Stack-CellxGene30M:donor_id:author_cell_type:true:feature_name"
4+
- "human:Stack-CellxGene45M:donor_id:author_cell_type:true:feature_name"
55
genelist_path: "Stack-Large-Aligned/basecount_1000per_15000max.pkl"
66
save_dir: "stack_checkpoints/bc_large_aligned"
77
max_epochs: 8
@@ -11,6 +11,7 @@ accumulate_grad_batches: 4
1111
num_workers: 4
1212
learning_rate: 0.00002
1313
weight_decay: 0.003
14+
scheduler_eta_min: 0.000005
1415
replacement_ratio: 0.75
1516
scheduler: cosine
1617
scheduler_T_max: 8

‎notebooks/.ipynb_checkpoints/tutorial-predict-checkpoint.ipynb‎

Lines changed: 1084 additions & 0 deletions
Large diffs are not rendered by default.

‎notebooks/tutorial-predict.ipynb‎

Lines changed: 399 additions & 277 deletions
Large diffs are not rendered by default.

‎pyproject.toml‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
44

55
[project]
66
name = "arc-stack"
7-
version = "0.1.0"
7+
version = "0.1.1"
88
description = "Stack is a single-cell foundation model that enables in-context learning at inference time."
99
readme = "README.md"
1010
license = { file = "LICENSE" }

0 commit comments

Comments
 (0)