Add files using upload-large-folder tool
Browse filesThis view is limited to 50 files because it contains too many changes. See raw diff
- .gitattributes +4 -0
- geneformer/Geneformer_code_only/config.json +24 -0
- geneformer/Geneformer_code_only/geneformer/gene_median_dictionary_gc104M.pkl +3 -0
- geneformer/Geneformer_code_only/geneformer/gene_name_id_dict_gc104M.pkl +3 -0
- geneformer/Geneformer_code_only/geneformer/token_dictionary_gc104M.pkl +3 -0
- geneformer/model.safetensors +3 -0
- kukanja_finetuned/Kukanja_EAE/checkpoints/uce_freeze_mouse_s3028/test_features.npz +3 -0
- kukanja_finetuned/Kukanja_EAE/checkpoints/uce_full_mouse_s3028/best_model.pth +3 -0
- kukanja_finetuned/Kukanja_EAE/checkpoints/uce_full_mouse_s3028/test_features.npz +3 -0
- kukanja_finetuned/Kukanja_MS/checkpoints/geneformer_freeze_s3028/best_model.pth +3 -0
- kukanja_finetuned/Kukanja_MS/checkpoints/geneformer_freeze_s3028/test_features.npz +3 -0
- kukanja_finetuned/Kukanja_MS/checkpoints/geneformer_full_s3028/test_features.npz +3 -0
- kukanja_finetuned/Kukanja_MS/checkpoints/scgpt_freeze_human_s3028/best_model.pth +3 -0
- kukanja_finetuned/Kukanja_MS/checkpoints/scgpt_freeze_human_s3028/test_features.npz +3 -0
- kukanja_finetuned/Kukanja_MS/checkpoints/scgpt_full_human_s3028/best_model.pth +3 -0
- kukanja_finetuned/Kukanja_MS/checkpoints/scgpt_full_human_s3028/test_features.npz +3 -0
- kukanja_finetuned/Kukanja_MS/checkpoints/uce_freeze_human_s3028/best_model.pth +3 -0
- kukanja_finetuned/Kukanja_MS/checkpoints/uce_full_human_s3028/best_model.pth +3 -0
- kukanja_finetuned/Kukanja_MS/checkpoints/uce_full_human_s3028/test_features.npz +3 -0
- kukanja_finetuned/main_nicheformer.py +105 -0
- nicheformer/gene_name_id_dict_gc104M.pkl +3 -0
- nicheformer/merfish_mean_script.npy +3 -0
- nicheformer/model.h5ad +3 -0
- nicheformer/nicheformer.ckpt +3 -0
- scANVI/SEA_AD_MTP_ST/checkpoints/s3028/scanvi_class/model.pt +3 -0
- scgpt/brain/args.json +43 -0
- scgpt/brain/best_model.pt +3 -0
- scgpt/brain/vocab.json +0 -0
- scgpt/human/args.json +53 -0
- scgpt/human/best_model.pt +3 -0
- scgpt/human/vocab.json +0 -0
- sklearn_baselines/lr/SEA_AD_MTP_ST/checkpoints/class_model.pkl +3 -0
- sklearn_baselines/lr/SEA_AD_MTP_ST/checkpoints/subclass_model.pkl +3 -0
- sklearn_baselines/lr/SEA_AD_MTP_ST/checkpoints/supertype_model.pkl +3 -0
- sklearn_baselines/svm/SEA_AD_MTP_ST/checkpoints/class_model.pkl +3 -0
- sklearn_baselines/svm/SEA_AD_MTP_ST/checkpoints/subclass_model.pkl +3 -0
- sklearn_baselines/svm/SEA_AD_MTP_ST/checkpoints/supertype_model.pkl +3 -0
- sklearn_baselines/xgboost/SEA_AD_MTP_ST/checkpoints/class_model.pkl +3 -0
- sklearn_baselines/xgboost/SEA_AD_MTP_ST/checkpoints/subclass_model.pkl +3 -0
- sklearn_baselines/xgboost/SEA_AD_MTP_ST/checkpoints/supertype_model.pkl +3 -0
- spatial_gnn/SEA_AD_MTP_ST/checkpoints/k15_gcn2_h256_z128_s3028/best_model.pth +3 -0
- spatial_gnn/SEA_AD_MTP_ST/checkpoints/k15_gcn2_h256_z128_s3028/test_features.npz +3 -0
- uce/all_tokens.torch +3 -0
- uce/species_chrom.csv +0 -0
- uce/species_offsets.pkl +3 -0
- uce_33l/all_tokens.torch +3 -0
- uce_33l/species_chrom.csv +0 -0
- uce_33l/species_offsets.pkl +3 -0
- uce_4l/all_tokens.torch +3 -0
- uce_4l/species_chrom.csv +0 -0
.gitattributes
CHANGED
|
@@ -33,3 +33,7 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
| 33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
| 36 |
+
nicheformer/model.h5ad filter=lfs diff=lfs merge=lfs -text
|
| 37 |
+
uce_33l/all_tokens.torch filter=lfs diff=lfs merge=lfs -text
|
| 38 |
+
uce/all_tokens.torch filter=lfs diff=lfs merge=lfs -text
|
| 39 |
+
uce_4l/all_tokens.torch filter=lfs diff=lfs merge=lfs -text
|
geneformer/Geneformer_code_only/config.json
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"BertForMaskedLM"
|
| 4 |
+
],
|
| 5 |
+
"attention_probs_dropout_prob": 0.1,
|
| 6 |
+
"classifier_dropout": null,
|
| 7 |
+
"hidden_act": "relu",
|
| 8 |
+
"hidden_dropout_prob": 0.1,
|
| 9 |
+
"hidden_size": 1152,
|
| 10 |
+
"initializer_range": 0.02,
|
| 11 |
+
"intermediate_size": 4608,
|
| 12 |
+
"layer_norm_eps": 1e-12,
|
| 13 |
+
"max_position_embeddings": 4096,
|
| 14 |
+
"model_type": "bert",
|
| 15 |
+
"num_attention_heads": 18,
|
| 16 |
+
"num_hidden_layers": 18,
|
| 17 |
+
"pad_token_id": 0,
|
| 18 |
+
"position_embedding_type": "absolute",
|
| 19 |
+
"torch_dtype": "float32",
|
| 20 |
+
"transformers_version": "4.44.2",
|
| 21 |
+
"type_vocab_size": 2,
|
| 22 |
+
"use_cache": true,
|
| 23 |
+
"vocab_size": 20275
|
| 24 |
+
}
|
geneformer/Geneformer_code_only/geneformer/gene_median_dictionary_gc104M.pkl
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a51c53f6a771d64508dfaf61529df70e394c53bd20856926117ae5d641a24bf5
|
| 3 |
+
size 1512661
|
geneformer/Geneformer_code_only/geneformer/gene_name_id_dict_gc104M.pkl
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fabfa0c2f49c598c59ae432a32c3499a5908c033756c663b5e0cddf58deea8e1
|
| 3 |
+
size 1660882
|
geneformer/Geneformer_code_only/geneformer/token_dictionary_gc104M.pkl
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:67c445f4385127adfc48dcc072320cd65d6822829bf27dd38070e6e787bc597f
|
| 3 |
+
size 425590
|
geneformer/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:965ceccea81953d362081ef3843560a0e4fef88d396c28017881f1e94b1246f3
|
| 3 |
+
size 1265455076
|
kukanja_finetuned/Kukanja_EAE/checkpoints/uce_freeze_mouse_s3028/test_features.npz
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cb2821e9f57c3759b8e76549dd78b853bc8a63327ae198559dcc4aa00476ce2a
|
| 3 |
+
size 257050497
|
kukanja_finetuned/Kukanja_EAE/checkpoints/uce_full_mouse_s3028/best_model.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7e6bb0b80c70466c29fba115087d4265aaaed6c589ef2ac1582005b21abf5045
|
| 3 |
+
size 3456528659
|
kukanja_finetuned/Kukanja_EAE/checkpoints/uce_full_mouse_s3028/test_features.npz
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:58212765d769baad43d24561bcec1c3bbd5d5bb7d29c94400e91449afcc960be
|
| 3 |
+
size 257118885
|
kukanja_finetuned/Kukanja_MS/checkpoints/geneformer_freeze_s3028/best_model.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:320daea1cfc60ef1c5203fa2b893a644b8813c6a343ccf0a8bb7439f1dc1dc9e
|
| 3 |
+
size 1303103393
|
kukanja_finetuned/Kukanja_MS/checkpoints/geneformer_freeze_s3028/test_features.npz
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1f473e0b245eac4ae0e8c40912a17be25a0510e9617c57a1514efb2b0c08ad08
|
| 3 |
+
size 1064481204
|
kukanja_finetuned/Kukanja_MS/checkpoints/geneformer_full_s3028/test_features.npz
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b2bfd5424d8537ff6891853a51c6180d507c2aa8d04eaa776bf7af99d090512b
|
| 3 |
+
size 1062018138
|
kukanja_finetuned/Kukanja_MS/checkpoints/scgpt_freeze_human_s3028/best_model.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c76000f538a24e3ad91230ddad95fbe5d86915ee60cce3417e487aac7c513d60
|
| 3 |
+
size 214014819
|
kukanja_finetuned/Kukanja_MS/checkpoints/scgpt_freeze_human_s3028/test_features.npz
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fbae090fbd45eff50246fb07a84c010a18abb557ae7dfa2428bf75a9df151a7e
|
| 3 |
+
size 473874197
|
kukanja_finetuned/Kukanja_MS/checkpoints/scgpt_full_human_s3028/best_model.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5f2d856fd16fdb92ed2960fdc7e589b5fec34506cef5da1be13552dc95761d58
|
| 3 |
+
size 214014819
|
kukanja_finetuned/Kukanja_MS/checkpoints/scgpt_full_human_s3028/test_features.npz
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:11b1828e85b3d296e96dbe85d6db553a93808830dae99a9dcd95cfcedb1e66cb
|
| 3 |
+
size 473893370
|
kukanja_finetuned/Kukanja_MS/checkpoints/uce_freeze_human_s3028/best_model.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0883abb1ec60e0bbb37f718e0386e125e6883687298320d2a855ae7401792bc4
|
| 3 |
+
size 3456483283
|
kukanja_finetuned/Kukanja_MS/checkpoints/uce_full_human_s3028/best_model.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9551c7ca5db4d5e41a26816d6f835f67e30f6e37dca50a4739c337d095b6b986
|
| 3 |
+
size 3456483283
|
kukanja_finetuned/Kukanja_MS/checkpoints/uce_full_human_s3028/test_features.npz
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f92e5a430cb367d604c009f38f3702333046a9f5dce7e5090ac37114f990a36d
|
| 3 |
+
size 1183526804
|
kukanja_finetuned/main_nicheformer.py
ADDED
|
@@ -0,0 +1,105 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
#!/usr/bin/env python
|
| 2 |
+
import sys, os
|
| 3 |
+
sys.path.append(os.path.abspath(__file__ + '/../../..'))
|
| 4 |
+
|
| 5 |
+
import random, json
|
| 6 |
+
import numpy as np
|
| 7 |
+
import torch
|
| 8 |
+
import anndata
|
| 9 |
+
|
| 10 |
+
from src.utils.dataloader_kukanja import build_dataloaders_kukanja
|
| 11 |
+
from src.models.fm_nlevel import NicheformerNLevel
|
| 12 |
+
from src.engines.fm_trainer_kukanja import FMTrainerKukanja
|
| 13 |
+
from src.utils.logger import get_logger
|
| 14 |
+
from src.utils.args import get_public_config
|
| 15 |
+
|
| 16 |
+
NICHEFORMER_ASSET_ROOT = (
|
| 17 |
+
"/data1/foundation_model_based_mas/tools_layer/mcp_tools/foundation_model_checkpoints/nicheformer"
|
| 18 |
+
)
|
| 19 |
+
|
| 20 |
+
|
| 21 |
+
def get_args():
|
| 22 |
+
parser = get_public_config()
|
| 23 |
+
parser.set_defaults(model_name="Nicheformer", dropout=0.2)
|
| 24 |
+
parser.add_argument('--npz_path', type=str, required=True)
|
| 25 |
+
parser.add_argument('--h5ad_path', type=str, required=True)
|
| 26 |
+
parser.add_argument('--checkpoint_path', type=str, default=f'{NICHEFORMER_ASSET_ROOT}/nicheformer.ckpt')
|
| 27 |
+
parser.add_argument('--vocab_path', type=str, default=f'{NICHEFORMER_ASSET_ROOT}/model.h5ad')
|
| 28 |
+
parser.add_argument('--merfish_mean_path', type=str, default=f'{NICHEFORMER_ASSET_ROOT}/merfish_mean_script.npy')
|
| 29 |
+
parser.add_argument('--gene_name_to_ens_path', type=str, default=f'{NICHEFORMER_ASSET_ROOT}/gene_name_id_dict_gc104M.pkl')
|
| 30 |
+
parser.add_argument('--freeze_backbone', action='store_true')
|
| 31 |
+
args = parser.parse_args()
|
| 32 |
+
|
| 33 |
+
if 'ms' in os.path.basename(args.npz_path).lower():
|
| 34 |
+
dataset_tag = 'Kukanja_MS'
|
| 35 |
+
elif 'eae' in os.path.basename(args.npz_path).lower():
|
| 36 |
+
dataset_tag = 'Kukanja_EAE'
|
| 37 |
+
else:
|
| 38 |
+
dataset_tag = 'Kukanja'
|
| 39 |
+
|
| 40 |
+
ft_tag = 'freeze' if args.freeze_backbone else 'full'
|
| 41 |
+
ckpt_tag = f"{ft_tag}_s{args.seed}"
|
| 42 |
+
output_root = os.environ.get('KUKANJA_OUT_ROOT', '/tmp/kukanja_runs')
|
| 43 |
+
log_dir = os.path.join(output_root, dataset_tag)
|
| 44 |
+
os.makedirs(log_dir, exist_ok=True)
|
| 45 |
+
logger = get_logger(log_dir, __name__, f'record_nicheformer_{ckpt_tag}.log')
|
| 46 |
+
logger.info(args)
|
| 47 |
+
args.save_dir = os.path.join(log_dir, 'checkpoints', f'nicheformer_{ckpt_tag}')
|
| 48 |
+
args.dataset_tag = dataset_tag
|
| 49 |
+
return args, log_dir, logger
|
| 50 |
+
|
| 51 |
+
|
| 52 |
+
def set_seed(seed):
|
| 53 |
+
random.seed(seed)
|
| 54 |
+
np.random.seed(seed)
|
| 55 |
+
torch.manual_seed(seed)
|
| 56 |
+
if torch.cuda.is_available():
|
| 57 |
+
torch.cuda.manual_seed_all(seed)
|
| 58 |
+
torch.backends.cudnn.deterministic = True
|
| 59 |
+
|
| 60 |
+
|
| 61 |
+
if __name__ == '__main__':
|
| 62 |
+
args, log_dir, logger = get_args()
|
| 63 |
+
set_seed(args.seed)
|
| 64 |
+
device = torch.device(args.device)
|
| 65 |
+
|
| 66 |
+
train_loader, val_loader, test_loader, info = build_dataloaders_kukanja(
|
| 67 |
+
args.npz_path, batch_size=args.bs, seed=args.seed, num_workers=args.num_workers)
|
| 68 |
+
|
| 69 |
+
args.output_num = info['output_num']
|
| 70 |
+
args.label_names = info['label_names']
|
| 71 |
+
logger.info(f"Dataset: {args.dataset_tag}, Levels: {list(zip(info['label_names'], info['output_num']))}")
|
| 72 |
+
|
| 73 |
+
adata = anndata.read_h5ad(args.h5ad_path, backed='r')
|
| 74 |
+
gene_names = list(adata.var_names)
|
| 75 |
+
logger.info(f"Gene panel: {len(gene_names)} genes")
|
| 76 |
+
|
| 77 |
+
model = NicheformerNLevel(
|
| 78 |
+
checkpoint_path=args.checkpoint_path,
|
| 79 |
+
vocab_path=args.vocab_path,
|
| 80 |
+
merfish_mean_path=args.merfish_mean_path,
|
| 81 |
+
gene_name_to_ens_path=args.gene_name_to_ens_path,
|
| 82 |
+
gene_names=gene_names,
|
| 83 |
+
output_num=info['output_num'],
|
| 84 |
+
dropout=args.dropout,
|
| 85 |
+
freeze_backbone=args.freeze_backbone,
|
| 86 |
+
).to(device)
|
| 87 |
+
|
| 88 |
+
n_params = sum(p.numel() for p in model.parameters() if p.requires_grad)
|
| 89 |
+
logger.info(f"Model: Nicheformer N-Level, trainable params: {n_params:,}")
|
| 90 |
+
|
| 91 |
+
optimizer = torch.optim.Adam(
|
| 92 |
+
filter(lambda p: p.requires_grad, model.parameters()),
|
| 93 |
+
lr=args.lr, weight_decay=args.weight_decay)
|
| 94 |
+
scheduler = torch.optim.lr_scheduler.CosineAnnealingLR(
|
| 95 |
+
optimizer, T_max=args.max_epochs, eta_min=1e-6)
|
| 96 |
+
|
| 97 |
+
engine = FMTrainerKukanja(
|
| 98 |
+
model, optimizer, scheduler, device, logger, args,
|
| 99 |
+
label_names=info['label_names'], output_num=info['output_num'])
|
| 100 |
+
|
| 101 |
+
if args.mode == 'train':
|
| 102 |
+
engine.train(train_loader, val_loader)
|
| 103 |
+
engine.test(test_loader)
|
| 104 |
+
else:
|
| 105 |
+
engine.test(test_loader)
|
nicheformer/gene_name_id_dict_gc104M.pkl
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fabfa0c2f49c598c59ae432a32c3499a5908c033756c663b5e0cddf58deea8e1
|
| 3 |
+
size 1660882
|
nicheformer/merfish_mean_script.npy
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0caa0ed9ad668cf6bf61bcfc8f75998d069d828198d4b7d41e3132afc1a2afbb
|
| 3 |
+
size 162608
|
nicheformer/model.h5ad
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a2658657dd9131c905dffe71e5d67c98accfde00f1b4bdf9d26fe3701b516322
|
| 3 |
+
size 1066912
|
nicheformer/nicheformer.ckpt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:67fa5f65b1a17431182d6714fcc4b5d6af86a37b8ac49ea2fb80d95cc46df2d1
|
| 3 |
+
size 572217913
|
scANVI/SEA_AD_MTP_ST/checkpoints/s3028/scanvi_class/model.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c763bc730bf43eec9287b83507c4f45a6aab7183d3ffe91eaab01589a2e68e62
|
| 3 |
+
size 28855255
|
scgpt/brain/args.json
ADDED
|
@@ -0,0 +1,43 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"data_source": "/scratch/ssd004/datasets/cellxgene/scb_strict/brain/all_counts",
|
| 3 |
+
"save_dir": "./save/cellxgene_census_brain-May19-02-00-2023",
|
| 4 |
+
"load_model": null,
|
| 5 |
+
"n_hvg": null,
|
| 6 |
+
"valid_size_or_ratio": 0.03,
|
| 7 |
+
"vocab_path": "/scratch/ssd004/datasets/cellxgene/scFormer/scformer/tokenizer/default_census_vocab.json",
|
| 8 |
+
"pad_token": "<pad>",
|
| 9 |
+
"input_style": "binned",
|
| 10 |
+
"input_emb_style": "continuous",
|
| 11 |
+
"n_bins": 51,
|
| 12 |
+
"max_seq_len": 1200,
|
| 13 |
+
"training_tasks": "both",
|
| 14 |
+
"mask_ratio": [
|
| 15 |
+
0.25,
|
| 16 |
+
0.5,
|
| 17 |
+
0.75
|
| 18 |
+
],
|
| 19 |
+
"trunc_by_sample": true,
|
| 20 |
+
"local_rank": 0,
|
| 21 |
+
"batch_size": 32,
|
| 22 |
+
"eval_batch_size": 64,
|
| 23 |
+
"grad_accu_steps": 1,
|
| 24 |
+
"epochs": 6,
|
| 25 |
+
"lr": 0.0001,
|
| 26 |
+
"scheduler_interval": 100,
|
| 27 |
+
"scheduler_factor": 0.99,
|
| 28 |
+
"warmup_ratio_or_step": 10000.0,
|
| 29 |
+
"no_cls": true,
|
| 30 |
+
"no_cce": true,
|
| 31 |
+
"fp16": true,
|
| 32 |
+
"fast_transformer": true,
|
| 33 |
+
"nlayers": 12,
|
| 34 |
+
"nheads": 8,
|
| 35 |
+
"embsize": 512,
|
| 36 |
+
"d_hid": 512,
|
| 37 |
+
"dropout": 0.2,
|
| 38 |
+
"n_layers_cls": 3,
|
| 39 |
+
"log_interval": 6000,
|
| 40 |
+
"save_interval": 18000,
|
| 41 |
+
"mask_value": -1,
|
| 42 |
+
"pad_value": -2
|
| 43 |
+
}
|
scgpt/brain/best_model.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bf2cf7dce9867f8174ad5186c0692d5a8fed41786ab44f4a3f9b736e066dbff8
|
| 3 |
+
size 205385258
|
scgpt/brain/vocab.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
scgpt/human/args.json
ADDED
|
@@ -0,0 +1,53 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"data_source": "/scratch/ssd004/datasets/cellxgene/scb_strict/human",
|
| 3 |
+
"save_dir": "/scratch/ssd004/datasets/cellxgene/save/cellxgene_census_human-May23-08-36-2023",
|
| 4 |
+
"load_model": null,
|
| 5 |
+
"n_hvg": null,
|
| 6 |
+
"valid_size_or_ratio": 0.003,
|
| 7 |
+
"dist_backend": "nccl",
|
| 8 |
+
"grad_accu_steps": 1,
|
| 9 |
+
"pad_token": "<pad>",
|
| 10 |
+
"input_style": "binned",
|
| 11 |
+
"input_emb_style": "continuous",
|
| 12 |
+
"n_bins": 51,
|
| 13 |
+
"max_seq_len": 1200,
|
| 14 |
+
"training_tasks": "both",
|
| 15 |
+
"dist_url": "tcp://gpu188.cluster.local:53833",
|
| 16 |
+
"mask_ratio": [
|
| 17 |
+
0.25,
|
| 18 |
+
0.5,
|
| 19 |
+
0.75
|
| 20 |
+
],
|
| 21 |
+
"trunc_by_sample": true,
|
| 22 |
+
"vocab_path": "/scratch/ssd004/datasets/cellxgene/scFormer/scformer/tokenizer/default_census_vocab.json",
|
| 23 |
+
"rank": 0,
|
| 24 |
+
"batch_size": 32,
|
| 25 |
+
"eval_batch_size": 64,
|
| 26 |
+
"epochs": 6,
|
| 27 |
+
"lr": 0.0001,
|
| 28 |
+
"scheduler_interval": 100,
|
| 29 |
+
"scheduler_factor": 0.99,
|
| 30 |
+
"warmup_ratio_or_step": 10000.0,
|
| 31 |
+
"no_cls": true,
|
| 32 |
+
"no_cce": true,
|
| 33 |
+
"fp16": true,
|
| 34 |
+
"fast_transformer": true,
|
| 35 |
+
"nlayers": 12,
|
| 36 |
+
"nheads": 8,
|
| 37 |
+
"embsize": 512,
|
| 38 |
+
"d_hid": 512,
|
| 39 |
+
"dropout": 0.2,
|
| 40 |
+
"n_layers_cls": 3,
|
| 41 |
+
"log_interval": 9000,
|
| 42 |
+
"save_interval": 27000,
|
| 43 |
+
"mask_value": -1,
|
| 44 |
+
"pad_value": -2,
|
| 45 |
+
"USE_CLS": false,
|
| 46 |
+
"USE_CCE": false,
|
| 47 |
+
"MVC": true,
|
| 48 |
+
"USE_GENERATIVE_TRAINING": true,
|
| 49 |
+
"world_size": 16,
|
| 50 |
+
"distributed": true,
|
| 51 |
+
"local_rank": 0,
|
| 52 |
+
"gpu": 0
|
| 53 |
+
}
|
scgpt/human/best_model.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6cb5d451ab5c4b33eb673adbe4fddc61d2389df1b89b7651a9fe2e557572b922
|
| 3 |
+
size 205385258
|
scgpt/human/vocab.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
sklearn_baselines/lr/SEA_AD_MTP_ST/checkpoints/class_model.pkl
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:80659a2ea0b764d01cfd59678d5485d7655143feef1d40931e9a8d62b2f550c1
|
| 3 |
+
size 4271
|
sklearn_baselines/lr/SEA_AD_MTP_ST/checkpoints/subclass_model.pkl
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:add516d8915947db8de9bf6f8760236cdf48fd0deb5c267aeca2476a6d9236cc
|
| 3 |
+
size 28119
|
sklearn_baselines/lr/SEA_AD_MTP_ST/checkpoints/supertype_model.pkl
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a7c41064f5b4aa45f036ceaeb1ed67973a19325e55f0e953c55e9951f386b788
|
| 3 |
+
size 156495
|
sklearn_baselines/svm/SEA_AD_MTP_ST/checkpoints/class_model.pkl
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d77be1bbd5c44feffc5fa1cee7ba59fe313a632461a697f9453a615ae852c31b
|
| 3 |
+
size 13243
|
sklearn_baselines/svm/SEA_AD_MTP_ST/checkpoints/subclass_model.pkl
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9118ca0a7bf36f81d816b81ed82bbedb2d7a20e6c6dbc6815921e02f23b9c1da
|
| 3 |
+
size 88995
|
sklearn_baselines/svm/SEA_AD_MTP_ST/checkpoints/supertype_model.pkl
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1d2566b34a24d95116921635ac199641f4e1d4dd0b8eb2180b6ff46dbfc20914
|
| 3 |
+
size 496715
|
sklearn_baselines/xgboost/SEA_AD_MTP_ST/checkpoints/class_model.pkl
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:789b26db578625590ab21cdc4e828e336d105872ac329e15bf60498f9abc0f66
|
| 3 |
+
size 4414218
|
sklearn_baselines/xgboost/SEA_AD_MTP_ST/checkpoints/subclass_model.pkl
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9ffbb1a71bf4dd426d087bcb837b3ebb77dd55736cdc7a3c1d9a2d9c42113ccd
|
| 3 |
+
size 33810521
|
sklearn_baselines/xgboost/SEA_AD_MTP_ST/checkpoints/supertype_model.pkl
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1237acb8d5ec1a23f98b4fcc80c66cb60d39d5b7e775bc01d1e865ca286cb2db
|
| 3 |
+
size 160796181
|
spatial_gnn/SEA_AD_MTP_ST/checkpoints/k15_gcn2_h256_z128_s3028/best_model.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2bb4a92943a55ea1f63229030c7299e940392e7c9f0e3e314e0a02e238a76ee3
|
| 3 |
+
size 1378101
|
spatial_gnn/SEA_AD_MTP_ST/checkpoints/k15_gcn2_h256_z128_s3028/test_features.npz
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3c8779ffb4602613abd6e401b47e016904c366c5d42a77315fdd1a3a3b533aba
|
| 3 |
+
size 196098571
|
uce/all_tokens.torch
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:07397ab3828502fb7d0bab658125c47f145f989f384db8f55fe106f6826b2a54
|
| 3 |
+
size 2979205876
|
uce/species_chrom.csv
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
uce/species_offsets.pkl
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:abda5b2bc4018187e408623b292686a061912f449daceb4c9c9603caf0d62538
|
| 3 |
+
size 139
|
uce_33l/all_tokens.torch
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:07397ab3828502fb7d0bab658125c47f145f989f384db8f55fe106f6826b2a54
|
| 3 |
+
size 2979205876
|
uce_33l/species_chrom.csv
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
uce_33l/species_offsets.pkl
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:abda5b2bc4018187e408623b292686a061912f449daceb4c9c9603caf0d62538
|
| 3 |
+
size 139
|
uce_4l/all_tokens.torch
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:07397ab3828502fb7d0bab658125c47f145f989f384db8f55fe106f6826b2a54
|
| 3 |
+
size 2979205876
|
uce_4l/species_chrom.csv
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|