UNOMT Torch batch_size=64 MPI [PC]
vibhatha opened this issue · comments
Vibhatha Lakmal Abeykoon commented
================================================================================
Training Epoch 1/ 5:
Drug Weighted QED Regression Loss: 0.028109
Drug Weighted QED Regression Loss: 0.028109
Drug Weighted QED Regression Loss: 0.028109
Drug Weighted QED Regression Loss: 0.028109
Drug Response Regression Loss: 5668.19
Epoch Running Time: 277.0 Seconds.
================================================================================
Training Epoch 2/ 5:
Drug Response Regression Loss: 5088.51
Epoch Running Time: 277.0 Seconds.
================================================================================
Training Epoch 2/ 5:
Drug Response Regression Loss: 5564.22
Epoch Running Time: 277.2 Seconds.
================================================================================
Training Epoch 2/ 5:
Drug Response Regression Loss: 5696.25
Epoch Running Time: 277.3 Seconds.
================================================================================
Training Epoch 2/ 5:
Drug Weighted QED Regression Loss: 0.019153
Drug Weighted QED Regression Loss: 0.018326
Drug Weighted QED Regression Loss: 0.019266
Drug Weighted QED Regression Loss: 0.018711
Drug Response Regression Loss: 3556.82
Epoch Running Time: 278.8 Seconds.
================================================================================
Training Epoch 3/ 5:
Drug Response Regression Loss: 3547.73
Epoch Running Time: 279.0 Seconds.
================================================================================
Training Epoch 3/ 5:
Drug Response Regression Loss: 4032.30
Epoch Running Time: 278.9 Seconds.
================================================================================
Training Epoch 3/ 5:
Drug Response Regression Loss: 3701.63
Epoch Running Time: 279.0 Seconds.
================================================================================
Training Epoch 3/ 5:
Drug Weighted QED Regression Loss: 0.014881
Drug Weighted QED Regression Loss: 0.013512
Drug Weighted QED Regression Loss: 0.015578
Drug Weighted QED Regression Loss: 0.014847
Drug Response Regression Loss: 3292.04
Drug Response Regression Loss: 3357.36
Drug Response Regression Loss: 3516.86
Drug Response Regression Loss: 3435.00
Drug Target Family Classification Accuracy: 31.37%
Drug Target Family Classification Accuracy: 7.84%
Drug Target Family Classification Accuracy: 35.29%
Drug Target Family Classification Accuracy: 9.80%
Drug Weighted QED Regression
MSE: 0.012924 MAE: 0.088674 R2: +0.65
Drug Response Regression:
Drug Weighted QED Regression
MSE: 0.012924 MAE: 0.088674 R2: +0.65
Drug Response Regression:
Drug Weighted QED Regression
MSE: 0.012924 MAE: 0.088674 R2: +0.65
Drug Response Regression:
Drug Weighted QED Regression
MSE: 0.012924 MAE: 0.088674 R2: +0.65
Drug Response Regression:
gCSI MSE: 3156.50 MAE: 43.79 R2: +0.19
gCSI MSE: 3130.61 MAE: 42.98 R2: +0.22
gCSI MSE: 3140.28 MAE: 43.38 R2: +0.21
gCSI MSE: 3228.83 MAE: 44.35 R2: +0.20
CCLE MSE: 3475.34 MAE: 45.41 R2: +0.15
Epoch Running Time: 876.4 Seconds.
================================================================================
Training Epoch 4/ 5:
CCLE MSE: 3154.87 MAE: 43.42 R2: +0.15
Epoch Running Time: 885.1 Seconds.
================================================================================
Training Epoch 4/ 5:
CCLE MSE: 3390.12 MAE: 44.87 R2: +0.14
Epoch Running Time: 886.7 Seconds.
================================================================================
Training Epoch 4/ 5:
CCLE MSE: 3435.81 MAE: 45.53 R2: +0.16
Epoch Running Time: 892.9 Seconds.
================================================================================
Training Epoch 4/ 5:
Drug Weighted QED Regression Loss: 0.008205
Drug Weighted QED Regression Loss: 0.008672
Drug Weighted QED Regression Loss: 0.007700
Drug Weighted QED Regression Loss: 0.008957
Drug Response Regression Loss: 3452.33
Drug Response Regression Loss: 3345.76
Drug Response Regression Loss: 3405.05
Drug Response Regression Loss: 3489.79
Drug Target Family Classification Accuracy: 9.80%
Drug Target Family Classification Accuracy: 19.61%
Drug Target Family Classification Accuracy: 23.53%
Drug Target Family Classification Accuracy: 27.45%
Drug Weighted QED Regression
MSE: 0.010913 MAE: 0.082061 R2: +0.70
Drug Response Regression:
Drug Weighted QED Regression
MSE: 0.010913 MAE: 0.082061 R2: +0.70
Drug Response Regression:
Drug Weighted QED Regression
MSE: 0.010913 MAE: 0.082061 R2: +0.70
Drug Response Regression:
Drug Weighted QED Regression
MSE: 0.010913 MAE: 0.082061 R2: +0.70
Drug Response Regression:
gCSI MSE: 3292.33 MAE: 48.25 R2: +0.15
gCSI MSE: 3221.04 MAE: 47.73 R2: +0.20
gCSI MSE: 3249.74 MAE: 47.59 R2: +0.19
gCSI MSE: 3317.77 MAE: 48.60 R2: +0.18
CCLE MSE: 3733.37 MAE: 51.38 R2: +0.09
Epoch Running Time: 891.5 Seconds.
================================================================================
Training Epoch 5/ 5:
CCLE MSE: 3522.12 MAE: 50.07 R2: +0.05
Epoch Running Time: 891.4 Seconds.
================================================================================
Training Epoch 5/ 5:
CCLE MSE: 3659.17 MAE: 51.00 R2: +0.07
Epoch Running Time: 891.3 Seconds.
================================================================================
Training Epoch 5/ 5:
CCLE MSE: 3728.04 MAE: 51.97 R2: +0.08
Epoch Running Time: 891.3 Seconds.
================================================================================
Training Epoch 5/ 5:
Drug Weighted QED Regression Loss: 0.006902
Drug Weighted QED Regression Loss: 0.007250
Drug Weighted QED Regression Loss: 0.006807
Drug Weighted QED Regression Loss: 0.006750
Drug Response Regression Loss: 2791.31
Drug Response Regression Loss: 3153.10
Drug Response Regression Loss: 3095.93
Drug Response Regression Loss: 3294.72
Drug Target Family Classification Accuracy: 37.25%
Drug Target Family Classification Accuracy: 39.22%
Drug Target Family Classification Accuracy: 7.84%
Drug Target Family Classification Accuracy: 11.76%
Drug Weighted QED Regression
MSE: 0.009588 MAE: 0.075388 R2: +0.74
Drug Response Regression:
Drug Weighted QED Regression
MSE: 0.009588 MAE: 0.075388 R2: +0.74
Drug Response Regression:
Drug Weighted QED Regression
MSE: 0.009588 MAE: 0.075388 R2: +0.74
Drug Response Regression:
Drug Weighted QED Regression
MSE: 0.009588 MAE: 0.075388 R2: +0.74
Drug Response Regression:
gCSI MSE: 3098.10 MAE: 45.38 R2: +0.20
gCSI MSE: 3015.94 MAE: 44.90 R2: +0.25
gCSI MSE: 3019.80 MAE: 44.59 R2: +0.24
gCSI MSE: 3083.02 MAE: 45.58 R2: +0.23
CCLE MSE: 3547.79 MAE: 48.90 R2: +0.14
Epoch Running Time: 889.9 Seconds.
/home/vibhatha/sandbox/UNO/Benchmarks/Pilot1/UnoMT/utils/datasets/drug_resp_dataset.py:243: UserWarning: Heterogeneous Cylon Table Detected!. Use Numpy operations with Caution.
self.__drug_resp_array = self.__drug_resp_tb.to_numpy(zero_copy_only=False)
/home/vibhatha/venv/ENVTORCH/lib/python3.8/site-packages/torch/optim/lr_scheduler.py:131: UserWarning: Detected call of `lr_scheduler.step()` before `optimizer.step()`. In PyTorch 1.1.0 and later, you should call them in the opposite order: `optimizer.step()` before `lr_scheduler.step()`. Failure to do this will result in PyTorch skipping the first value of the learning rate schedule. See more details at https://pytorch.org/docs/stable/optim.html#how-to-adjust-learning-rate
warnings.warn("Detected call of `lr_scheduler.step()` before `optimizer.step()`. "
/home/vibhatha/venv/ENVTORCH/lib/python3.8/site-packages/torch/optim/lr_scheduler.py:156: UserWarning: The epoch parameter in `scheduler.step()` was not necessary and is being deprecated where possible. Please use `scheduler.step()` to step the scheduler. During the deprecation, if epoch is different from None, the closed form is used instead of the new chainable form, where available. Please open an issue if you are unable to replicate your use case: https://github.com/pytorch/pytorch/issues/new/choose.
warnings.warn(EPOCH_DEPRECATION_WARNING, UserWarning)
/home/vibhatha/venv/ENVTORCH/lib/python3.8/site-packages/torch/optim/lr_scheduler.py:156: UserWarning: The epoch parameter in `scheduler.step()` was not necessary and is being deprecated where possible. Please use `scheduler.step()` to step the scheduler. During the deprecation, if epoch is different from None, the closed form is used instead of the new chainable form, where available. Please open an issue if you are unable to replicate your use case: https://github.com/pytorch/pytorch/issues/new/choose.
warnings.warn(EPOCH_DEPRECATION_WARNING, UserWarning)
/home/vibhatha/venv/ENVTORCH/lib/python3.8/site-packages/torch/optim/lr_scheduler.py:156: UserWarning: The epoch parameter in `scheduler.step()` was not necessary and is being deprecated where possible. Please use `scheduler.step()` to step the scheduler. During the deprecation, if epoch is different from None, the closed form is used instead of the new chainable form, where available. Please open an issue if you are unable to replicate your use case: https://github.com/pytorch/pytorch/issues/new/choose.
warnings.warn(EPOCH_DEPRECATION_WARNING, UserWarning)
CCLE MSE: 3322.74 MAE: 47.47 R2: +0.11
Epoch Running Time: 889.7 Seconds.
/home/vibhatha/sandbox/UNO/Benchmarks/Pilot1/UnoMT/utils/datasets/drug_resp_dataset.py:243: UserWarning: Heterogeneous Cylon Table Detected!. Use Numpy operations with Caution.
self.__drug_resp_array = self.__drug_resp_tb.to_numpy(zero_copy_only=False)
/home/vibhatha/venv/ENVTORCH/lib/python3.8/site-packages/torch/optim/lr_scheduler.py:131: UserWarning: Detected call of `lr_scheduler.step()` before `optimizer.step()`. In PyTorch 1.1.0 and later, you should call them in the opposite order: `optimizer.step()` before `lr_scheduler.step()`. Failure to do this will result in PyTorch skipping the first value of the learning rate schedule. See more details at https://pytorch.org/docs/stable/optim.html#how-to-adjust-learning-rate
warnings.warn("Detected call of `lr_scheduler.step()` before `optimizer.step()`. "
/home/vibhatha/venv/ENVTORCH/lib/python3.8/site-packages/torch/optim/lr_scheduler.py:156: UserWarning: The epoch parameter in `scheduler.step()` was not necessary and is being deprecated where possible. Please use `scheduler.step()` to step the scheduler. During the deprecation, if epoch is different from None, the closed form is used instead of the new chainable form, where available. Please open an issue if you are unable to replicate your use case: https://github.com/pytorch/pytorch/issues/new/choose.
warnings.warn(EPOCH_DEPRECATION_WARNING, UserWarning)
/home/vibhatha/venv/ENVTORCH/lib/python3.8/site-packages/torch/optim/lr_scheduler.py:156: UserWarning: The epoch parameter in `scheduler.step()` was not necessary and is being deprecated where possible. Please use `scheduler.step()` to step the scheduler. During the deprecation, if epoch is different from None, the closed form is used instead of the new chainable form, where available. Please open an issue if you are unable to replicate your use case: https://github.com/pytorch/pytorch/issues/new/choose.
warnings.warn(EPOCH_DEPRECATION_WARNING, UserWarning)
/home/vibhatha/venv/ENVTORCH/lib/python3.8/site-packages/torch/optim/lr_scheduler.py:156: UserWarning: The epoch parameter in `scheduler.step()` was not necessary and is being deprecated where possible. Please use `scheduler.step()` to step the scheduler. During the deprecation, if epoch is different from None, the closed form is used instead of the new chainable form, where available. Please open an issue if you are unable to replicate your use case: https://github.com/pytorch/pytorch/issues/new/choose.
warnings.warn(EPOCH_DEPRECATION_WARNING, UserWarning)
CCLE MSE: 3480.50 MAE: 48.50 R2: +0.12
Epoch Running Time: 890.0 Seconds.
/home/vibhatha/sandbox/UNO/Benchmarks/Pilot1/UnoMT/utils/datasets/drug_resp_dataset.py:243: UserWarning: Heterogeneous Cylon Table Detected!. Use Numpy operations with Caution.
self.__drug_resp_array = self.__drug_resp_tb.to_numpy(zero_copy_only=False)
/home/vibhatha/venv/ENVTORCH/lib/python3.8/site-packages/torch/optim/lr_scheduler.py:131: UserWarning: Detected call of `lr_scheduler.step()` before `optimizer.step()`. In PyTorch 1.1.0 and later, you should call them in the opposite order: `optimizer.step()` before `lr_scheduler.step()`. Failure to do this will result in PyTorch skipping the first value of the learning rate schedule. See more details at https://pytorch.org/docs/stable/optim.html#how-to-adjust-learning-rate
warnings.warn("Detected call of `lr_scheduler.step()` before `optimizer.step()`. "
/home/vibhatha/venv/ENVTORCH/lib/python3.8/site-packages/torch/optim/lr_scheduler.py:156: UserWarning: The epoch parameter in `scheduler.step()` was not necessary and is being deprecated where possible. Please use `scheduler.step()` to step the scheduler. During the deprecation, if epoch is different from None, the closed form is used instead of the new chainable form, where available. Please open an issue if you are unable to replicate your use case: https://github.com/pytorch/pytorch/issues/new/choose.
warnings.warn(EPOCH_DEPRECATION_WARNING, UserWarning)
/home/vibhatha/venv/ENVTORCH/lib/python3.8/site-packages/torch/optim/lr_scheduler.py:156: UserWarning: The epoch parameter in `scheduler.step()` was not necessary and is being deprecated where possible. Please use `scheduler.step()` to step the scheduler. During the deprecation, if epoch is different from None, the closed form is used instead of the new chainable form, where available. Please open an issue if you are unable to replicate your use case: https://github.com/pytorch/pytorch/issues/new/choose.
warnings.warn(EPOCH_DEPRECATION_WARNING, UserWarning)
/home/vibhatha/venv/ENVTORCH/lib/python3.8/site-packages/torch/optim/lr_scheduler.py:156: UserWarning: The epoch parameter in `scheduler.step()` was not necessary and is being deprecated where possible. Please use `scheduler.step()` to step the scheduler. During the deprecation, if epoch is different from None, the closed form is used instead of the new chainable form, where available. Please open an issue if you are unable to replicate your use case: https://github.com/pytorch/pytorch/issues/new/choose.
warnings.warn(EPOCH_DEPRECATION_WARNING, UserWarning)
CCLE MSE: 3533.67 MAE: 49.41 R2: +0.13
Epoch Running Time: 890.0 Seconds.
/home/vibhatha/sandbox/UNO/Benchmarks/Pilot1/UnoMT/utils/datasets/drug_resp_dataset.py:243: UserWarning: Heterogeneous Cylon Table Detected!. Use Numpy operations with Caution.
self.__drug_resp_array = self.__drug_resp_tb.to_numpy(zero_copy_only=False)
/home/vibhatha/venv/ENVTORCH/lib/python3.8/site-packages/torch/optim/lr_scheduler.py:131: UserWarning: Detected call of `lr_scheduler.step()` before `optimizer.step()`. In PyTorch 1.1.0 and later, you should call them in the opposite order: `optimizer.step()` before `lr_scheduler.step()`. Failure to do this will result in PyTorch skipping the first value of the learning rate schedule. See more details at https://pytorch.org/docs/stable/optim.html#how-to-adjust-learning-rate
warnings.warn("Detected call of `lr_scheduler.step()` before `optimizer.step()`. "
/home/vibhatha/venv/ENVTORCH/lib/python3.8/site-packages/torch/optim/lr_scheduler.py:156: UserWarning: The epoch parameter in `scheduler.step()` was not necessary and is being deprecated where possible. Please use `scheduler.step()` to step the scheduler. During the deprecation, if epoch is different from None, the closed form is used instead of the new chainable form, where available. Please open an issue if you are unable to replicate your use case: https://github.com/pytorch/pytorch/issues/new/choose.
warnings.warn(EPOCH_DEPRECATION_WARNING, UserWarning)
/home/vibhatha/venv/ENVTORCH/lib/python3.8/site-packages/torch/optim/lr_scheduler.py:156: UserWarning: The epoch parameter in `scheduler.step()` was not necessary and is being deprecated where possible. Please use `scheduler.step()` to step the scheduler. During the deprecation, if epoch is different from None, the closed form is used instead of the new chainable form, where available. Please open an issue if you are unable to replicate your use case: https://github.com/pytorch/pytorch/issues/new/choose.
warnings.warn(EPOCH_DEPRECATION_WARNING, UserWarning)
/home/vibhatha/venv/ENVTORCH/lib/python3.8/site-packages/torch/optim/lr_scheduler.py:156: UserWarning: The epoch parameter in `scheduler.step()` was not necessary and is being deprecated where possible. Please use `scheduler.step()` to step the scheduler. During the deprecation, if epoch is different from None, the closed form is used instead of the new chainable form, where available. Please open an issue if you are unable to replicate your use case: https://github.com/pytorch/pytorch/issues/new/choose.
warnings.warn(EPOCH_DEPRECATION_WARNING, UserWarning)