VeloVAE benchmark on dyngen data#
Notebook benchmarks velocity and latent time inference using VeloVAE on dyngen-generated data.
Library imports#
import velovae as vv
import numpy as np
import pandas as pd
import torch
import anndata as ad
import scvelo as scv
from rgv_tools import DATA_DIR
from rgv_tools.benchmarking import get_velocity_correlation
2024-12-14 02:31:21.846239: I tensorflow/core/util/port.cc:153] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.
2024-12-14 02:31:26.831024: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:477] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered
WARNING: All log messages before absl::InitializeLog() is called are written to STDERR
E0000 00:00:1734139887.884843 3085998 cuda_dnn.cc:8310] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
E0000 00:00:1734139888.345119 3085998 cuda_blas.cc:1418] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
2024-12-14 02:31:29.908273: I tensorflow/core/platform/cpu_feature_guard.cc:210] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.
To enable the following instructions: AVX2 AVX512F AVX512_VNNI FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.
/home/icb/weixu.wang/miniconda3/envs/regvelo_test/lib/python3.10/site-packages/anndata/utils.py:429: FutureWarning: Importing read_csv from `anndata` is deprecated. Import anndata.io.read_csv instead.
warnings.warn(msg, FutureWarning)
/home/icb/weixu.wang/miniconda3/envs/regvelo_test/lib/python3.10/site-packages/anndata/utils.py:429: FutureWarning: Importing read_excel from `anndata` is deprecated. Import anndata.io.read_excel instead.
warnings.warn(msg, FutureWarning)
/home/icb/weixu.wang/miniconda3/envs/regvelo_test/lib/python3.10/site-packages/anndata/utils.py:429: FutureWarning: Importing read_hdf from `anndata` is deprecated. Import anndata.io.read_hdf instead.
warnings.warn(msg, FutureWarning)
/home/icb/weixu.wang/miniconda3/envs/regvelo_test/lib/python3.10/site-packages/anndata/utils.py:429: FutureWarning: Importing read_loom from `anndata` is deprecated. Import anndata.io.read_loom instead.
warnings.warn(msg, FutureWarning)
/home/icb/weixu.wang/miniconda3/envs/regvelo_test/lib/python3.10/site-packages/anndata/utils.py:429: FutureWarning: Importing read_mtx from `anndata` is deprecated. Import anndata.io.read_mtx instead.
warnings.warn(msg, FutureWarning)
/home/icb/weixu.wang/miniconda3/envs/regvelo_test/lib/python3.10/site-packages/anndata/utils.py:429: FutureWarning: Importing read_text from `anndata` is deprecated. Import anndata.io.read_text instead.
warnings.warn(msg, FutureWarning)
/home/icb/weixu.wang/miniconda3/envs/regvelo_test/lib/python3.10/site-packages/anndata/utils.py:429: FutureWarning: Importing read_umi_tools from `anndata` is deprecated. Import anndata.io.read_umi_tools instead.
warnings.warn(msg, FutureWarning)
General settings#
scv.settings.verbosity = 3
Constants#
torch.manual_seed(0)
np.random.seed(0)
DATASET = "dyngen"
SAVE_DATA = True
if SAVE_DATA:
(DATA_DIR / DATASET / "results").mkdir(parents=True, exist_ok=True)
(DATA_DIR / DATASET / "processed" / "velovae_vae").mkdir(parents=True, exist_ok=True)
Velocity pipeline#
velocity_correlation = []
for filename in (DATA_DIR / DATASET / "processed").iterdir():
if filename.suffix != ".zarr":
continue
adata = ad.io.read_zarr(filename)
try:
vae = vv.VAE(adata, tmax=20, dim_z=5, device="cuda:0")
config = {}
vae.train(adata, config=config, plot=False, embed="pca")
# Output velocity to adata object
vae.save_anndata(adata, "vae", DATA_DIR / DATASET / "processed" / "velovae_vae", file_name="velovae.h5ad")
adata.layers["velocity"] = adata.layers["vae_velocity"].copy()
velocity_correlation.append(
get_velocity_correlation(
ground_truth=adata.layers["true_velocity"], estimated=adata.layers["velocity"], aggregation=np.mean
)
)
except Exception as e: # noqa: BLE001
# Append np.nan in case of an error and optionally log the error
print(f"An error occurred: {e}")
velocity_correlation.append(np.nan)
Estimating ODE parameters...
Detected 59 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
(0.42, 0.2890740448785467), (0.58, 0.7827494240291312)
KS-test result: [1. 0. 1.]
Initial induction: 84, repression: 21/105
Learning Rate based on Data Sparsity: 0.0001
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 6, test iteration: 10
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.035
Average Set Size: 20
********* Round 1: Early Stop Triggered at epoch 1197. *********
Change in noise variance: 0.0522
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 1284. *********
Change in noise variance: 0.0202
Change in x0: 0.2577
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 1558. *********
Change in noise variance: 0.0021
Change in x0: 0.2075
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 1659. *********
Change in noise variance: 0.0022
Change in x0: 0.1758
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 1696. *********
Change in noise variance: 0.0014
Change in x0: 0.1505
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 1848. *********
Change in noise variance: 0.0005
Change in x0: 0.0920
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 1897. *********
Change in noise variance: 0.0000
Change in x0: 0.0793
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 1904. *********
Change in noise variance: 0.0000
Change in x0: 0.0672
********* Velocity Refinement Round 9 *********
********* Round 9: Early Stop Triggered at epoch 1985. *********
Change in noise variance: 0.0000
Change in x0: 0.0616
********* Velocity Refinement Round 10 *********
Stage 2: Early Stop Triggered at round 9.
********* Finished. Total Time = 0 h : 3 m : 28 s *********
Final: Train ELBO = 119.917, Test ELBO = 117.292
Estimating ODE parameters...
Detected 25 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
KS-test result: [1. 2. 1.]
Initial induction: 35, repression: 24/59
Learning Rate based on Data Sparsity: 0.0000
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 3, test iteration: 4
********* Stage 1: Early Stop Triggered at epoch 9. *********
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.028
Average Set Size: 18
********* Round 1: Early Stop Triggered at epoch 336. *********
Change in noise variance: 0.2841
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 358. *********
Change in noise variance: 0.0281
Change in x0: 0.3996
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 372. *********
Change in noise variance: 0.0069
Change in x0: 0.3231
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 385. *********
Change in noise variance: 0.0288
Change in x0: 0.2682
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 396. *********
Change in noise variance: 0.0067
Change in x0: 0.1960
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 409. *********
Change in noise variance: 0.0044
Change in x0: 0.1752
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 422. *********
Change in noise variance: 0.0026
Change in x0: 0.1730
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 435. *********
Change in noise variance: 0.0023
Change in x0: 0.1750
********* Velocity Refinement Round 9 *********
********* Round 9: Early Stop Triggered at epoch 446. *********
Change in noise variance: 0.0018
Change in x0: 0.1787
********* Velocity Refinement Round 10 *********
********* Round 10: Early Stop Triggered at epoch 471. *********
Change in noise variance: 0.0013
Change in x0: 0.1775
********* Velocity Refinement Round 11 *********
********* Round 11: Early Stop Triggered at epoch 488. *********
Change in noise variance: 0.0008
Change in x0: 0.1746
********* Velocity Refinement Round 12 *********
Stage 2: Early Stop Triggered at round 11.
********* Finished. Total Time = 0 h : 0 m : 33 s *********
Final: Train ELBO = -17687.199, Test ELBO = -17872.184
Estimating ODE parameters...
Detected 37 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
(0.39, 0.1200757575748825), (0.61, 0.7766228691348986)
KS-test result: [0. 1. 1.]
Initial induction: 73, repression: 13/86
Learning Rate based on Data Sparsity: 0.0000
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 2, test iteration: 2
********* Stage 1: Early Stop Triggered at epoch 11. *********
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.027
Average Set Size: 16
********* Round 1: Early Stop Triggered at epoch 17. *********
Change in noise variance: 0.3412
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 22. *********
Change in noise variance: 0.0173
Change in x0: 0.3557
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 28. *********
Change in noise variance: 0.0057
Change in x0: 0.1894
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 34. *********
Change in noise variance: 0.0012
Change in x0: 0.1652
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 40. *********
Change in noise variance: 0.0009
Change in x0: 0.1457
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 46. *********
Change in noise variance: 0.0000
Change in x0: 0.1203
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 52. *********
Change in noise variance: 0.0000
Change in x0: 0.1090
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 58. *********
Change in noise variance: 0.0000
Change in x0: 0.1062
********* Velocity Refinement Round 9 *********
Stage 2: Early Stop Triggered at round 8.
********* Finished. Total Time = 0 h : 0 m : 18 s *********
Final: Train ELBO = -17602.459, Test ELBO = -17961.113
Estimating ODE parameters...
Detected 35 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
KS-test result: [1. 1. 1.]
Assign cluster 0 to repressive
Initial induction: 50, repression: 31/81
Learning Rate based on Data Sparsity: 0.0000
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 2, test iteration: 2
********* Stage 1: Early Stop Triggered at epoch 13. *********
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.027
Average Set Size: 17
********* Round 1: Early Stop Triggered at epoch 19. *********
Change in noise variance: 0.2821
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 55. *********
Change in noise variance: 0.4924
Change in x0: 1.1602
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 70. *********
Change in noise variance: 0.0222
Change in x0: 0.6248
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 76. *********
Change in noise variance: 0.0062
Change in x0: 0.4080
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 82. *********
Change in noise variance: 0.0023
Change in x0: 0.2941
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 93. *********
Change in noise variance: 0.0011
Change in x0: 0.2334
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 99. *********
Change in noise variance: 0.0006
Change in x0: 0.1944
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 107. *********
Change in noise variance: 0.0000
Change in x0: 0.1640
********* Velocity Refinement Round 9 *********
********* Round 9: Early Stop Triggered at epoch 118. *********
Change in noise variance: 0.0000
Change in x0: 0.1349
********* Velocity Refinement Round 10 *********
********* Round 10: Early Stop Triggered at epoch 123. *********
Change in noise variance: 0.0000
Change in x0: 0.1116
********* Velocity Refinement Round 11 *********
********* Round 11: Early Stop Triggered at epoch 128. *********
Change in noise variance: 0.0000
Change in x0: 0.0961
********* Velocity Refinement Round 12 *********
********* Round 12: Early Stop Triggered at epoch 140. *********
Change in noise variance: 0.0000
Change in x0: 0.0875
********* Velocity Refinement Round 13 *********
Stage 2: Early Stop Triggered at round 12.
********* Finished. Total Time = 0 h : 0 m : 5 s *********
Final: Train ELBO = -18108.508, Test ELBO = -17719.186
Estimating ODE parameters...
Detected 35 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
KS-test result: [1. 1. 1.]
Assign cluster 0 to repressive
Initial induction: 51, repression: 29/80
Learning Rate based on Data Sparsity: 0.0000
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 2, test iteration: 2
********* Stage 1: Early Stop Triggered at epoch 822. *********
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.027
Average Set Size: 18
Change in noise variance: 0.1961
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 1402. *********
Change in noise variance: 0.2211
Change in x0: 1.2181
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 1500. *********
Change in noise variance: 0.0065
Change in x0: 0.3970
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 1602. *********
Change in noise variance: 0.0040
Change in x0: 0.3289
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 1653. *********
Change in noise variance: 0.0048
Change in x0: 0.3318
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 1686. *********
Change in noise variance: 0.0071
Change in x0: 0.3488
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 1858. *********
Change in noise variance: 0.0036
Change in x0: 0.3185
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 1885. *********
Change in noise variance: 0.0053
Change in x0: 0.2814
********* Velocity Refinement Round 9 *********
********* Round 9: Early Stop Triggered at epoch 1968. *********
Change in noise variance: 0.0069
Change in x0: 0.1984
********* Velocity Refinement Round 10 *********
********* Round 10: Early Stop Triggered at epoch 2002. *********
Change in noise variance: 0.0054
Change in x0: 0.1864
********* Velocity Refinement Round 11 *********
********* Round 11: Early Stop Triggered at epoch 2029. *********
Change in noise variance: 0.0042
Change in x0: 0.1192
********* Velocity Refinement Round 12 *********
********* Round 12: Early Stop Triggered at epoch 2167. *********
Change in noise variance: 0.0040
Change in x0: 0.0985
********* Velocity Refinement Round 13 *********
********* Round 13: Early Stop Triggered at epoch 2299. *********
Change in noise variance: 0.0076
Change in x0: 0.2069
********* Velocity Refinement Round 14 *********
********* Round 14: Early Stop Triggered at epoch 2389. *********
Change in noise variance: 0.0076
Change in x0: 0.1763
********* Velocity Refinement Round 15 *********
********* Round 15: Early Stop Triggered at epoch 2489. *********
Change in noise variance: 0.0028
Change in x0: 0.1887
********* Velocity Refinement Round 16 *********
********* Round 16: Early Stop Triggered at epoch 2534. *********
Change in noise variance: 0.0029
Change in x0: 0.1844
********* Velocity Refinement Round 17 *********
********* Round 17: Early Stop Triggered at epoch 2606. *********
Change in noise variance: 0.0053
Change in x0: 0.1728
********* Velocity Refinement Round 18 *********
********* Round 18: Early Stop Triggered at epoch 2656. *********
Change in noise variance: 0.0058
Change in x0: 0.1727
********* Velocity Refinement Round 19 *********
********* Round 19: Early Stop Triggered at epoch 2794. *********
Change in noise variance: 0.0049
Change in x0: 0.1381
********* Velocity Refinement Round 20 *********
********* Round 20: Early Stop Triggered at epoch 2819. *********
Change in noise variance: 0.0000
Change in x0: 0.4219
********* Finished. Total Time = 0 h : 2 m : 9 s *********
Final: Train ELBO = -988.844, Test ELBO = -1027.358
Estimating ODE parameters...
Detected 67 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
KS-test result: [1. 1. 1.]
Assign cluster 1 to repressive
Initial induction: 72, repression: 37/109
Learning Rate based on Data Sparsity: 0.0001
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 6, test iteration: 10
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.063
Average Set Size: 20
********* Round 1: Early Stop Triggered at epoch 1247. *********
Change in noise variance: 0.0386
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 1411. *********
Change in noise variance: 0.0092
Change in x0: 0.2743
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 1537. *********
Change in noise variance: 0.0033
Change in x0: 0.1806
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 1593. *********
Change in noise variance: 0.0023
Change in x0: 0.1571
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 1650. *********
Change in noise variance: 0.0009
Change in x0: 0.1232
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 1712. *********
Change in noise variance: 0.0000
Change in x0: 0.0968
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 1793. *********
Change in noise variance: 0.0000
Change in x0: 0.0810
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 2065. *********
Change in noise variance: 0.0000
Change in x0: 0.0707
********* Velocity Refinement Round 9 *********
********* Round 9: Early Stop Triggered at epoch 2164. *********
Change in noise variance: 0.0000
Change in x0: 0.0819
********* Velocity Refinement Round 10 *********
Stage 2: Early Stop Triggered at round 9.
********* Finished. Total Time = 0 h : 3 m : 30 s *********
Final: Train ELBO = 126.385, Test ELBO = 136.408
Estimating ODE parameters...
Detected 18 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
(0.47, 0.8701520422929223), (0.53, 0.2900017161456463)
KS-test result: [1. 2. 0.]
Initial induction: 20, repression: 41/61
Learning Rate based on Data Sparsity: 0.0001
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 2, test iteration: 2
********* Stage 1: Early Stop Triggered at epoch 8. *********
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.060
Average Set Size: 16
********* Round 1: Early Stop Triggered at epoch 14. *********
Change in noise variance: 0.5563
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 119. *********
Change in noise variance: 0.0420
Change in x0: 0.4501
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 150. *********
Change in noise variance: 0.0017
Change in x0: 0.3890
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 158. *********
Change in noise variance: 0.0033
Change in x0: 0.3775
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 165. *********
Change in noise variance: 0.0030
Change in x0: 0.2746
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 171. *********
Change in noise variance: 0.0016
Change in x0: 0.2153
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 178. *********
Change in noise variance: 0.0006
Change in x0: 0.2057
********* Velocity Refinement Round 8 *********
Stage 2: Early Stop Triggered at round 7.
********* Finished. Total Time = 0 h : 0 m : 20 s *********
Final: Train ELBO = -13756.597, Test ELBO = -13157.369
Estimating ODE parameters...
Detected 34 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
KS-test result: [1. 1. 1.]
Assign cluster 2 to repressive
Initial induction: 45, repression: 18/63
Learning Rate based on Data Sparsity: 0.0000
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 2, test iteration: 2
********* Stage 1: Early Stop Triggered at epoch 8. *********
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.023
Average Set Size: 17
********* Round 1: Early Stop Triggered at epoch 14. *********
Change in noise variance: 0.1852
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 20. *********
Change in noise variance: 0.0144
Change in x0: 0.7071
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 26. *********
Change in noise variance: 0.0210
Change in x0: 0.4340
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 31. *********
Change in noise variance: 0.0193
Change in x0: 0.3304
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 36. *********
Change in noise variance: 0.0129
Change in x0: 0.2647
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 41. *********
Change in noise variance: 0.0063
Change in x0: 0.2390
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 47. *********
Change in noise variance: 0.0072
Change in x0: 0.2218
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 53. *********
Change in noise variance: 0.0059
Change in x0: 0.2065
********* Velocity Refinement Round 9 *********
********* Round 9: Early Stop Triggered at epoch 59. *********
Change in noise variance: 0.0038
Change in x0: 0.1955
********* Velocity Refinement Round 10 *********
********* Round 10: Early Stop Triggered at epoch 64. *********
Change in noise variance: 0.0022
Change in x0: 0.1854
********* Velocity Refinement Round 11 *********
********* Round 11: Early Stop Triggered at epoch 69. *********
Change in noise variance: 0.0013
Change in x0: 0.1735
********* Velocity Refinement Round 12 *********
********* Round 12: Early Stop Triggered at epoch 74. *********
Change in noise variance: 0.0008
Change in x0: 0.1638
********* Velocity Refinement Round 13 *********
Stage 2: Early Stop Triggered at round 12.
********* Finished. Total Time = 0 h : 0 m : 16 s *********
Final: Train ELBO = -18537.670, Test ELBO = -18844.469
Estimating ODE parameters...
Detected 32 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
(0.56, 0.3239376112994556), (0.44, 0.9106431281299979)
(0.52, 0.2605808478636782), (0.48, 0.805314977942315)
KS-test result: [1. 0. 0.]
Initial induction: 43, repression: 37/80
Learning Rate based on Data Sparsity: 0.0000
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 2, test iteration: 2
********* Stage 1: Early Stop Triggered at epoch 5. *********
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.063
Average Set Size: 16
********* Round 1: Early Stop Triggered at epoch 21. *********
Change in noise variance: 0.4645
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 26. *********
Change in noise variance: 0.0397
Change in x0: 0.4195
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 32. *********
Change in noise variance: 0.0093
Change in x0: 0.2190
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 38. *********
Change in noise variance: 0.0051
Change in x0: 0.1891
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 44. *********
Change in noise variance: 0.0041
Change in x0: 0.1706
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 50. *********
Change in noise variance: 0.0021
Change in x0: 0.1658
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 56. *********
Change in noise variance: 0.0007
Change in x0: 0.1556
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 61. *********
Change in noise variance: 0.0000
Change in x0: 0.1380
********* Velocity Refinement Round 9 *********
********* Round 9: Early Stop Triggered at epoch 66. *********
Change in noise variance: 0.0000
Change in x0: 0.1268
********* Velocity Refinement Round 10 *********
********* Round 10: Early Stop Triggered at epoch 71. *********
Change in noise variance: 0.0000
Change in x0: 0.1199
********* Velocity Refinement Round 11 *********
Stage 2: Early Stop Triggered at round 10.
********* Finished. Total Time = 0 h : 0 m : 15 s *********
Final: Train ELBO = -13288.914, Test ELBO = -13578.303
Estimating ODE parameters...
Detected 20 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
(0.53, 0.06264460485018672), (0.47, 0.7010424470363079)
(0.58, 0.707388107196621), (0.42, 0.09198273775867975)
KS-test result: [0. 0. 2.]
Initial induction: 27, repression: 49/76
Learning Rate based on Data Sparsity: 0.0000
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 2, test iteration: 2
********* Stage 1: Early Stop Triggered at epoch 9. *********
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.017
Average Set Size: 16
********* Round 1: Early Stop Triggered at epoch 15. *********
Change in noise variance: 0.2935
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 65. *********
Change in noise variance: 0.0327
Change in x0: 0.7020
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 71. *********
Change in noise variance: 0.0079
Change in x0: 0.3804
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 77. *********
Change in noise variance: 0.0046
Change in x0: 0.2670
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 83. *********
Change in noise variance: 0.0010
Change in x0: 0.2560
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 89. *********
Change in noise variance: 0.0012
Change in x0: 0.2223
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 95. *********
Change in noise variance: 0.0011
Change in x0: 0.2148
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 101. *********
Change in noise variance: 0.0007
Change in x0: 0.2009
********* Velocity Refinement Round 9 *********
********* Round 9: Early Stop Triggered at epoch 106. *********
Change in noise variance: 0.0000
Change in x0: 0.1751
********* Velocity Refinement Round 10 *********
********* Round 10: Early Stop Triggered at epoch 112. *********
Change in noise variance: 0.0000
Change in x0: 0.1636
********* Velocity Refinement Round 11 *********
********* Round 11: Early Stop Triggered at epoch 118. *********
Change in noise variance: 0.0000
Change in x0: 0.1660
********* Velocity Refinement Round 12 *********
Stage 2: Early Stop Triggered at round 11.
********* Finished. Total Time = 0 h : 0 m : 15 s *********
Final: Train ELBO = -15358.680, Test ELBO = -14971.483
Estimating ODE parameters...
Detected 35 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
(0.56, 0.24297477479231852), (0.44, 0.7779664767114342)
KS-test result: [0. 1. 1.]
Initial induction: 60, repression: 23/83
Learning Rate based on Data Sparsity: 0.0000
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 2, test iteration: 2
********* Stage 1: Early Stop Triggered at epoch 7. *********
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.030
Average Set Size: 16
********* Round 1: Early Stop Triggered at epoch 13. *********
Change in noise variance: 0.5683
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 20. *********
Change in noise variance: 0.2858
Change in x0: 0.4193
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 26. *********
Change in noise variance: 0.0066
Change in x0: 0.3290
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 32. *********
Change in noise variance: 0.0037
Change in x0: 0.2685
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 38. *********
Change in noise variance: 0.0024
Change in x0: 0.2072
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 45. *********
Change in noise variance: 0.0017
Change in x0: 0.1657
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 51. *********
Change in noise variance: 0.0008
Change in x0: 0.1436
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 56. *********
Change in noise variance: 0.0000
Change in x0: 0.1297
********* Velocity Refinement Round 9 *********
********* Round 9: Early Stop Triggered at epoch 61. *********
Change in noise variance: 0.0000
Change in x0: 0.1180
********* Velocity Refinement Round 10 *********
********* Round 10: Early Stop Triggered at epoch 66. *********
Change in noise variance: 0.0000
Change in x0: 0.1084
********* Velocity Refinement Round 11 *********
Stage 2: Early Stop Triggered at round 10.
********* Finished. Total Time = 0 h : 0 m : 15 s *********
Final: Train ELBO = -18305.402, Test ELBO = -18420.889
Estimating ODE parameters...
Detected 30 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
(0.62, 0.1494859643683501), (0.38, 0.8554410718115751)
KS-test result: [0. 1. 1.]
Initial induction: 49, repression: 23/72
Learning Rate based on Data Sparsity: 0.0000
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 3, test iteration: 4
********* Stage 1: Early Stop Triggered at epoch 10. *********
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.045
Average Set Size: 18
********* Round 1: Early Stop Triggered at epoch 57. *********
Change in noise variance: 0.2595
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 184. *********
Change in noise variance: 0.0213
Change in x0: 0.4230
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 195. *********
Change in noise variance: 0.0013
Change in x0: 0.4526
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 202. *********
Change in noise variance: 0.0008
Change in x0: 0.3712
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 217. *********
Change in noise variance: 0.0000
Change in x0: 0.2914
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 224. *********
Change in noise variance: 0.0000
Change in x0: 0.2443
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 235. *********
Change in noise variance: 0.0000
Change in x0: 0.2246
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 242. *********
Change in noise variance: 0.0000
Change in x0: 0.2088
********* Velocity Refinement Round 9 *********
********* Round 9: Early Stop Triggered at epoch 249. *********
Change in noise variance: 0.0000
Change in x0: 0.1991
********* Velocity Refinement Round 10 *********
Stage 2: Early Stop Triggered at round 9.
********* Finished. Total Time = 0 h : 0 m : 21 s *********
Final: Train ELBO = -12114.753, Test ELBO = -12328.426
Estimating ODE parameters...
Detected 31 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
(0.62, 0.7680112510853812), (0.38, 0.1779761449705523)
KS-test result: [1. 1. 0.]
Initial induction: 72, repression: 13/85
Learning Rate based on Data Sparsity: 0.0000
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 2, test iteration: 2
********* Stage 1: Early Stop Triggered at epoch 215. *********
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.023
Average Set Size: 16
********* Round 1: Early Stop Triggered at epoch 625. *********
Change in noise variance: 0.7099
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 697. *********
Change in noise variance: 0.0840
Change in x0: 0.2672
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 714. *********
Change in noise variance: 0.0085
Change in x0: 0.2250
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 720. *********
Change in noise variance: 0.0042
Change in x0: 0.1654
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 725. *********
Change in noise variance: 0.0017
Change in x0: 0.1122
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 731. *********
Change in noise variance: 0.0010
Change in x0: 0.0804
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 737. *********
Change in noise variance: 0.0000
Change in x0: 0.0653
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 799. *********
Change in noise variance: 0.0000
Change in x0: 0.0524
********* Velocity Refinement Round 9 *********
********* Round 9: Early Stop Triggered at epoch 808. *********
Change in noise variance: 0.0000
Change in x0: 0.0711
********* Velocity Refinement Round 10 *********
Stage 2: Early Stop Triggered at round 9.
********* Finished. Total Time = 0 h : 0 m : 52 s *********
Final: Train ELBO = -8640.196, Test ELBO = -8722.279
Estimating ODE parameters...
Detected 37 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
KS-test result: [1. 1. 1.]
Assign cluster 1 to repressive
Initial induction: 50, repression: 29/79
Learning Rate based on Data Sparsity: 0.0000
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 2, test iteration: 2
********* Stage 1: Early Stop Triggered at epoch 10. *********
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.033
Average Set Size: 17
********* Round 1: Early Stop Triggered at epoch 16. *********
Change in noise variance: 0.2294
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 21. *********
Change in noise variance: 0.0072
Change in x0: 0.5413
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 30. *********
Change in noise variance: 0.0046
Change in x0: 0.3871
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 36. *********
Change in noise variance: 0.0114
Change in x0: 0.3332
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 45. *********
Change in noise variance: 0.0150
Change in x0: 0.2683
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 54. *********
Change in noise variance: 0.0011
Change in x0: 0.2226
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 62. *********
Change in noise variance: 0.0041
Change in x0: 0.1993
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 68. *********
Change in noise variance: 0.0036
Change in x0: 0.1888
********* Velocity Refinement Round 9 *********
********* Round 9: Early Stop Triggered at epoch 73. *********
Change in noise variance: 0.0025
Change in x0: 0.1799
********* Velocity Refinement Round 10 *********
********* Round 10: Early Stop Triggered at epoch 85. *********
Change in noise variance: 0.0016
Change in x0: 0.1698
********* Velocity Refinement Round 11 *********
********* Round 11: Early Stop Triggered at epoch 92. *********
Change in noise variance: 0.0010
Change in x0: 0.1602
********* Velocity Refinement Round 12 *********
********* Round 12: Early Stop Triggered at epoch 101. *********
Change in noise variance: 0.0007
Change in x0: 0.1550
********* Velocity Refinement Round 13 *********
Stage 2: Early Stop Triggered at round 12.
********* Finished. Total Time = 0 h : 0 m : 18 s *********
Final: Train ELBO = -16939.322, Test ELBO = -17156.008
Estimating ODE parameters...
Detected 38 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
KS-test result: [1. 1. 1.]
Assign cluster 1 to repressive
Initial induction: 56, repression: 27/83
Learning Rate based on Data Sparsity: 0.0000
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 2, test iteration: 2
********* Stage 1: Early Stop Triggered at epoch 24. *********
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.030
Average Set Size: 17
********* Round 1: Early Stop Triggered at epoch 39. *********
Change in noise variance: 0.2136
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 44. *********
Change in noise variance: 0.0086
Change in x0: 0.6531
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 224. *********
Change in noise variance: 0.0015
Change in x0: 0.4806
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 254. *********
Change in noise variance: 0.0004
Change in x0: 0.3275
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 298. *********
Change in noise variance: 0.0000
Change in x0: 0.2862
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 322. *********
Change in noise variance: 0.0000
Change in x0: 0.2618
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 347. *********
Change in noise variance: 0.0000
Change in x0: 0.2411
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 432. *********
Change in noise variance: 0.0000
Change in x0: 0.2238
********* Velocity Refinement Round 9 *********
********* Round 9: Early Stop Triggered at epoch 442. *********
Change in noise variance: 0.0000
Change in x0: 0.2127
********* Velocity Refinement Round 10 *********
********* Round 10: Early Stop Triggered at epoch 478. *********
Change in noise variance: 0.0000
Change in x0: 0.2050
********* Velocity Refinement Round 11 *********
Stage 2: Early Stop Triggered at round 10.
********* Finished. Total Time = 0 h : 0 m : 37 s *********
Final: Train ELBO = -13098.929, Test ELBO = -13426.184
Estimating ODE parameters...
Detected 35 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
KS-test result: [1. 1. 1.]
Assign cluster 2 to repressive
Initial induction: 48, repression: 33/81
Learning Rate based on Data Sparsity: 0.0000
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 2, test iteration: 2
********* Stage 1: Early Stop Triggered at epoch 940. *********
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.033
Average Set Size: 23
********* Round 1: Early Stop Triggered at epoch 1145. *********
Change in noise variance: 0.2259
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 1151. *********
Change in noise variance: 0.2420
Change in x0: 0.9584
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 1156. *********
Change in noise variance: 0.0045
Change in x0: 0.7704
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 1161. *********
Change in noise variance: 0.0063
Change in x0: 0.7077
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 1166. *********
Change in noise variance: 0.0023
Change in x0: 0.5915
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 1171. *********
Change in noise variance: 0.0040
Change in x0: 0.4506
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 1176. *********
Change in noise variance: 0.0052
Change in x0: 0.3248
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 1181. *********
Change in noise variance: 0.0035
Change in x0: 0.2641
********* Velocity Refinement Round 9 *********
********* Round 9: Early Stop Triggered at epoch 1186. *********
Change in noise variance: 0.0042
Change in x0: 0.2259
********* Velocity Refinement Round 10 *********
********* Round 10: Early Stop Triggered at epoch 1403. *********
Change in noise variance: 0.0026
Change in x0: 0.1946
********* Velocity Refinement Round 11 *********
********* Round 11: Early Stop Triggered at epoch 1408. *********
Change in noise variance: 0.0060
Change in x0: 0.1857
********* Velocity Refinement Round 12 *********
********* Round 12: Early Stop Triggered at epoch 1413. *********
Change in noise variance: 0.0029
Change in x0: 0.1329
********* Velocity Refinement Round 13 *********
********* Round 13: Early Stop Triggered at epoch 1501. *********
Change in noise variance: 0.0042
Change in x0: 0.0993
********* Velocity Refinement Round 14 *********
********* Round 14: Early Stop Triggered at epoch 1506. *********
Change in noise variance: 0.0026
Change in x0: 0.0917
********* Velocity Refinement Round 15 *********
********* Round 15: Early Stop Triggered at epoch 1512. *********
Change in noise variance: 0.0027
Change in x0: 0.0536
********* Velocity Refinement Round 16 *********
********* Round 16: Early Stop Triggered at epoch 1753. *********
Change in noise variance: 0.0038
Change in x0: 0.0348
********* Velocity Refinement Round 17 *********
********* Round 17: Early Stop Triggered at epoch 1768. *********
Change in noise variance: 0.0020
Change in x0: 0.1141
********* Velocity Refinement Round 18 *********
********* Round 18: Early Stop Triggered at epoch 1774. *********
Change in noise variance: 0.0012
Change in x0: 0.0552
********* Velocity Refinement Round 19 *********
********* Round 19: Early Stop Triggered at epoch 1847. *********
Change in noise variance: 0.0014
Change in x0: 0.0458
********* Velocity Refinement Round 20 *********
Stage 2: Early Stop Triggered at round 19.
********* Finished. Total Time = 0 h : 1 m : 33 s *********
Final: Train ELBO = -1605.839, Test ELBO = -1667.010
Estimating ODE parameters...
Detected 30 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
(0.40, 0.17999999999639998), (0.60, 0.7935524750489109)
KS-test result: [1. 1. 0.]
Initial induction: 60, repression: 14/74
Learning Rate based on Data Sparsity: 0.0000
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 2, test iteration: 2
********* Stage 1: Early Stop Triggered at epoch 9. *********
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.020
Average Set Size: 12
********* Round 1: Early Stop Triggered at epoch 200. *********
Change in noise variance: 1.1207
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 229. *********
Change in noise variance: 0.0484
Change in x0: 0.4921
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 240. *********
Change in noise variance: 0.0027
Change in x0: 0.3630
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 266. *********
Change in noise variance: 0.0016
Change in x0: 0.3171
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 282. *********
Change in noise variance: 0.0009
Change in x0: 0.2796
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 303. *********
Change in noise variance: 0.0000
Change in x0: 0.2425
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 309. *********
Change in noise variance: 0.0000
Change in x0: 0.2244
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 337. *********
Change in noise variance: 0.0000
Change in x0: 0.2111
********* Velocity Refinement Round 9 *********
********* Round 9: Early Stop Triggered at epoch 363. *********
Change in noise variance: 0.0000
Change in x0: 0.1923
********* Velocity Refinement Round 10 *********
********* Round 10: Early Stop Triggered at epoch 372. *********
Change in noise variance: 0.0000
Change in x0: 0.1785
********* Velocity Refinement Round 11 *********
********* Round 11: Early Stop Triggered at epoch 384. *********
Change in noise variance: 0.0000
Change in x0: 0.1698
********* Velocity Refinement Round 12 *********
Stage 2: Early Stop Triggered at round 11.
********* Finished. Total Time = 0 h : 0 m : 26 s *********
Final: Train ELBO = -15334.994, Test ELBO = -15484.329
Estimating ODE parameters...
Detected 31 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
KS-test result: [1. 1. 1.]
Assign cluster 0 to repressive
Initial induction: 36, repression: 14/50
Learning Rate based on Data Sparsity: 0.0000
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 2, test iteration: 2
********* Stage 1: Early Stop Triggered at epoch 8. *********
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.030
Average Set Size: 17
********* Round 1: Early Stop Triggered at epoch 14. *********
Change in noise variance: 0.2023
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 19. *********
Change in noise variance: 0.0102
Change in x0: 0.6114
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 25. *********
Change in noise variance: 0.0015
Change in x0: 0.4109
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 31. *********
Change in noise variance: 0.0010
Change in x0: 0.2966
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 36. *********
Change in noise variance: 0.0000
Change in x0: 0.2537
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 41. *********
Change in noise variance: 0.0000
Change in x0: 0.2370
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 46. *********
Change in noise variance: 0.0000
Change in x0: 0.2231
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 51. *********
Change in noise variance: 0.0000
Change in x0: 0.2072
********* Velocity Refinement Round 9 *********
********* Round 9: Early Stop Triggered at epoch 59. *********
Change in noise variance: 0.0000
Change in x0: 0.1936
********* Velocity Refinement Round 10 *********
********* Round 10: Early Stop Triggered at epoch 67. *********
Change in noise variance: 0.0000
Change in x0: 0.1863
********* Velocity Refinement Round 11 *********
Stage 2: Early Stop Triggered at round 10.
********* Finished. Total Time = 0 h : 0 m : 18 s *********
Final: Train ELBO = -19504.984, Test ELBO = -19992.318
Estimating ODE parameters...
Detected 28 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
(0.36, 0.9320509281825696), (0.64, 0.2539013763998727)
(0.30, 0.05787037036969842), (0.70, 0.8620463316579644)
KS-test result: [1. 0. 0.]
Initial induction: 37, repression: 35/72
Learning Rate based on Data Sparsity: 0.0000
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 2, test iteration: 2
********* Stage 1: Early Stop Triggered at epoch 7. *********
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.023
Average Set Size: 17
********* Round 1: Early Stop Triggered at epoch 13. *********
Change in noise variance: 0.2301
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 18. *********
Change in noise variance: 0.0175
Change in x0: 0.6791
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 24. *********
Change in noise variance: 0.0021
Change in x0: 0.4718
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 30. *********
Change in noise variance: 0.0061
Change in x0: 0.3152
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 36. *********
Change in noise variance: 0.0028
Change in x0: 0.2411
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 42. *********
Change in noise variance: 0.0006
Change in x0: 0.2266
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 48. *********
Change in noise variance: 0.0000
Change in x0: 0.2186
********* Velocity Refinement Round 8 *********
Stage 2: Early Stop Triggered at round 7.
********* Finished. Total Time = 0 h : 0 m : 18 s *********
Final: Train ELBO = -16154.742, Test ELBO = -15867.226
Estimating ODE parameters...
Detected 37 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
(0.56, 0.24578835857334116), (0.44, 0.8411598564162884)
KS-test result: [1. 0. 1.]
Initial induction: 55, repression: 22/77
Learning Rate based on Data Sparsity: 0.0000
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 2, test iteration: 2
********* Stage 1: Early Stop Triggered at epoch 10. *********
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.037
Average Set Size: 17
********* Round 1: Early Stop Triggered at epoch 24. *********
Change in noise variance: 0.3507
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 29. *********
Change in noise variance: 0.0316
Change in x0: 0.4174
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 35. *********
Change in noise variance: 0.0019
Change in x0: 0.3224
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 41. *********
Change in noise variance: 0.0016
Change in x0: 0.2422
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 47. *********
Change in noise variance: 0.0008
Change in x0: 0.2128
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 53. *********
Change in noise variance: 0.0000
Change in x0: 0.1934
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 59. *********
Change in noise variance: 0.0000
Change in x0: 0.1704
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 67. *********
Change in noise variance: 0.0000
Change in x0: 0.1488
********* Velocity Refinement Round 9 *********
********* Round 9: Early Stop Triggered at epoch 73. *********
Change in noise variance: 0.0000
Change in x0: 0.1394
********* Velocity Refinement Round 10 *********
Stage 2: Early Stop Triggered at round 9.
********* Finished. Total Time = 0 h : 0 m : 18 s *********
Final: Train ELBO = -14676.509, Test ELBO = -15549.430
Estimating ODE parameters...
Detected 33 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
KS-test result: [1. 1. 1.]
Assign cluster 0 to repressive
Initial induction: 47, repression: 26/73
Learning Rate based on Data Sparsity: 0.0000
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 2, test iteration: 2
********* Stage 1: Early Stop Triggered at epoch 6. *********
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.023
Average Set Size: 17
********* Round 1: Early Stop Triggered at epoch 13. *********
Change in noise variance: 0.3068
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 334. *********
Change in noise variance: 0.0243
Change in x0: 0.6400
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 408. *********
Change in noise variance: 0.0046
Change in x0: 0.4568
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 437. *********
Change in noise variance: 0.0017
Change in x0: 0.3252
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 466. *********
Change in noise variance: 0.0013
Change in x0: 0.2669
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 489. *********
Change in noise variance: 0.0010
Change in x0: 0.2328
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 494. *********
Change in noise variance: 0.0000
Change in x0: 0.2068
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 504. *********
Change in noise variance: 0.0000
Change in x0: 0.1984
********* Velocity Refinement Round 9 *********
Stage 2: Early Stop Triggered at round 8.
********* Finished. Total Time = 0 h : 0 m : 23 s *********
Final: Train ELBO = -15790.699, Test ELBO = -15499.406
Estimating ODE parameters...
Detected 35 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
An error occurred: alpha <= 0
Estimating ODE parameters...
Detected 31 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
(0.61, 0.26568115737011694), (0.39, 0.8432651228926057)
KS-test result: [0. 1. 1.]
Initial induction: 50, repression: 29/79
Learning Rate based on Data Sparsity: 0.0000
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 2, test iteration: 2
********* Stage 1: Early Stop Triggered at epoch 6. *********
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.053
Average Set Size: 16
********* Round 1: Early Stop Triggered at epoch 12. *********
Change in noise variance: 0.3293
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 17. *********
Change in noise variance: 0.0461
Change in x0: 0.2776
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 22. *********
Change in noise variance: 0.0068
Change in x0: 0.2148
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 27. *********
Change in noise variance: 0.0019
Change in x0: 0.2173
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 33. *********
Change in noise variance: 0.0067
Change in x0: 0.2049
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 39. *********
Change in noise variance: 0.0038
Change in x0: 0.1761
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 45. *********
Change in noise variance: 0.0015
Change in x0: 0.1639
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 51. *********
Change in noise variance: 0.0006
Change in x0: 0.1546
********* Velocity Refinement Round 9 *********
Stage 2: Early Stop Triggered at round 8.
********* Finished. Total Time = 0 h : 0 m : 14 s *********
Final: Train ELBO = -19448.113, Test ELBO = -19306.715
Estimating ODE parameters...
Detected 21 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
KS-test result: [2. 1. 1.]
Initial induction: 28, repression: 26/54
Learning Rate based on Data Sparsity: 0.0001
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 2, test iteration: 2
********* Stage 1: Early Stop Triggered at epoch 249. *********
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.037
Average Set Size: 16
********* Round 1: Early Stop Triggered at epoch 255. *********
Change in noise variance: 0.2184
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 322. *********
Change in noise variance: 0.0163
Change in x0: 0.8697
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 327. *********
Change in noise variance: 0.0055
Change in x0: 0.6548
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 355. *********
Change in noise variance: 0.0160
Change in x0: 0.4874
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 372. *********
Change in noise variance: 0.0071
Change in x0: 0.3342
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 377. *********
Change in noise variance: 0.0011
Change in x0: 0.2571
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 387. *********
Change in noise variance: 0.0011
Change in x0: 0.2222
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 392. *********
Change in noise variance: 0.0009
Change in x0: 0.2013
********* Velocity Refinement Round 9 *********
********* Round 9: Early Stop Triggered at epoch 404. *********
Change in noise variance: 0.0000
Change in x0: 0.1815
********* Velocity Refinement Round 10 *********
********* Round 10: Early Stop Triggered at epoch 409. *********
Change in noise variance: 0.0000
Change in x0: 0.1691
********* Velocity Refinement Round 11 *********
********* Round 11: Early Stop Triggered at epoch 414. *********
Change in noise variance: 0.0000
Change in x0: 0.1666
********* Velocity Refinement Round 12 *********
Stage 2: Early Stop Triggered at round 11.
********* Finished. Total Time = 0 h : 0 m : 33 s *********
Final: Train ELBO = -847.994, Test ELBO = -912.455
Estimating ODE parameters...
Detected 81 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
(0.49, 0.23313747351401098), (0.51, 0.8250112585464345)
KS-test result: [1. 0. 1.]
Initial induction: 104, repression: 36/140
Learning Rate based on Data Sparsity: 0.0000
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 6, test iteration: 10
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.037
Average Set Size: 20
********* Round 1: Early Stop Triggered at epoch 1009. *********
Change in noise variance: 0.1328
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 1051. *********
Change in noise variance: 0.0066
Change in x0: 0.5858
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 1058. *********
Change in noise variance: 0.0054
Change in x0: 0.4545
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 1065. *********
Change in noise variance: 0.0026
Change in x0: 0.3529
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 1072. *********
Change in noise variance: 0.0021
Change in x0: 0.3174
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 1079. *********
Change in noise variance: 0.0011
Change in x0: 0.2811
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 1086. *********
Change in noise variance: 0.0007
Change in x0: 0.2527
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 1093. *********
Change in noise variance: 0.0000
Change in x0: 0.2334
********* Velocity Refinement Round 9 *********
********* Round 9: Early Stop Triggered at epoch 1125. *********
Change in noise variance: 0.0000
Change in x0: 0.2108
********* Velocity Refinement Round 10 *********
********* Round 10: Early Stop Triggered at epoch 1191. *********
Change in noise variance: 0.0000
Change in x0: 0.1929
********* Velocity Refinement Round 11 *********
********* Round 11: Early Stop Triggered at epoch 1227. *********
Change in noise variance: 0.0000
Change in x0: 0.1785
********* Velocity Refinement Round 12 *********
********* Round 12: Early Stop Triggered at epoch 1234. *********
Change in noise variance: 0.0000
Change in x0: 0.1787
********* Velocity Refinement Round 13 *********
Stage 2: Early Stop Triggered at round 12.
********* Finished. Total Time = 0 h : 3 m : 19 s *********
Final: Train ELBO = -228.650, Test ELBO = -247.470
Estimating ODE parameters...
Detected 35 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
KS-test result: [1. 1. 1.]
Assign cluster 0 to repressive
Initial induction: 45, repression: 27/72
Learning Rate based on Data Sparsity: 0.0000
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 2, test iteration: 2
********* Stage 1: Early Stop Triggered at epoch 9. *********
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.020
Average Set Size: 18
********* Round 1: Early Stop Triggered at epoch 15. *********
Change in noise variance: 0.3534
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 20. *********
Change in noise variance: 0.0696
Change in x0: 0.4406
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 35. *********
Change in noise variance: 0.0021
Change in x0: 0.3764
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 41. *********
Change in noise variance: 0.0032
Change in x0: 0.3248
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 47. *********
Change in noise variance: 0.0029
Change in x0: 0.2676
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 53. *********
Change in noise variance: 0.0014
Change in x0: 0.2334
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 59. *********
Change in noise variance: 0.0007
Change in x0: 0.2122
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 65. *********
Change in noise variance: 0.0000
Change in x0: 0.1937
********* Velocity Refinement Round 9 *********
********* Round 9: Early Stop Triggered at epoch 71. *********
Change in noise variance: 0.0000
Change in x0: 0.1762
********* Velocity Refinement Round 10 *********
********* Round 10: Early Stop Triggered at epoch 77. *********
Change in noise variance: 0.0000
Change in x0: 0.1606
********* Velocity Refinement Round 11 *********
********* Round 11: Early Stop Triggered at epoch 83. *********
Change in noise variance: 0.0000
Change in x0: 0.1500
********* Velocity Refinement Round 12 *********
********* Round 12: Early Stop Triggered at epoch 89. *********
Change in noise variance: 0.0000
Change in x0: 0.1462
********* Velocity Refinement Round 13 *********
Stage 2: Early Stop Triggered at round 12.
********* Finished. Total Time = 0 h : 0 m : 18 s *********
Final: Train ELBO = -14246.879, Test ELBO = -14316.127
Estimating ODE parameters...
Detected 106 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
KS-test result: [1. 1. 1.]
Assign cluster 2 to repressive
Initial induction: 111, repression: 43/154
Learning Rate based on Data Sparsity: 0.0001
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 6, test iteration: 10
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.044
Average Set Size: 21
********* Round 1: Early Stop Triggered at epoch 1127. *********
Change in noise variance: 0.0615
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 1236. *********
Change in noise variance: 0.0089
Change in x0: 0.3694
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 1320. *********
Change in noise variance: 0.0034
Change in x0: 0.2746
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 1526. *********
Change in noise variance: 0.0015
Change in x0: 0.2268
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 1680. *********
Change in noise variance: 0.0009
Change in x0: 0.1868
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 1936. *********
Change in noise variance: 0.0000
Change in x0: 0.1435
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 2020. *********
Change in noise variance: 0.0000
Change in x0: 0.1323
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 2157. *********
Change in noise variance: 0.0000
Change in x0: 0.1100
********* Velocity Refinement Round 9 *********
********* Round 9: Early Stop Triggered at epoch 2258. *********
Change in noise variance: 0.0000
Change in x0: 0.1042
********* Velocity Refinement Round 10 *********
Stage 2: Early Stop Triggered at round 9.
********* Finished. Total Time = 0 h : 5 m : 35 s *********
Final: Train ELBO = 108.060, Test ELBO = 96.440
Estimating ODE parameters...
Detected 34 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
(0.59, 0.11364537925550108), (0.41, 0.9060473661761724)
KS-test result: [1. 1. 0.]
Initial induction: 54, repression: 19/73
Learning Rate based on Data Sparsity: 0.0000
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 2, test iteration: 2
********* Stage 1: Early Stop Triggered at epoch 8. *********
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.030
Average Set Size: 16
********* Round 1: Early Stop Triggered at epoch 14. *********
Change in noise variance: 0.2778
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 19. *********
Change in noise variance: 0.0288
Change in x0: 0.6292
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 25. *********
Change in noise variance: 0.0010
Change in x0: 0.3971
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 31. *********
Change in noise variance: 0.0000
Change in x0: 0.3093
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 37. *********
Change in noise variance: 0.0000
Change in x0: 0.2501
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 43. *********
Change in noise variance: 0.0000
Change in x0: 0.2238
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 49. *********
Change in noise variance: 0.0000
Change in x0: 0.2013
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 55. *********
Change in noise variance: 0.0000
Change in x0: 0.1853
********* Velocity Refinement Round 9 *********
********* Round 9: Early Stop Triggered at epoch 61. *********
Change in noise variance: 0.0000
Change in x0: 0.1730
********* Velocity Refinement Round 10 *********
********* Round 10: Early Stop Triggered at epoch 67. *********
Change in noise variance: 0.0000
Change in x0: 0.1671
********* Velocity Refinement Round 11 *********
Stage 2: Early Stop Triggered at round 10.
********* Finished. Total Time = 0 h : 0 m : 18 s *********
Final: Train ELBO = -14860.914, Test ELBO = -14663.303
Estimating ODE parameters...
Detected 62 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
KS-test result: [1. 1. 1.]
Assign cluster 0 to repressive
Initial induction: 68, repression: 49/117
Learning Rate based on Data Sparsity: 0.0001
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 6, test iteration: 10
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.039
Average Set Size: 20
********* Round 1: Early Stop Triggered at epoch 1159. *********
Change in noise variance: 0.0512
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 1513. *********
Change in noise variance: 0.0108
Change in x0: 0.3184
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 1687. *********
Change in noise variance: 0.0063
Change in x0: 0.2393
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 1766. *********
Change in noise variance: 0.0031
Change in x0: 0.1870
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 1983. *********
Change in noise variance: 0.0010
Change in x0: 0.1339
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 2135. *********
Change in noise variance: 0.0000
Change in x0: 0.1055
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 2149. *********
Change in noise variance: 0.0000
Change in x0: 0.0870
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 2171. *********
Change in noise variance: 0.0000
Change in x0: 0.0733
********* Velocity Refinement Round 9 *********
********* Round 9: Early Stop Triggered at epoch 2207. *********
Change in noise variance: 0.0000
Change in x0: 0.0647
********* Velocity Refinement Round 10 *********
Stage 2: Early Stop Triggered at round 9.
********* Finished. Total Time = 0 h : 3 m : 48 s *********
Final: Train ELBO = 177.682, Test ELBO = 170.667
Estimating ODE parameters...
Detected 28 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
KS-test result: [1. 1. 1.]
Assign cluster 1 to repressive
Initial induction: 44, repression: 26/70
Learning Rate based on Data Sparsity: 0.0000
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 2, test iteration: 2
********* Stage 1: Early Stop Triggered at epoch 16. *********
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.027
Average Set Size: 17
********* Round 1: Early Stop Triggered at epoch 235. *********
Change in noise variance: 0.5609
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 313. *********
Change in noise variance: 0.0216
Change in x0: 0.1737
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 330. *********
Change in noise variance: 0.0028
Change in x0: 0.1804
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 336. *********
Change in noise variance: 0.0010
Change in x0: 0.1564
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 342. *********
Change in noise variance: 0.0009
Change in x0: 0.1409
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 383. *********
Change in noise variance: 0.0000
Change in x0: 0.1138
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 389. *********
Change in noise variance: 0.0000
Change in x0: 0.1130
********* Velocity Refinement Round 8 *********
Stage 2: Early Stop Triggered at round 7.
********* Finished. Total Time = 0 h : 0 m : 28 s *********
Final: Train ELBO = -11489.504, Test ELBO = -11277.100
Estimating ODE parameters...
Detected 35 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
(0.36, 0.25070307343442405), (0.64, 0.7640605172671028)
(0.57, 0.8475524475473406), (0.43, 0.10044955044863953)
KS-test result: [0. 0. 1.]
Initial induction: 62, repression: 22/84
Learning Rate based on Data Sparsity: 0.0000
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 2, test iteration: 2
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.030
Average Set Size: 16
********* Round 1: Early Stop Triggered at epoch 1349. *********
Change in noise variance: 0.2281
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 1522. *********
Change in noise variance: 0.0052
Change in x0: 0.6921
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 1663. *********
Change in noise variance: 0.0064
Change in x0: 0.3447
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 1710. *********
Change in noise variance: 0.0064
Change in x0: 0.3396
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 1734. *********
Change in noise variance: 0.0022
Change in x0: 0.2616
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 1783. *********
Change in noise variance: 0.0020
Change in x0: 0.2183
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 1799. *********
Change in noise variance: 0.0013
Change in x0: 0.2125
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 1835. *********
Change in noise variance: 0.0007
Change in x0: 0.2014
********* Velocity Refinement Round 9 *********
********* Round 9: Early Stop Triggered at epoch 1860. *********
Change in noise variance: 0.0000
Change in x0: 0.1965
********* Velocity Refinement Round 10 *********
Stage 2: Early Stop Triggered at round 9.
********* Finished. Total Time = 0 h : 1 m : 32 s *********
Final: Train ELBO = -1606.755, Test ELBO = -1609.287
Estimating ODE parameters...
Detected 32 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
(0.62, 0.25073704656682283), (0.38, 0.8864542002036675)
KS-test result: [0. 1. 1.]
Initial induction: 44, repression: 27/71
Learning Rate based on Data Sparsity: 0.0000
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 2, test iteration: 2
********* Stage 1: Early Stop Triggered at epoch 6. *********
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.037
Average Set Size: 17
********* Round 1: Early Stop Triggered at epoch 212. *********
Change in noise variance: 0.7725
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 245. *********
Change in noise variance: 0.0427
Change in x0: 0.6257
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 254. *********
Change in noise variance: 0.0046
Change in x0: 0.3750
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 260. *********
Change in noise variance: 0.0027
Change in x0: 0.3104
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 268. *********
Change in noise variance: 0.0009
Change in x0: 0.2714
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 274. *********
Change in noise variance: 0.0000
Change in x0: 0.2441
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 280. *********
Change in noise variance: 0.0000
Change in x0: 0.2225
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 286. *********
Change in noise variance: 0.0000
Change in x0: 0.2113
********* Velocity Refinement Round 9 *********
********* Round 9: Early Stop Triggered at epoch 292. *********
Change in noise variance: 0.0000
Change in x0: 0.2111
********* Velocity Refinement Round 10 *********
Stage 2: Early Stop Triggered at round 9.
********* Finished. Total Time = 0 h : 0 m : 27 s *********
Final: Train ELBO = -14000.091, Test ELBO = -14185.506
Estimating ODE parameters...
Detected 40 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
(0.50, 0.7798950426379351), (0.50, 0.36970058925632504)
KS-test result: [0. 1. 1.]
Initial induction: 68, repression: 17/85
Learning Rate based on Data Sparsity: 0.0000
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 2, test iteration: 2
********* Stage 1: Early Stop Triggered at epoch 8. *********
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.060
Average Set Size: 11
********* Round 1: Early Stop Triggered at epoch 14. *********
Change in noise variance: 1.2757
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 20. *********
Change in noise variance: 0.0801
Change in x0: 0.4768
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 26. *********
Change in noise variance: 0.0039
Change in x0: 0.2874
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 46. *********
Change in noise variance: 0.0015
Change in x0: 0.2370
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 52. *********
Change in noise variance: 0.0008
Change in x0: 0.2077
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 59. *********
Change in noise variance: 0.0000
Change in x0: 0.1950
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 65. *********
Change in noise variance: 0.0000
Change in x0: 0.1838
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 71. *********
Change in noise variance: 0.0000
Change in x0: 0.1655
********* Velocity Refinement Round 9 *********
********* Round 9: Early Stop Triggered at epoch 79. *********
Change in noise variance: 0.0000
Change in x0: 0.1513
********* Velocity Refinement Round 10 *********
********* Round 10: Early Stop Triggered at epoch 85. *********
Change in noise variance: 0.0000
Change in x0: 0.1465
********* Velocity Refinement Round 11 *********
Stage 2: Early Stop Triggered at round 10.
********* Finished. Total Time = 0 h : 0 m : 14 s *********
Final: Train ELBO = -20587.127, Test ELBO = -20604.510
Estimating ODE parameters...
Detected 31 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
(0.61, 0.25328421223351666), (0.39, 0.8536592818877626)
(0.50, 0.13934589145383208), (0.50, 0.7024048173998951)
KS-test result: [0. 1. 0.]
Initial induction: 43, repression: 38/81
Learning Rate based on Data Sparsity: 0.0000
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 2, test iteration: 2
********* Stage 1: Early Stop Triggered at epoch 5. *********
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.033
Average Set Size: 16
********* Round 1: Early Stop Triggered at epoch 11. *********
Change in noise variance: 0.3671
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 16. *********
Change in noise variance: 0.0477
Change in x0: 0.6078
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 22. *********
Change in noise variance: 0.0024
Change in x0: 0.3214
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 28. *********
Change in noise variance: 0.0022
Change in x0: 0.2438
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 34. *********
Change in noise variance: 0.0012
Change in x0: 0.2155
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 40. *********
Change in noise variance: 0.0008
Change in x0: 0.1997
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 46. *********
Change in noise variance: 0.0000
Change in x0: 0.1880
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 52. *********
Change in noise variance: 0.0000
Change in x0: 0.1739
********* Velocity Refinement Round 9 *********
********* Round 9: Early Stop Triggered at epoch 58. *********
Change in noise variance: 0.0000
Change in x0: 0.1664
********* Velocity Refinement Round 10 *********
Stage 2: Early Stop Triggered at round 9.
********* Finished. Total Time = 0 h : 0 m : 5 s *********
Final: Train ELBO = -14216.592, Test ELBO = -14286.768
Estimating ODE parameters...
Detected 31 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
KS-test result: [1. 1. 1.]
Assign cluster 0 to repressive
Initial induction: 42, repression: 31/73
Learning Rate based on Data Sparsity: 0.0000
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 2, test iteration: 2
********* Stage 1: Early Stop Triggered at epoch 13. *********
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.030
Average Set Size: 17
********* Round 1: Early Stop Triggered at epoch 19. *********
Change in noise variance: 0.2287
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 26. *********
Change in noise variance: 0.0118
Change in x0: 0.4770
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 141. *********
Change in noise variance: 0.0030
Change in x0: 0.3930
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 191. *********
Change in noise variance: 0.0022
Change in x0: 0.3111
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 221. *********
Change in noise variance: 0.0012
Change in x0: 0.2692
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 262. *********
Change in noise variance: 0.0011
Change in x0: 0.2309
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 274. *********
Change in noise variance: 0.0009
Change in x0: 0.1985
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 290. *********
Change in noise variance: 0.0000
Change in x0: 0.1709
********* Velocity Refinement Round 9 *********
********* Round 9: Early Stop Triggered at epoch 295. *********
Change in noise variance: 0.0000
Change in x0: 0.1484
********* Velocity Refinement Round 10 *********
********* Round 10: Early Stop Triggered at epoch 305. *********
Change in noise variance: 0.0000
Change in x0: 0.1350
********* Velocity Refinement Round 11 *********
********* Round 11: Early Stop Triggered at epoch 314. *********
Change in noise variance: 0.0000
Change in x0: 0.1269
********* Velocity Refinement Round 12 *********
Stage 2: Early Stop Triggered at round 11.
********* Finished. Total Time = 0 h : 0 m : 27 s *********
Final: Train ELBO = -14776.117, Test ELBO = -14653.228
Estimating ODE parameters...
Detected 32 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
(0.62, 0.20672961138115753), (0.38, 0.8812856125296097)
(0.47, 0.15060377505382813), (0.53, 0.8300993269683058)
(0.55, 0.05115333441901289), (0.45, 0.8035111864172376)
KS-test result: [0. 0. 0.]
Initial induction: 37, repression: 48/85
Learning Rate based on Data Sparsity: 0.0000
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 2, test iteration: 2
********* Stage 1: Early Stop Triggered at epoch 16. *********
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.083
Average Set Size: 15
********* Round 1: Early Stop Triggered at epoch 22. *********
Change in noise variance: 0.2995
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 27. *********
Change in noise variance: 0.0712
Change in x0: 0.4024
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 32. *********
Change in noise variance: 0.0131
Change in x0: 0.3061
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 38. *********
Change in noise variance: 0.0088
Change in x0: 0.2449
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 44. *********
Change in noise variance: 0.0033
Change in x0: 0.2076
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 50. *********
Change in noise variance: 0.0013
Change in x0: 0.1996
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 56. *********
Change in noise variance: 0.0009
Change in x0: 0.1879
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 61. *********
Change in noise variance: 0.0000
Change in x0: 0.1675
********* Velocity Refinement Round 9 *********
********* Round 9: Early Stop Triggered at epoch 66. *********
Change in noise variance: 0.0000
Change in x0: 0.1580
********* Velocity Refinement Round 10 *********
Stage 2: Early Stop Triggered at round 9.
********* Finished. Total Time = 0 h : 0 m : 15 s *********
Final: Train ELBO = -16159.053, Test ELBO = -16439.312
Estimating ODE parameters...
Detected 30 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
KS-test result: [1. 1. 1.]
Assign cluster 2 to repressive
Initial induction: 48, repression: 25/73
Learning Rate based on Data Sparsity: 0.0000
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 2, test iteration: 2
********* Stage 1: Early Stop Triggered at epoch 5. *********
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.023
Average Set Size: 17
********* Round 1: Early Stop Triggered at epoch 11. *********
Change in noise variance: 0.3156
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 16. *********
Change in noise variance: 0.0792
Change in x0: 0.4805
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 22. *********
Change in noise variance: 0.0043
Change in x0: 0.3847
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 28. *********
Change in noise variance: 0.0088
Change in x0: 0.2860
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 34. *********
Change in noise variance: 0.0124
Change in x0: 0.2093
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 40. *********
Change in noise variance: 0.0065
Change in x0: 0.1651
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 46. *********
Change in noise variance: 0.0027
Change in x0: 0.1482
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 52. *********
Change in noise variance: 0.0047
Change in x0: 0.1399
********* Velocity Refinement Round 9 *********
********* Round 9: Early Stop Triggered at epoch 58. *********
Change in noise variance: 0.0033
Change in x0: 0.1331
********* Velocity Refinement Round 10 *********
********* Round 10: Early Stop Triggered at epoch 64. *********
Change in noise variance: 0.0020
Change in x0: 0.1280
********* Velocity Refinement Round 11 *********
********* Round 11: Early Stop Triggered at epoch 70. *********
Change in noise variance: 0.0014
Change in x0: 0.1245
********* Velocity Refinement Round 12 *********
********* Round 12: Early Stop Triggered at epoch 76. *********
Change in noise variance: 0.0010
Change in x0: 0.1219
********* Velocity Refinement Round 13 *********
Stage 2: Early Stop Triggered at round 12.
********* Finished. Total Time = 0 h : 0 m : 19 s *********
Final: Train ELBO = -14586.744, Test ELBO = -14794.264
Estimating ODE parameters...
Detected 33 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
(0.59, 0.3174184423133425), (0.41, 0.8289720081868958)
KS-test result: [0. 2. 1.]
Initial induction: 33, repression: 41/74
Learning Rate based on Data Sparsity: 0.0000
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 2, test iteration: 2
********* Stage 1: Early Stop Triggered at epoch 8. *********
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.025
Average Set Size: 11
********* Round 1: Early Stop Triggered at epoch 15. *********
Change in noise variance: 1.1601
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 249. *********
Change in noise variance: 0.0809
Change in x0: 0.5851
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 386. *********
Change in noise variance: 0.0074
Change in x0: 0.3659
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 429. *********
Change in noise variance: 0.0025
Change in x0: 0.3050
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 477. *********
Change in noise variance: 0.0010
Change in x0: 0.2527
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 493. *********
Change in noise variance: 0.0000
Change in x0: 0.2137
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 509. *********
Change in noise variance: 0.0000
Change in x0: 0.1961
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 533. *********
Change in noise variance: 0.0000
Change in x0: 0.1938
********* Velocity Refinement Round 9 *********
Stage 2: Early Stop Triggered at round 8.
********* Finished. Total Time = 0 h : 0 m : 31 s *********
Final: Train ELBO = -14918.545, Test ELBO = -16750.279
Estimating ODE parameters...
Detected 32 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
(0.48, 0.14413919413760584), (0.52, 0.9094301356860764)
KS-test result: [1. 0. 1.]
Initial induction: 60, repression: 17/77
Learning Rate based on Data Sparsity: 0.0000
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 2, test iteration: 2
********* Stage 1: Early Stop Triggered at epoch 5. *********
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.017
Average Set Size: 17
********* Round 1: Early Stop Triggered at epoch 151. *********
Change in noise variance: 0.6006
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 291. *********
Change in noise variance: 0.0873
Change in x0: 0.5995
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 319. *********
Change in noise variance: 0.0039
Change in x0: 0.5194
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 325. *********
Change in noise variance: 0.0036
Change in x0: 0.3645
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 331. *********
Change in noise variance: 0.0024
Change in x0: 0.2908
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 337. *********
Change in noise variance: 0.0017
Change in x0: 0.2739
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 343. *********
Change in noise variance: 0.0016
Change in x0: 0.2460
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 349. *********
Change in noise variance: 0.0015
Change in x0: 0.1965
********* Velocity Refinement Round 9 *********
********* Round 9: Early Stop Triggered at epoch 355. *********
Change in noise variance: 0.0008
Change in x0: 0.1640
********* Velocity Refinement Round 10 *********
********* Round 10: Early Stop Triggered at epoch 361. *********
Change in noise variance: 0.0000
Change in x0: 0.1623
********* Velocity Refinement Round 11 *********
Stage 2: Early Stop Triggered at round 10.
********* Finished. Total Time = 0 h : 0 m : 26 s *********
Final: Train ELBO = -15776.562, Test ELBO = -15699.438
Estimating ODE parameters...
Detected 32 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
KS-test result: [1. 1. 1.]
Assign cluster 1 to repressive
Initial induction: 43, repression: 26/69
Learning Rate based on Data Sparsity: 0.0000
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 2, test iteration: 2
********* Stage 1: Early Stop Triggered at epoch 7. *********
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.017
Average Set Size: 17
********* Round 1: Early Stop Triggered at epoch 13. *********
Change in noise variance: 0.3804
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 18. *********
Change in noise variance: 0.0383
Change in x0: 0.4629
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 63. *********
Change in noise variance: 0.0023
Change in x0: 0.3529
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 86. *********
Change in noise variance: 0.0020
Change in x0: 0.2845
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 101. *********
Change in noise variance: 0.0012
Change in x0: 0.2306
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 107. *********
Change in noise variance: 0.0006
Change in x0: 0.1938
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 113. *********
Change in noise variance: 0.0000
Change in x0: 0.1721
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 119. *********
Change in noise variance: 0.0000
Change in x0: 0.1692
********* Velocity Refinement Round 9 *********
Stage 2: Early Stop Triggered at round 8.
********* Finished. Total Time = 0 h : 0 m : 21 s *********
Final: Train ELBO = -15506.851, Test ELBO = -15114.594
Estimating ODE parameters...
Detected 43 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
KS-test result: [1. 1. 1.]
Assign cluster 1 to repressive
Initial induction: 55, repression: 24/79
Learning Rate based on Data Sparsity: 0.0000
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 2, test iteration: 2
********* Stage 1: Early Stop Triggered at epoch 7. *********
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.030
Average Set Size: 17
Change in noise variance: 0.3149
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 536. *********
Change in noise variance: 0.0409
Change in x0: 0.5604
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 549. *********
Change in noise variance: 0.0115
Change in x0: 0.3583
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 573. *********
Change in noise variance: 0.0167
Change in x0: 0.2715
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 596. *********
Change in noise variance: 0.0089
Change in x0: 0.2361
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 634. *********
Change in noise variance: 0.0026
Change in x0: 0.2195
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 677. *********
Change in noise variance: 0.0054
Change in x0: 0.2167
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 706. *********
Change in noise variance: 0.0046
Change in x0: 0.2121
********* Velocity Refinement Round 9 *********
********* Round 9: Early Stop Triggered at epoch 753. *********
Change in noise variance: 0.0030
Change in x0: 0.1999
********* Velocity Refinement Round 10 *********
********* Round 10: Early Stop Triggered at epoch 767. *********
Change in noise variance: 0.0017
Change in x0: 0.1882
********* Velocity Refinement Round 11 *********
********* Round 11: Early Stop Triggered at epoch 811. *********
Change in noise variance: 0.0009
Change in x0: 0.1820
********* Velocity Refinement Round 12 *********
Stage 2: Early Stop Triggered at round 11.
********* Finished. Total Time = 0 h : 0 m : 46 s *********
Final: Train ELBO = -17525.242, Test ELBO = -17494.643
Estimating ODE parameters...
Detected 28 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
KS-test result: [1. 1. 1.]
Assign cluster 2 to repressive
Initial induction: 46, repression: 16/62
Learning Rate based on Data Sparsity: 0.0000
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 2, test iteration: 2
********* Stage 1: Early Stop Triggered at epoch 5. *********
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.043
Average Set Size: 16
********* Round 1: Early Stop Triggered at epoch 11. *********
Change in noise variance: 0.2553
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 16. *********
Change in noise variance: 0.0305
Change in x0: 0.6486
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 22. *********
Change in noise variance: 0.0164
Change in x0: 0.3742
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 28. *********
Change in noise variance: 0.0181
Change in x0: 0.2555
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 33. *********
Change in noise variance: 0.0008
Change in x0: 0.2440
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 38. *********
Change in noise variance: 0.0000
Change in x0: 0.2472
********* Velocity Refinement Round 7 *********
Stage 2: Early Stop Triggered at round 6.
********* Finished. Total Time = 0 h : 0 m : 15 s *********
Final: Train ELBO = -20341.797, Test ELBO = -20511.984
Estimating ODE parameters...
Detected 25 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
(0.44, 0.7659555497733345), (0.56, 0.1878343494524111)
KS-test result: [1. 0. 1.]
Initial induction: 44, repression: 20/64
Learning Rate based on Data Sparsity: 0.0000
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 2, test iteration: 2
********* Stage 1: Early Stop Triggered at epoch 5. *********
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.003
Average Set Size: 16
********* Round 1: Early Stop Triggered at epoch 166. *********
Change in noise variance: 0.7913
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 282. *********
Change in noise variance: 0.0707
Change in x0: 0.6809
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 291. *********
Change in noise variance: 0.0609
Change in x0: 0.3917
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 306. *********
Change in noise variance: 0.0193
Change in x0: 0.3149
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 313. *********
Change in noise variance: 0.0160
Change in x0: 0.2764
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 323. *********
Change in noise variance: 0.0042
Change in x0: 0.2416
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 332. *********
Change in noise variance: 0.0018
Change in x0: 0.2053
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 340. *********
Change in noise variance: 0.0007
Change in x0: 0.2041
********* Velocity Refinement Round 9 *********
Stage 2: Early Stop Triggered at round 8.
********* Finished. Total Time = 0 h : 0 m : 27 s *********
Final: Train ELBO = -15492.789, Test ELBO = -15448.032
Estimating ODE parameters...
Detected 84 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
KS-test result: [1. 1. 1.]
Assign cluster 1 to repressive
Initial induction: 84, repression: 66/150
Learning Rate based on Data Sparsity: 0.0000
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 6, test iteration: 10
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.029
Average Set Size: 20
********* Round 1: Early Stop Triggered at epoch 1009. *********
Change in noise variance: 0.1517
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 1018. *********
Change in noise variance: 0.0034
Change in x0: 0.6934
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 1027. *********
Change in noise variance: 0.0060
Change in x0: 0.5286
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 1034. *********
Change in noise variance: 0.0179
Change in x0: 0.4225
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 1041. *********
Change in noise variance: 0.0064
Change in x0: 0.3509
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 1048. *********
Change in noise variance: 0.0042
Change in x0: 0.3036
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 1055. *********
Change in noise variance: 0.0019
Change in x0: 0.2672
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 1062. *********
Change in noise variance: 0.0008
Change in x0: 0.2463
********* Velocity Refinement Round 9 *********
********* Round 9: Early Stop Triggered at epoch 1069. *********
Change in noise variance: 0.0000
Change in x0: 0.2563
********* Velocity Refinement Round 10 *********
Stage 2: Early Stop Triggered at round 9.
********* Finished. Total Time = 0 h : 3 m : 8 s *********
Final: Train ELBO = -290.949, Test ELBO = -311.508
Estimating ODE parameters...
Detected 35 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
(0.38, 0.18558028707247598), (0.62, 0.7691498527847566)
KS-test result: [1. 0. 1.]
Initial induction: 68, repression: 13/81
Learning Rate based on Data Sparsity: 0.0000
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 2, test iteration: 2
********* Stage 1: Early Stop Triggered at epoch 9. *********
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.033
Average Set Size: 16
********* Round 1: Early Stop Triggered at epoch 15. *********
Change in noise variance: 0.2443
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 20. *********
Change in noise variance: 0.0150
Change in x0: 0.6972
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 26. *********
Change in noise variance: 0.0030
Change in x0: 0.4751
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 32. *********
Change in noise variance: 0.0016
Change in x0: 0.3783
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 38. *********
Change in noise variance: 0.0041
Change in x0: 0.2802
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 44. *********
Change in noise variance: 0.0025
Change in x0: 0.2150
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 50. *********
Change in noise variance: 0.0008
Change in x0: 0.1977
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 55. *********
Change in noise variance: 0.0000
Change in x0: 0.1902
********* Velocity Refinement Round 9 *********
Stage 2: Early Stop Triggered at round 8.
********* Finished. Total Time = 0 h : 0 m : 18 s *********
Final: Train ELBO = -20595.697, Test ELBO = -20799.051
Estimating ODE parameters...
Detected 33 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
KS-test result: [1. 1. 1.]
Assign cluster 2 to repressive
Initial induction: 47, repression: 32/79
Learning Rate based on Data Sparsity: 0.0000
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 2, test iteration: 2
********* Stage 1: Early Stop Triggered at epoch 453. *********
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.023
Average Set Size: 23
Change in noise variance: 0.1903
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 1136. *********
Change in noise variance: 0.1648
Change in x0: 1.0268
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 1187. *********
Change in noise variance: 0.0067
Change in x0: 0.6616
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 1192. *********
Change in noise variance: 0.0038
Change in x0: 0.6284
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 1366. *********
Change in noise variance: 0.0058
Change in x0: 0.4408
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 1535. *********
Change in noise variance: 0.0065
Change in x0: 0.5481
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 1572. *********
Change in noise variance: 0.0043
Change in x0: 0.5013
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 1655. *********
Change in noise variance: 0.0040
Change in x0: 0.3870
********* Velocity Refinement Round 9 *********
********* Round 9: Early Stop Triggered at epoch 1776. *********
Change in noise variance: 0.0031
Change in x0: 0.3702
********* Velocity Refinement Round 10 *********
********* Round 10: Early Stop Triggered at epoch 1863. *********
Change in noise variance: 0.0032
Change in x0: 0.3509
********* Velocity Refinement Round 11 *********
********* Round 11: Early Stop Triggered at epoch 1870. *********
Change in noise variance: 0.0024
Change in x0: 0.3240
********* Velocity Refinement Round 12 *********
********* Round 12: Early Stop Triggered at epoch 2058. *********
Change in noise variance: 0.0026
Change in x0: 0.2263
********* Velocity Refinement Round 13 *********
********* Round 13: Early Stop Triggered at epoch 2109. *********
Change in noise variance: 0.0024
Change in x0: 0.4064
********* Velocity Refinement Round 14 *********
********* Round 14: Early Stop Triggered at epoch 2201. *********
Change in noise variance: 0.0033
Change in x0: 0.2971
********* Velocity Refinement Round 15 *********
********* Round 15: Early Stop Triggered at epoch 2293. *********
Change in noise variance: 0.0032
Change in x0: 0.3209
********* Velocity Refinement Round 16 *********
********* Round 16: Early Stop Triggered at epoch 2486. *********
Change in noise variance: 0.0024
Change in x0: 0.2835
********* Velocity Refinement Round 17 *********
********* Round 17: Early Stop Triggered at epoch 2587. *********
Change in noise variance: 0.0022
Change in x0: 0.3458
********* Velocity Refinement Round 18 *********
********* Round 18: Early Stop Triggered at epoch 2739. *********
Change in noise variance: 0.0021
Change in x0: 0.2973
********* Velocity Refinement Round 19 *********
********* Round 19: Early Stop Triggered at epoch 2796. *********
Change in noise variance: 0.0018
Change in x0: 0.2869
********* Velocity Refinement Round 20 *********
********* Round 20: Early Stop Triggered at epoch 2940. *********
Change in noise variance: 0.0000
Change in x0: 0.3014
********* Finished. Total Time = 0 h : 2 m : 4 s *********
Final: Train ELBO = -1905.243, Test ELBO = -2021.629
Estimating ODE parameters...
Detected 32 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
KS-test result: [1. 1. 1.]
Assign cluster 1 to repressive
Initial induction: 48, repression: 23/71
Learning Rate based on Data Sparsity: 0.0000
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 3, test iteration: 4
********* Stage 1: Early Stop Triggered at epoch 6. *********
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.025
Average Set Size: 20
********* Round 1: Early Stop Triggered at epoch 13. *********
Change in noise variance: 0.1940
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 20. *********
Change in noise variance: 0.0185
Change in x0: 0.7663
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 27. *********
Change in noise variance: 0.0148
Change in x0: 0.4259
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 34. *********
Change in noise variance: 0.0021
Change in x0: 0.3094
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 41. *********
Change in noise variance: 0.0039
Change in x0: 0.2825
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 48. *********
Change in noise variance: 0.0031
Change in x0: 0.2522
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 55. *********
Change in noise variance: 0.0025
Change in x0: 0.2267
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 62. *********
Change in noise variance: 0.0019
Change in x0: 0.2150
********* Velocity Refinement Round 9 *********
********* Round 9: Early Stop Triggered at epoch 69. *********
Change in noise variance: 0.0013
Change in x0: 0.2022
********* Velocity Refinement Round 10 *********
********* Round 10: Early Stop Triggered at epoch 76. *********
Change in noise variance: 0.0008
Change in x0: 0.1883
********* Velocity Refinement Round 11 *********
********* Round 11: Early Stop Triggered at epoch 82. *********
Change in noise variance: 0.0000
Change in x0: 0.1784
********* Velocity Refinement Round 12 *********
Stage 2: Early Stop Triggered at round 11.
********* Finished. Total Time = 0 h : 0 m : 10 s *********
Final: Train ELBO = -14711.658, Test ELBO = -15284.556
Estimating ODE parameters...
Detected 34 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
KS-test result: [1. 1. 1.]
Assign cluster 2 to repressive
Initial induction: 46, repression: 23/69
Learning Rate based on Data Sparsity: 0.0000
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 2, test iteration: 2
********* Stage 1: Early Stop Triggered at epoch 8. *********
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.050
Average Set Size: 16
********* Round 1: Early Stop Triggered at epoch 14. *********
Change in noise variance: 0.2069
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 20. *********
Change in noise variance: 0.0319
Change in x0: 0.5998
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 26. *********
Change in noise variance: 0.0025
Change in x0: 0.4039
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 32. *********
Change in noise variance: 0.0012
Change in x0: 0.3299
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 38. *********
Change in noise variance: 0.0007
Change in x0: 0.2737
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 44. *********
Change in noise variance: 0.0000
Change in x0: 0.2441
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 50. *********
Change in noise variance: 0.0000
Change in x0: 0.2284
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 56. *********
Change in noise variance: 0.0000
Change in x0: 0.2224
********* Velocity Refinement Round 9 *********
Stage 2: Early Stop Triggered at round 8.
********* Finished. Total Time = 0 h : 0 m : 18 s *********
Final: Train ELBO = -15914.265, Test ELBO = -15941.758
Estimating ODE parameters...
Detected 36 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
(0.46, 0.815320187953141), (0.54, 0.18212739641125106)
(0.61, 0.2646005089450621), (0.39, 0.871184004409053)
KS-test result: [0. 1. 0.]
Initial induction: 34, repression: 42/76
Learning Rate based on Data Sparsity: 0.0000
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 2, test iteration: 2
********* Stage 1: Early Stop Triggered at epoch 9. *********
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.027
Average Set Size: 17
********* Round 1: Early Stop Triggered at epoch 15. *********
Change in noise variance: 0.8975
********* Velocity Refinement Round 2 *********
********* Round 2: Early Stop Triggered at epoch 362. *********
Change in noise variance: 0.0808
Change in x0: 0.2958
********* Velocity Refinement Round 3 *********
********* Round 3: Early Stop Triggered at epoch 405. *********
Change in noise variance: 0.0011
Change in x0: 0.2908
********* Velocity Refinement Round 4 *********
********* Round 4: Early Stop Triggered at epoch 425. *********
Change in noise variance: 0.0007
Change in x0: 0.2674
********* Velocity Refinement Round 5 *********
********* Round 5: Early Stop Triggered at epoch 432. *********
Change in noise variance: 0.0000
Change in x0: 0.2239
********* Velocity Refinement Round 6 *********
********* Round 6: Early Stop Triggered at epoch 475. *********
Change in noise variance: 0.0000
Change in x0: 0.1955
********* Velocity Refinement Round 7 *********
********* Round 7: Early Stop Triggered at epoch 482. *********
Change in noise variance: 0.0000
Change in x0: 0.1796
********* Velocity Refinement Round 8 *********
********* Round 8: Early Stop Triggered at epoch 492. *********
Change in noise variance: 0.0000
Change in x0: 0.1643
********* Velocity Refinement Round 9 *********
********* Round 9: Early Stop Triggered at epoch 502. *********
Change in noise variance: 0.0000
Change in x0: 0.1513
********* Velocity Refinement Round 10 *********
********* Round 10: Early Stop Triggered at epoch 522. *********
Change in noise variance: 0.0000
Change in x0: 0.1452
********* Velocity Refinement Round 11 *********
Stage 2: Early Stop Triggered at round 10.
********* Finished. Total Time = 0 h : 0 m : 35 s *********
Final: Train ELBO = -18069.350, Test ELBO = -17271.164
Estimating ODE parameters...
Detected 25 velocity genes.
Estimating the variance...
Initialization using the steady-state and dynamical models.
Reinitialize the regular ODE parameters based on estimated global latent time.
3 clusters detected based on gene co-expression.
(0.41, 0.08570857886357829), (0.59, 0.9175531644155611)
KS-test result: [2. 2. 0.]
Initial induction: 23, repression: 49/72
Learning Rate based on Data Sparsity: 0.0000
--------------------------- Train a VeloVAE ---------------------------
********* Creating Training/Validation Datasets *********
********* Finished. *********
********* Creating optimizers *********
********* Finished. *********
********* Start training *********
********* Stage 1 *********
Total Number of Iterations Per Epoch: 2, test iteration: 2
********* Stage 1: Early Stop Triggered at epoch 5. *********
********* Stage 2 *********
********* Velocity Refinement Round 1 *********
Percentage of Invalid Sets: 0.033
Average Set Size: 16
********* Round 1: Early Stop Triggered at epoch 251. *********
Change in noise variance: 0.3936
********* Velocity Refinement Round 2 *********
Data saving#
if SAVE_DATA:
pd.DataFrame({"velocity": velocity_correlation}).to_parquet(
path=DATA_DIR / DATASET / "results" / "velovae_vae_correlation.parquet"
)