Show code cell content

Hide code cell content

import os
from pathlib import Path
from functools import partial
from urllib import request
import contextlib
import time

import numpy as onp

import matplotlib.pyplot as plt

import optax

import haiku as hk

from jax import numpy as jnp, random

from jax_md_mod import io, custom_space, custom_energy
from jax_md_mod.model import neural_networks, layers
from jax_md import space, energy, partition, simulate

from chemtrain import trainers, ensemble, util, quantity

%config InlineBackend.figure_formats = ['svg']
%matplotlib inline

DiffTRe: CG Water on Structural Data

This example, based on the paper Learning neural network potentials from experimental data via Differentiable Trajectory Reweighting, employs DiffTRe with an adaptive scaling of the learning rate to learn a structurally accurate coarse-grained model for water.

Introduction

Classical approaches to Top-Down learning of potential models, e.g., Iterative Boltzmann Inversion [1] or Inverse Monte Carlo [2], only learn pair-wise dependencies between particles. However, to correctly predict the anomalous properties of water, it is necessary that the water model correctly represents triplet interactions [3].

Therefore, the DiffTRe algorithm [4] leverages umbrella sampling [5] to flexibly train arbitrary potential models on experimental reference data.

We present the details of the approach in the toy example Differentiable Trajectory Reweighting (DiffTRe).

Load Data

# Download if not present
data_urls = {
    "O_O_O_ADF.csv": "https://github.com/tummfm/difftre/blob/92c0790b89f0d570ed9f79663e6c06580f598345/data/experimental/O_O_O_ADF.csv?raw=true",
    "O_O_RDF.csv": "https://github.com/tummfm/difftre/blob/92c0790b89f0d570ed9f79663e6c06580f598345/data/experimental/O_O_RDF.csv?raw=true",
    "Water_experimental_3nm.gro": "https://github.com/tummfm/difftre/blob/92c0790b89f0d570ed9f79663e6c06580f598345/data/confs/Water_experimental.gro?raw=true"
}
data_path = Path("../_data/water_data")
data_path.mkdir(exist_ok=True, parents=True)

for fname, furl in data_urls.items():
    if not (data_path / fname).exists():
        request.urlretrieve(furl, data_path / fname)

Setup Training

First, we need to create a reference system, in which we can predict macroscopic measurements. Therefore, we load a box of 901 water particles and construct a periodic space. Additionally, we initialize a neighbor list, which enables accelerating the simulation and the extraction of a neighborlist graph for the Graph Neural Network (GNN) model.

System

system: quantity.targets.InitArguments = {}

file = data_path / 'Water_experimental_3nm.gro'

with open(file, "r") as f:
    lines = f.readlines()

# Add time to read file with mdtraj
lines[0] = "system description here, t= 1200.000000, length unit=nm\n"

with open(file, "w") as f:
    f.writelines(lines)


box, r_init, _, _ = io.load_box(file)  # initial configuration
mass = 18.0154

box_tensor, scale_fn = custom_space.init_fractional_coordinates(box)
displacement_fn, shift_fn = space.periodic_general(
    box_tensor, fractional_coordinates=True, wrapped=True)
r_init = scale_fn(r_init)

neighbor_fn = partition.neighbor_list(
    displacement_fn, box, r_cutoff=0.5, dr_threshold=0.0,
    disable_cell_list=True, fractional_coordinates=True
)
nbrs_init = neighbor_fn.allocate(r_init, capacity_multiplier=1.5, extra_capacity=1)

system.update({
    "displacement_fn": displacement_fn,
    "reference_box": box_tensor,
    "nbrs_init": nbrs_init,
    "r_init": r_init
})

Model

Next, we construct a potential model for the reference system. We use the expressive DimeNet architecture [6] and employ a $\Delta$-learning approach based on a prior. The prior stabilizes the simulation in the first epochs and additionally should enforce physical constraints. Like the paper, we choose a simple repulsive prior based on the repulsive term of the Lennard Jones potential. Consequently, the system directly respects the Pauli exclusion principle.

mlp_init = {
    'b_init': hk.initializers.Constant(0.),
    'w_init': layers.OrthogonalVarianceScalingInit(scale=1.)
    }

model_init, model_apply = neural_networks.dimenetpp_neighborlist(
    displacement_fn, 0.5, 1, r_init, nbrs_init,
    embed_size=32, init_kwargs=mlp_init,
)

prior_energy_apply = custom_energy.generic_repulsion_neighborlist(
    displacement_fn, box, sigma=0.3165, epsilon=1.0, exp=12.0,
    r_onset=0.45, r_cutoff=0.5, initialize_neighbor_list=False)

def energy_fn_template(energy_params):
    def energy_fn(pos, neighbor, **dyn_kwargs):
        gnn_energy = model_apply(energy_params, pos, neighbor, **dyn_kwargs)

        prior_energy = prior_energy_apply(pos, neighbor, **dyn_kwargs)

        return prior_energy + gnn_energy

    return energy_fn

init_model_params = model_init(random.PRNGKey(11), r_init, nbrs_init)

Show code cell output

Hide code cell output

Capping edges and triplets. Beware of overflow, which is currently not being detected.

/home/paul/miniconda3/envs/chemtrain_execute/lib/python3.10/site-packages/jax/_src/numpy/lax_numpy.py:166: UserWarning: Explicitly requested dtype float64 requested in asarray is not available, and will be truncated to dtype float32. To enable more dtypes, set the jax_enable_x64 configuration option or the JAX_ENABLE_X64 shell environment variable. See https://github.com/google/jax#current-gotchas for more.
  return asarray(x, dtype=self.dtype)

/home/paul/miniconda3/envs/chemtrain_execute/lib/python3.10/site-packages/jax/_src/numpy/lax_numpy.py:166: UserWarning: Explicitly requested dtype float64 requested in asarray is not available, and will be truncated to dtype float32. To enable more dtypes, set the jax_enable_x64 configuration option or the JAX_ENABLE_X64 shell environment variable. See https://github.com/google/jax#current-gotchas for more.
  return asarray(x, dtype=self.dtype)

/home/paul/miniconda3/envs/chemtrain_execute/lib/python3.10/site-packages/jax/_src/numpy/lax_numpy.py:166: UserWarning: Explicitly requested dtype float64 requested in asarray is not available, and will be truncated to dtype float32. To enable more dtypes, set the jax_enable_x64 configuration option or the JAX_ENABLE_X64 shell environment variable. See https://github.com/google/jax#current-gotchas for more.
  return asarray(x, dtype=self.dtype)
/home/paul/miniconda3/envs/chemtrain_execute/lib/python3.10/site-packages/jax/_src/numpy/lax_numpy.py:166: UserWarning: Explicitly requested dtype float64 requested in asarray is not available, and will be truncated to dtype float32. To enable more dtypes, set the jax_enable_x64 configuration option or the JAX_ENABLE_X64 shell environment variable. See https://github.com/google/jax#current-gotchas for more.
  return asarray(x, dtype=self.dtype)
/home/paul/miniconda3/envs/chemtrain_execute/lib/python3.10/site-packages/jax/_src/numpy/lax_numpy.py:166: UserWarning: Explicitly requested dtype float64 requested in asarray is not available, and will be truncated to dtype float32. To enable more dtypes, set the jax_enable_x64 configuration option or the JAX_ENABLE_X64 shell environment variable. See https://github.com/google/jax#current-gotchas for more.
  return asarray(x, dtype=self.dtype)
/home/paul/miniconda3/envs/chemtrain_execute/lib/python3.10/site-packages/jax/_src/numpy/lax_numpy.py:166: UserWarning: Explicitly requested dtype float64 requested in asarray is not available, and will be truncated to dtype float32. To enable more dtypes, set the jax_enable_x64 configuration option or the JAX_ENABLE_X64 shell environment variable. See https://github.com/google/jax#current-gotchas for more.
  return asarray(x, dtype=self.dtype)
/home/paul/miniconda3/envs/chemtrain_execute/lib/python3.10/site-packages/jax/_src/numpy/lax_numpy.py:166: UserWarning: Explicitly requested dtype float64 requested in asarray is not available, and will be truncated to dtype float32. To enable more dtypes, set the jax_enable_x64 configuration option or the JAX_ENABLE_X64 shell environment variable. See https://github.com/google/jax#current-gotchas for more.
  return asarray(x, dtype=self.dtype)
/home/paul/miniconda3/envs/chemtrain_execute/lib/python3.10/site-packages/jax/_src/numpy/lax_numpy.py:166: UserWarning: Explicitly requested dtype float64 requested in asarray is not available, and will be truncated to dtype float32. To enable more dtypes, set the jax_enable_x64 configuration option or the JAX_ENABLE_X64 shell environment variable. See https://github.com/google/jax#current-gotchas for more.
  return asarray(x, dtype=self.dtype)
/home/paul/miniconda3/envs/chemtrain_execute/lib/python3.10/site-packages/jax/_src/numpy/lax_numpy.py:166: UserWarning: Explicitly requested dtype float64 requested in asarray is not available, and will be truncated to dtype float32. To enable more dtypes, set the jax_enable_x64 configuration option or the JAX_ENABLE_X64 shell environment variable. See https://github.com/google/jax#current-gotchas for more.
  return asarray(x, dtype=self.dtype)
/home/paul/miniconda3/envs/chemtrain_execute/lib/python3.10/site-packages/jax/_src/numpy/reductions.py:221: UserWarning: Explicitly requested dtype <class 'jax.numpy.float64'> requested in sum is not available, and will be truncated to dtype float32. To enable more dtypes, set the jax_enable_x64 configuration option or the JAX_ENABLE_X64 shell environment variable. See https://github.com/google/jax#current-gotchas for more.
  return _reduction(a, "sum", np.sum, lax.add, 0, preproc=_cast_to_numeric,

Simulator

As outlined in the Differentiable Trajectory Reweighting (DiffTRe) example, a molecular dynamics simulation is an efficient method to sample particle position from the equilibrium distribution. Hence, we set up a simulation routine based on a Nosé-Hoover integrator for a total time of $70\ \text{ps}$. Since the samples should be uncorrelated, we subsample the trajectory every $0.1\ \text{ps}$ and additionally substract the first $10\ \text{ps}$ to account for the changing potential model.

system_temperature = 296.15  # Kelvin
boltzmann_constant = 0.0083145107  # in kJ / mol K
kbt = system_temperature * boltzmann_constant

time_step = 0.01
total_time = 70.
t_equilib = 20.
print_every = 0.1

timings = ensemble.sampling.process_printouts(
    time_step, total_time, t_equilib, print_every
)

# The simulator template enables to initialize a 
# new simulation routine whenever the potential model
# changes.
simulator_template = partial(
    simulate.nvt_nose_hoover, shift_fn=shift_fn,
    dt=time_step, kT=kbt, tau=50. * time_step, chain_length=3, chain_steps=1
)

# We need an initial state of the simulator, where
# we can start the simulations from.
simulator_init, _ = simulator_template(
    energy_fn_template(init_model_params)
)

init_simulator_state = simulator_init(
    random.PRNGKey(11), r_init, neighbor=nbrs_init, mass=mass
)
reference_simulation_state = ensemble.sampling.SimulatorState(
    sim_state=init_simulator_state, nbrs=nbrs_init
)

/home/paul/miniconda3/envs/chemtrain_execute/lib/python3.10/site-packages/jax/_src/numpy/lax_numpy.py:166: UserWarning: Explicitly requested dtype float64 requested in asarray is not available, and will be truncated to dtype float32. To enable more dtypes, set the jax_enable_x64 configuration option or the JAX_ENABLE_X64 shell environment variable. See https://github.com/google/jax#current-gotchas for more.
  return asarray(x, dtype=self.dtype)

/home/paul/miniconda3/envs/chemtrain_execute/lib/python3.10/site-packages/jax/_src/numpy/lax_numpy.py:166: UserWarning: Explicitly requested dtype float64 requested in asarray is not available, and will be truncated to dtype float32. To enable more dtypes, set the jax_enable_x64 configuration option or the JAX_ENABLE_X64 shell environment variable. See https://github.com/google/jax#current-gotchas for more.
  return asarray(x, dtype=self.dtype)
/home/paul/miniconda3/envs/chemtrain_execute/lib/python3.10/site-packages/jax/_src/numpy/lax_numpy.py:166: UserWarning: Explicitly requested dtype float64 requested in asarray is not available, and will be truncated to dtype float32. To enable more dtypes, set the jax_enable_x64 configuration option or the JAX_ENABLE_X64 shell environment variable. See https://github.com/google/jax#current-gotchas for more.
  return asarray(x, dtype=self.dtype)
/home/paul/miniconda3/envs/chemtrain_execute/lib/python3.10/site-packages/jax/_src/numpy/reductions.py:221: UserWarning: Explicitly requested dtype <class 'jax.numpy.float64'> requested in sum is not available, and will be truncated to dtype float32. To enable more dtypes, set the jax_enable_x64 configuration option or the JAX_ENABLE_X64 shell environment variable. See https://github.com/google/jax#current-gotchas for more.
  return _reduction(a, "sum", np.sum, lax.add, 0, preproc=_cast_to_numeric,

Targets

Next, we can set up the macroscopic properties we want to match. We use the experimentally determined RDF and ADF at $296.15\ \text{K}$ [7] and additionally the pressure target of $1\ \text{bar}$.

reference_rdf = onp.loadtxt(data_path / "O_O_RDF.csv")
reference_adf = onp.loadtxt(data_path / "O_O_O_ADF.csv")
pressure_target = 1. / 16.6054   # 1 bar in kJ / mol nm^3

target_builder = quantity.targets.TargetBuilder(system, strict=True)

target_builder["rdf"] = quantity.targets.init_radial_distribution_target(
    reference_rdf, nbins=300, rdf_cut=1.0, gamma=1.0
)
target_builder["adf"] = quantity.targets.init_angular_distribution_target(
    reference_adf, nbins=150, r_outer=0.318, gamma=1.0
)
target_builder["pressure"] = quantity.targets.init_pressure_target(
    energy_fn_template, include_kinetic=False, gamma=1.0e-7, target=pressure_target
)

targets, compute_fns = target_builder.build()

[ADF] Estimates 498644 max. triplets in neighbor list and 23308 max. triplets in cutoff-shell.

Trainer

Finally, we can use the potential model, simulation routines, and targets to initialize the DiffTRe algorithm. Therefore, we use the high-level interface, called trainer. It is possible to add multiple statepoints to the trainer, e.g., to train different systems at different thermodynamic state points. However, here we only set up a single state points. We set the reweighting ratio to 0.9, such that the computed trajectories might be re-used if the updated potential and the reference potential are sufficiently close.

check_freq = None
num_updates = 300

initial_lr = 0.005

lr_schedule = optax.exponential_decay(initial_lr, num_updates, 0.01)
optimizer = optax.chain(
    optax.scale_by_adam(0.1, 0.4),
    optax.scale_by_schedule(lr_schedule),
    optax.scale_by_learning_rate(1.0, flip_sign=True)
)

trainer = trainers.Difftre(init_model_params,
                           optimizer,
                           energy_fn_template=energy_fn_template)

trainer.add_statepoint(energy_fn_template, simulator_template,
                       neighbor_fn, timings, {'kT': kbt}, compute_fns,
                       reference_simulation_state, targets)

Show code cell output

Hide code cell output

/home/paul/chemtrain_execute/chemtrain/ensemble/reweighting.py:777: UserWarning: Propagation function is not safe by default. Do not forget to use the wrapper around the compute function to ensure that the neighborlist does not overflow.
  warnings.warn(

/home/paul/miniconda3/envs/chemtrain_execute/lib/python3.10/site-packages/jax/_src/numpy/lax_numpy.py:166: UserWarning: Explicitly requested dtype float64 requested in asarray is not available, and will be truncated to dtype float32. To enable more dtypes, set the jax_enable_x64 configuration option or the JAX_ENABLE_X64 shell environment variable. See https://github.com/google/jax#current-gotchas for more.
  return asarray(x, dtype=self.dtype)

/home/paul/miniconda3/envs/chemtrain_execute/lib/python3.10/site-packages/jax/_src/numpy/lax_numpy.py:166: UserWarning: Explicitly requested dtype float64 requested in asarray is not available, and will be truncated to dtype float32. To enable more dtypes, set the jax_enable_x64 configuration option or the JAX_ENABLE_X64 shell environment variable. See https://github.com/google/jax#current-gotchas for more.
  return asarray(x, dtype=self.dtype)

/home/paul/miniconda3/envs/chemtrain_execute/lib/python3.10/site-packages/jax/_src/numpy/lax_numpy.py:166: UserWarning: Explicitly requested dtype float64 requested in asarray is not available, and will be truncated to dtype float32. To enable more dtypes, set the jax_enable_x64 configuration option or the JAX_ENABLE_X64 shell environment variable. See https://github.com/google/jax#current-gotchas for more.
  return asarray(x, dtype=self.dtype)

/home/paul/miniconda3/envs/chemtrain_execute/lib/python3.10/site-packages/jax/_src/numpy/reductions.py:221: UserWarning: Explicitly requested dtype <class 'jax.numpy.float64'> requested in sum is not available, and will be truncated to dtype float32. To enable more dtypes, set the jax_enable_x64 configuration option or the JAX_ENABLE_X64 shell environment variable. See https://github.com/google/jax#current-gotchas for more.
  return _reduction(a, "sum", np.sum, lax.add, 0, preproc=_cast_to_numeric,

/home/paul/miniconda3/envs/chemtrain_execute/lib/python3.10/site-packages/jax/_src/numpy/lax_numpy.py:166: UserWarning: Explicitly requested dtype float64 requested in asarray is not available, and will be truncated to dtype float32. To enable more dtypes, set the jax_enable_x64 configuration option or the JAX_ENABLE_X64 shell environment variable. See https://github.com/google/jax#current-gotchas for more.
  return asarray(x, dtype=self.dtype)
/home/paul/miniconda3/envs/chemtrain_execute/lib/python3.10/site-packages/jax/_src/numpy/lax_numpy.py:166: UserWarning: Explicitly requested dtype float64 requested in asarray is not available, and will be truncated to dtype float32. To enable more dtypes, set the jax_enable_x64 configuration option or the JAX_ENABLE_X64 shell environment variable. See https://github.com/google/jax#current-gotchas for more.
  return asarray(x, dtype=self.dtype)
/home/paul/miniconda3/envs/chemtrain_execute/lib/python3.10/site-packages/jax/_src/numpy/lax_numpy.py:166: UserWarning: Explicitly requested dtype float64 requested in asarray is not available, and will be truncated to dtype float32. To enable more dtypes, set the jax_enable_x64 configuration option or the JAX_ENABLE_X64 shell environment variable. See https://github.com/google/jax#current-gotchas for more.
  return asarray(x, dtype=self.dtype)
/home/paul/miniconda3/envs/chemtrain_execute/lib/python3.10/site-packages/jax/_src/numpy/lax_numpy.py:166: UserWarning: Explicitly requested dtype float64 requested in asarray is not available, and will be truncated to dtype float32. To enable more dtypes, set the jax_enable_x64 configuration option or the JAX_ENABLE_X64 shell environment variable. See https://github.com/google/jax#current-gotchas for more.
  return asarray(x, dtype=self.dtype)
/home/paul/miniconda3/envs/chemtrain_execute/lib/python3.10/site-packages/jax/_src/numpy/lax_numpy.py:166: UserWarning: Explicitly requested dtype float64 requested in asarray is not available, and will be truncated to dtype float32. To enable more dtypes, set the jax_enable_x64 configuration option or the JAX_ENABLE_X64 shell environment variable. See https://github.com/google/jax#current-gotchas for more.
  return asarray(x, dtype=self.dtype)

Time for trajectory initialization 0: 0.790979262193044 mins

trainer.init_step_size_adaption(0.10)

[Step size] Use 5 iterations for 10 interior points.

Train

if os.environ.get("TRAIN", "False").lower() == "true":
    Path("../_data/output").mkdir(parents=True, exist_ok=True)
       
    # Save the training log
    with open("../_data/output/difftre_water.log", "w") as f:
        with contextlib.redirect_stdout(f), contextlib.redirect_stderr(f):

            start = time.time()
            trainer.train(num_updates)
            print(f"Total training time: {(time.time() - start) / 3600 : .1f} hours")
    
    trainer.save_energy_params("../_data/output/water_difftre_params.pkl", '.pkl')
    trainer.save_trainer("../_data/output/water_difftre_trainer.pkl", '.pkl')
    
results = onp.load("../_data/output/water_difftre_trainer.pkl", allow_pickle=True)

with open("../_data/output/difftre_water.log") as f:
    print(f.read())

Show code cell output

Hide code cell output

/home/paul/miniconda3/envs/chemtrain_execute/lib/python3.10/site-packages/jax/_src/numpy/reductions.py:221: UserWarning: Explicitly requested dtype <class 'jax.numpy.float64'> requested in sum is not available, and will be truncated to dtype float32. To enable more dtypes, set the jax_enable_x64 configuration option or the JAX_ENABLE_X64 shell environment variable. See https://github.com/google/jax#current-gotchas for more.
  return _reduction(a, "sum", np.sum, lax.add, 0, preproc=_cast_to_numeric,
/home/paul/miniconda3/envs/chemtrain_execute/lib/python3.10/site-packages/jax/_src/numpy/reductions.py:221: UserWarning: Explicitly requested dtype <class 'jax.numpy.float64'> requested in sum is not available, and will be truncated to dtype float32. To enable more dtypes, set the jax_enable_x64 configuration option or the JAX_ENABLE_X64 shell environment variable. See https://github.com/google/jax#current-gotchas for more.
  return _reduction(a, "sum", np.sum, lax.add, 0, preproc=_cast_to_numeric,
/home/paul/miniconda3/envs/chemtrain_execute/lib/python3.10/site-packages/jax/_src/numpy/lax_numpy.py:166: UserWarning: Explicitly requested dtype float64 requested in asarray is not available, and will be truncated to dtype float32. To enable more dtypes, set the jax_enable_x64 configuration option or the JAX_ENABLE_X64 shell environment variable. See https://github.com/google/jax#current-gotchas for more.
  return asarray(x, dtype=self.dtype)
/home/paul/miniconda3/envs/chemtrain_execute/lib/python3.10/site-packages/jax/_src/numpy/lax_numpy.py:166: UserWarning: Explicitly requested dtype float64 requested in asarray is not available, and will be truncated to dtype float32. To enable more dtypes, set the jax_enable_x64 configuration option or the JAX_ENABLE_X64 shell environment variable. See https://github.com/google/jax#current-gotchas for more.
  return asarray(x, dtype=self.dtype)
[Step Size] Found optimal step size 0.05929497629404068 with residual 0.0011034011840820312

[DiffTRe] Epoch 0
	Epoch loss = 0.19014
	Gradient norm: 0.12823013961315155
	Elapsed time = 3.404 min
[Statepoint 0]
	kT = 2.464 ref_kT = 2.462
	Predicted entropy: 1.6329046275131986e-06
	Predicted free_energy: 4.696545147453435e-06
	Predicted pressure: 669.3466796875
[Step Size] Found optimal step size 0.05493614822626114 with residual 0.0005590915679931641

[DiffTRe] Epoch 1
	Epoch loss = 0.17469
	Gradient norm: 0.12685701251029968
	Elapsed time = 1.772 min
[Statepoint 0]
	kT = 2.458 ref_kT = 2.462
	Predicted entropy: 0.4574434161186218
	Predicted free_energy: -116.05196380615234
	Predicted pressure: 642.810302734375
[Step Size] Found optimal step size 0.0442129522562027 with residual 0.003111600875854492

[DiffTRe] Epoch 2
	Epoch loss = 0.16052
	Gradient norm: 0.140895813703537
	Elapsed time = 1.773 min
[Statepoint 0]
	kT = 2.467 ref_kT = 2.462
	Predicted entropy: 0.8821817636489868
	Predicted free_energy: -270.0376281738281
	Predicted pressure: 614.1198120117188
[Step Size] Found optimal step size 0.043188437819480896 with residual 5.0067901611328125e-05

[DiffTRe] Epoch 3
	Epoch loss = 0.14806
	Gradient norm: 0.2002345621585846
	Elapsed time = 1.775 min
[Statepoint 0]
	kT = 2.469 ref_kT = 2.462
	Predicted entropy: 1.209067940711975
	Predicted free_energy: -404.0960388183594
	Predicted pressure: 592.4592895507812
[Step Size] Found optimal step size 0.05656294897198677 with residual 0.00038909912109375

[DiffTRe] Epoch 4
	Epoch loss = 0.13578
	Gradient norm: 0.18344427645206451
	Elapsed time = 1.777 min
[Statepoint 0]
	kT = 2.461 ref_kT = 2.462
	Predicted entropy: 1.5147840976715088
	Predicted free_energy: -547.8067016601562
	Predicted pressure: 566.3475952148438
[Step Size] Found optimal step size 0.036718495190143585 with residual 0.0017390251159667969

[DiffTRe] Epoch 5
	Epoch loss = 0.12027
	Gradient norm: 0.18248403072357178
	Elapsed time = 1.771 min
[Statepoint 0]
	kT = 2.452 ref_kT = 2.462
	Predicted entropy: 1.8580281734466553
	Predicted free_energy: -748.82470703125
	Predicted pressure: 531.8713989257812
[Step Size] Found optimal step size 0.0359361469745636 with residual 0.0009217262268066406

[DiffTRe] Epoch 6
	Epoch loss = 0.11041
	Gradient norm: 0.18522265553474426
	Elapsed time = 1.775 min
[Statepoint 0]
	kT = 2.464 ref_kT = 2.462
	Predicted entropy: 2.079115390777588
	Predicted free_energy: -894.3207397460938
	Predicted pressure: 509.0533752441406
[Step Size] Found optimal step size 0.03794170543551445 with residual 0.0016269683837890625

[DiffTRe] Epoch 7
	Epoch loss = 0.10117
	Gradient norm: 0.21723368763923645
	Elapsed time = 1.778 min
[Statepoint 0]
	kT = 2.457 ref_kT = 2.462
	Predicted entropy: 2.322474956512451
	Predicted free_energy: -1034.0592041015625
	Predicted pressure: 489.0552673339844
[Step Size] Found optimal step size 0.027057064697146416 with residual 6.890296936035156e-05

[DiffTRe] Epoch 8
	Epoch loss = 0.09125
	Gradient norm: 0.19186505675315857
	Elapsed time = 1.777 min
[Statepoint 0]
	kT = 2.450 ref_kT = 2.462
	Predicted entropy: 2.548941135406494
	Predicted free_energy: -1219.4130859375
	Predicted pressure: 462.0693054199219
[Step Size] Found optimal step size 0.032198235392570496 with residual 3.24249267578125e-05

[DiffTRe] Epoch 9
	Epoch loss = 0.08488
	Gradient norm: 0.1502356082201004
	Elapsed time = 1.775 min
[Statepoint 0]
	kT = 2.460 ref_kT = 2.462
	Predicted entropy: 2.6691977977752686
	Predicted free_energy: -1386.005126953125
	Predicted pressure: 439.672119140625
[Step Size] Found optimal step size 0.03910281881690025 with residual 0.0006232261657714844

[DiffTRe] Epoch 10
	Epoch loss = 0.07871
	Gradient norm: 0.15819592773914337
	Elapsed time = 1.776 min
[Statepoint 0]
	kT = 2.465 ref_kT = 2.462
	Predicted entropy: 2.8133163452148438
	Predicted free_energy: -1524.0372314453125
	Predicted pressure: 424.2068786621094
[Step Size] Found optimal step size 0.04287797957658768 with residual 0.0002224445343017578

[DiffTRe] Epoch 11
	Epoch loss = 0.07101
	Gradient norm: 0.1506803333759308
	Elapsed time = 1.778 min
[Statepoint 0]
	kT = 2.462 ref_kT = 2.462
	Predicted entropy: 2.9710891246795654
	Predicted free_energy: -1653.4901123046875
	Predicted pressure: 406.1748046875
[Step Size] Found optimal step size 0.037295952439308167 with residual 0.0009670257568359375

[DiffTRe] Epoch 12
	Epoch loss = 0.06355
	Gradient norm: 0.061876192688941956
	Elapsed time = 1.774 min
[Statepoint 0]
	kT = 2.461 ref_kT = 2.462
	Predicted entropy: 3.0779943466186523
	Predicted free_energy: -1922.0245361328125
	Predicted pressure: 375.1353454589844
[Step Size] Found optimal step size 0.041406407952308655 with residual 0.0003685951232910156

[DiffTRe] Epoch 13
	Epoch loss = 0.05876
	Gradient norm: 0.053959302604198456
	Elapsed time = 1.775 min
[Statepoint 0]
	kT = 2.470 ref_kT = 2.462
	Predicted entropy: 3.1934802532196045
	Predicted free_energy: -2007.9583740234375
	Predicted pressure: 362.7120361328125
[Step Size] Found optimal step size 0.03642046079039574 with residual 3.337860107421875e-06

[DiffTRe] Epoch 14
	Epoch loss = 0.05333
	Gradient norm: 0.06847898662090302
	Elapsed time = 1.771 min
[Statepoint 0]
	kT = 2.460 ref_kT = 2.462
	Predicted entropy: 3.312819004058838
	Predicted free_energy: -2057.07470703125
	Predicted pressure: 352.8749084472656
[Step Size] Found optimal step size 0.037264905869960785 with residual 0.0003142356872558594

[DiffTRe] Epoch 15
	Epoch loss = 0.04943
	Gradient norm: 0.03620465099811554
	Elapsed time = 1.775 min
[Statepoint 0]
	kT = 2.458 ref_kT = 2.462
	Predicted entropy: 3.3637630939483643
	Predicted free_energy: -2193.57080078125
	Predicted pressure: 340.0343933105469
[Step Size] Found optimal step size 0.05103059485554695 with residual 0.0012035369873046875

[DiffTRe] Epoch 16
	Epoch loss = 0.04692
	Gradient norm: 0.028800420463085175
	Elapsed time = 1.772 min
[Statepoint 0]
	kT = 2.472 ref_kT = 2.462
	Predicted entropy: 3.411777973175049
	Predicted free_energy: -2326.30322265625
	Predicted pressure: 328.76190185546875
[Step Size] Found optimal step size 0.022778958082199097 with residual 0.0005266666412353516

[DiffTRe] Epoch 17
	Epoch loss = 0.04380
	Gradient norm: 0.03282453119754791
	Elapsed time = 1.775 min
[Statepoint 0]
	kT = 2.458 ref_kT = 2.462
	Predicted entropy: 3.5940446853637695
	Predicted free_energy: -2198.33642578125
	Predicted pressure: 344.1204528808594
[Step Size] Found optimal step size 0.042089417576789856 with residual 0.0003368854522705078

[DiffTRe] Epoch 18
	Epoch loss = 0.04272
	Gradient norm: 0.01439292449504137
	Elapsed time = 1.773 min
[Statepoint 0]
	kT = 2.475 ref_kT = 2.462
	Predicted entropy: 3.51694917678833
	Predicted free_energy: -2463.50927734375
	Predicted pressure: 323.17425537109375
[Step Size] Found optimal step size 0.030478304252028465 with residual 0.0018916130065917969

[DiffTRe] Epoch 19
	Epoch loss = 0.04142
	Gradient norm: 0.03944658860564232
	Elapsed time = 1.776 min
[Statepoint 0]
	kT = 2.465 ref_kT = 2.462
	Predicted entropy: 3.6841933727264404
	Predicted free_energy: -2403.00634765625
	Predicted pressure: 331.78375244140625
[Step Size] Found optimal step size 0.029037784785032272 with residual 0.003751039505004883

[DiffTRe] Epoch 20
	Epoch loss = 0.04037
	Gradient norm: 0.036156658083200455
	Elapsed time = 1.771 min
[Statepoint 0]
	kT = 2.460 ref_kT = 2.462
	Predicted entropy: 3.667158603668213
	Predicted free_energy: -2621.615234375
	Predicted pressure: 311.8834228515625
[Step Size] Found optimal step size 0.046243343502283096 with residual 7.653236389160156e-05

[DiffTRe] Epoch 21
	Epoch loss = 0.03969
	Gradient norm: 0.009144319221377373
	Elapsed time = 1.771 min
[Statepoint 0]
	kT = 2.464 ref_kT = 2.462
	Predicted entropy: 3.8678700923919678
	Predicted free_energy: -2359.218994140625
	Predicted pressure: 336.85516357421875
[Step Size] Found optimal step size 0.03873026743531227 with residual 0.0018033981323242188

[DiffTRe] Epoch 22
	Epoch loss = 0.03900
	Gradient norm: 0.013603090308606625
	Elapsed time = 1.772 min
[Statepoint 0]
	kT = 2.464 ref_kT = 2.462
	Predicted entropy: 4.040920734405518
	Predicted free_energy: -2205.065185546875
	Predicted pressure: 351.02789306640625
[Step Size] Found optimal step size 0.02787666954100132 with residual 0.0009717941284179688

[DiffTRe] Epoch 23
	Epoch loss = 0.03771
	Gradient norm: 0.01812558062374592
	Elapsed time = 1.772 min
[Statepoint 0]
	kT = 2.466 ref_kT = 2.462
	Predicted entropy: 3.9732563495635986
	Predicted free_energy: -2650.9296875
	Predicted pressure: 317.5606994628906
[Step Size] Found optimal step size 0.15628191828727722 with residual 0.00869607925415039

[DiffTRe] Epoch 24
	Epoch loss = 0.03743
	Gradient norm: 0.004155977629125118
	Elapsed time = 1.773 min
[Statepoint 0]
	kT = 2.460 ref_kT = 2.462
	Predicted entropy: 4.013966083526611
	Predicted free_energy: -2648.177734375
	Predicted pressure: 322.7134094238281
[Step Size] Found optimal step size 0.10997407138347626 with residual 0.0003933906555175781

[DiffTRe] Epoch 25
	Epoch loss = 0.03468
	Gradient norm: 0.004014665726572275
	Elapsed time = 1.778 min
[Statepoint 0]
	kT = 2.461 ref_kT = 2.462
	Predicted entropy: 4.032769680023193
	Predicted free_energy: -3662.820068359375
	Predicted pressure: 276.0408935546875
[Step Size] Found optimal step size 0.04175412654876709 with residual 0.00078582763671875

[DiffTRe] Epoch 26
	Epoch loss = 0.03312
	Gradient norm: 0.01278436928987503
	Elapsed time = 1.776 min
[Statepoint 0]
	kT = 2.468 ref_kT = 2.462
	Predicted entropy: 3.942852020263672
	Predicted free_energy: -4499.44921875
	Predicted pressure: 232.03041076660156
[Step Size] Found optimal step size 0.03838876262307167 with residual 0.0007216930389404297

[DiffTRe] Epoch 27
	Epoch loss = 0.03229
	Gradient norm: 0.012335007078945637
	Elapsed time = 1.777 min
[Statepoint 0]
	kT = 2.462 ref_kT = 2.462
	Predicted entropy: 4.069455623626709
	Predicted free_energy: -4221.78564453125
	Predicted pressure: 254.82298278808594
[Step Size] Found optimal step size 0.04727406054735184 with residual 0.00501251220703125

[DiffTRe] Epoch 28
	Epoch loss = 0.03128
	Gradient norm: 0.00808016862720251
	Elapsed time = 1.774 min
[Statepoint 0]
	kT = 2.462 ref_kT = 2.462
	Predicted entropy: 4.016363620758057
	Predicted free_energy: -4605.6005859375
	Predicted pressure: 229.2788848876953
[Step Size] Found optimal step size 0.0360044501721859 with residual 0.0009479522705078125

[DiffTRe] Epoch 29
	Epoch loss = 0.03194
	Gradient norm: 0.007689347490668297
	Elapsed time = 1.772 min
[Statepoint 0]
	kT = 2.461 ref_kT = 2.462
	Predicted entropy: 4.148475646972656
	Predicted free_energy: -4295.06689453125
	Predicted pressure: 259.0669860839844
[Step Size] Found optimal step size 0.044150859117507935 with residual 0.0034461021423339844

[DiffTRe] Epoch 30
	Epoch loss = 0.03118
	Gradient norm: 0.007173060905188322
	Elapsed time = 1.778 min
[Statepoint 0]
	kT = 2.468 ref_kT = 2.462
	Predicted entropy: 4.053627014160156
	Predicted free_energy: -4744.66552734375
	Predicted pressure: 231.0420684814453
[Step Size] Found optimal step size 0.12943977117538452 with residual 0.0004916191101074219

[DiffTRe] Epoch 31
	Epoch loss = 0.03030
	Gradient norm: 0.0024557311553508043
	Elapsed time = 1.778 min
[Statepoint 0]
	kT = 2.466 ref_kT = 2.462
	Predicted entropy: 4.000321388244629
	Predicted free_energy: -4885.921875
	Predicted pressure: 225.909423828125
[Step Size] Found optimal step size 0.035942357033491135 with residual 0.0019385814666748047

[DiffTRe] Epoch 32
	Epoch loss = 0.02873
	Gradient norm: 0.0050691901706159115
	Elapsed time = 1.771 min
[Statepoint 0]
	kT = 2.467 ref_kT = 2.462
	Predicted entropy: 3.967437505722046
	Predicted free_energy: -5591.423828125
	Predicted pressure: 201.7595977783203
[Step Size] Found optimal step size 0.14904826879501343 with residual 0.00017261505126953125

[DiffTRe] Epoch 33
	Epoch loss = 0.02857
	Gradient norm: 0.00201980397105217
	Elapsed time = 1.777 min
[Statepoint 0]
	kT = 2.471 ref_kT = 2.462
	Predicted entropy: 3.971832036972046
	Predicted free_energy: -5390.779296875
	Predicted pressure: 216.9598846435547
[Step Size] Found optimal step size 0.045336805284023285 with residual 0.00010538101196289062

[DiffTRe] Epoch 34
	Epoch loss = 0.02723
	Gradient norm: 0.00808955729007721
	Elapsed time = 1.773 min
[Statepoint 0]
	kT = 2.465 ref_kT = 2.462
	Predicted entropy: 3.8606112003326416
	Predicted free_energy: -5391.849609375
	Predicted pressure: 240.05393981933594
[Step Size] Found optimal step size 0.05608484148979187 with residual 0.0007383823394775391

[DiffTRe] Epoch 35
	Epoch loss = 0.02642
	Gradient norm: 0.002505391137674451
	Elapsed time = 1.774 min
[Statepoint 0]
	kT = 2.454 ref_kT = 2.462
	Predicted entropy: 3.7820117473602295
	Predicted free_energy: -6062.89892578125
	Predicted pressure: 199.27593994140625
[Step Size] Found optimal step size 0.14888061583042145 with residual 0.0009665489196777344

[DiffTRe] Epoch 36
	Epoch loss = 0.02569
	Gradient norm: 0.001884533790871501
	Elapsed time = 1.775 min
[Statepoint 0]
	kT = 2.463 ref_kT = 2.462
	Predicted entropy: 3.790658712387085
	Predicted free_energy: -6169.99072265625
	Predicted pressure: 198.36306762695312
[Step Size] Found optimal step size 0.03149660676717758 with residual 0.004755973815917969

[DiffTRe] Epoch 37
	Epoch loss = 0.02430
	Gradient norm: 0.006879275199025869
	Elapsed time = 1.778 min
[Statepoint 0]
	kT = 2.463 ref_kT = 2.462
	Predicted entropy: 3.7984230518341064
	Predicted free_energy: -6855.427734375
	Predicted pressure: 185.25912475585938
[Step Size] Found optimal step size 0.02828647568821907 with residual 0.0012350082397460938

[DiffTRe] Epoch 38
	Epoch loss = 0.02506
	Gradient norm: 0.03258499503135681
	Elapsed time = 1.770 min
[Statepoint 0]
	kT = 2.469 ref_kT = 2.462
	Predicted entropy: 3.7360639572143555
	Predicted free_energy: -6673.1865234375
	Predicted pressure: 200.61831665039062
[Step Size] Found optimal step size 0.07684825360774994 with residual 0.00167083740234375

[DiffTRe] Epoch 39
	Epoch loss = 0.02391
	Gradient norm: 0.0018712279852479696
	Elapsed time = 1.775 min
[Statepoint 0]
	kT = 2.460 ref_kT = 2.462
	Predicted entropy: 3.6745448112487793
	Predicted free_energy: -7154.45361328125
	Predicted pressure: 167.97518920898438
[Step Size] Found optimal step size 0.043790727853775024 with residual 0.0021452903747558594

[DiffTRe] Epoch 40
	Epoch loss = 0.02326
	Gradient norm: 0.004330873489379883
	Elapsed time = 1.772 min
[Statepoint 0]
	kT = 2.458 ref_kT = 2.462
	Predicted entropy: 3.571645975112915
	Predicted free_energy: -7411.57421875
	Predicted pressure: 154.16590881347656
[Step Size] Found optimal step size 0.04433713108301163 with residual 0.0006072521209716797

[DiffTRe] Epoch 41
	Epoch loss = 0.02339
	Gradient norm: 0.0024997121654450893
	Elapsed time = 1.773 min
[Statepoint 0]
	kT = 2.468 ref_kT = 2.462
	Predicted entropy: 3.561678171157837
	Predicted free_energy: -7620.24267578125
	Predicted pressure: 150.06204223632812
[Step Size] Found optimal step size 0.026610001921653748 with residual 0.005694866180419922

[DiffTRe] Epoch 42
	Epoch loss = 0.02276
	Gradient norm: 0.007961194030940533
	Elapsed time = 1.769 min
[Statepoint 0]
	kT = 2.459 ref_kT = 2.462
	Predicted entropy: 3.5570547580718994
	Predicted free_energy: -7530.6181640625
	Predicted pressure: 157.1514892578125
[Step Size] Found optimal step size 0.040580593049526215 with residual 0.012398242950439453

[DiffTRe] Epoch 43
	Epoch loss = 0.02331
	Gradient norm: 0.008803782053291798
	Elapsed time = 1.774 min
[Statepoint 0]
	kT = 2.457 ref_kT = 2.462
	Predicted entropy: 3.546964645385742
	Predicted free_energy: -7275.82373046875
	Predicted pressure: 172.4276580810547
[Step Size] Found optimal step size 0.03412307798862457 with residual 0.0007264614105224609

[DiffTRe] Epoch 44
	Epoch loss = 0.02267
	Gradient norm: 0.007650857791304588
	Elapsed time = 1.774 min
[Statepoint 0]
	kT = 2.458 ref_kT = 2.462
	Predicted entropy: 3.542985200881958
	Predicted free_energy: -7642.275390625
	Predicted pressure: 160.2111358642578
[Step Size] Found optimal step size 0.13435739278793335 with residual 0.0003256797790527344

[DiffTRe] Epoch 45
	Epoch loss = 0.02212
	Gradient norm: 0.0022009985987097025
	Elapsed time = 1.770 min
[Statepoint 0]
	kT = 2.457 ref_kT = 2.462
	Predicted entropy: 3.598086357116699
	Predicted free_energy: -7654.90625
	Predicted pressure: 154.29722595214844
[Step Size] Found optimal step size 0.06178484112024307 with residual 0.002057313919067383

[DiffTRe] Epoch 46
	Epoch loss = 0.02078
	Gradient norm: 0.0029711571987718344
	Elapsed time = 1.772 min
[Statepoint 0]
	kT = 2.462 ref_kT = 2.462
	Predicted entropy: 3.513646125793457
	Predicted free_energy: -7559.49951171875
	Predicted pressure: 160.2257080078125
[Step Size] Found optimal step size 0.04467242583632469 with residual 7.62939453125e-06

[DiffTRe] Epoch 47
	Epoch loss = 0.02054
	Gradient norm: 0.003572851186618209
	Elapsed time = 1.770 min
[Statepoint 0]
	kT = 2.459 ref_kT = 2.462
	Predicted entropy: 3.4854729175567627
	Predicted free_energy: -7805.75244140625
	Predicted pressure: 147.81568908691406
[Step Size] Found optimal step size 0.04303941875696182 with residual 0.006447792053222656

[DiffTRe] Epoch 48
	Epoch loss = 0.02002
	Gradient norm: 0.009161151014268398
	Elapsed time = 1.772 min
[Statepoint 0]
	kT = 2.467 ref_kT = 2.462
	Predicted entropy: 3.3846256732940674
	Predicted free_energy: -7781.650390625
	Predicted pressure: 151.1243438720703
[Step Size] Found optimal step size 0.02950967848300934 with residual 0.0008013248443603516

[DiffTRe] Epoch 49
	Epoch loss = 0.02019
	Gradient norm: 0.010777447372674942
	Elapsed time = 1.770 min
[Statepoint 0]
	kT = 2.468 ref_kT = 2.462
	Predicted entropy: 3.4592113494873047
	Predicted free_energy: -7645.369140625
	Predicted pressure: 164.0375213623047
[Step Size] Found optimal step size 0.03108680248260498 with residual 0.000347137451171875

[DiffTRe] Epoch 50
	Epoch loss = 0.01964
	Gradient norm: 0.01743745058774948
	Elapsed time = 1.769 min
[Statepoint 0]
	kT = 2.464 ref_kT = 2.462
	Predicted entropy: 3.4200611114501953
	Predicted free_energy: -7682.56591796875
	Predicted pressure: 161.92794799804688
[Step Size] Found optimal step size 0.14412438869476318 with residual 0.0010724067687988281

[DiffTRe] Epoch 51
	Epoch loss = 0.01917
	Gradient norm: 0.0012696434278041124
	Elapsed time = 1.774 min
[Statepoint 0]
	kT = 2.466 ref_kT = 2.462
	Predicted entropy: 3.3434698581695557
	Predicted free_energy: -8140.92431640625
	Predicted pressure: 135.39723205566406
[Step Size] Found optimal step size 0.02919922024011612 with residual 0.0013861656188964844

[DiffTRe] Epoch 52
	Epoch loss = 0.01881
	Gradient norm: 0.02161470055580139
	Elapsed time = 1.773 min
[Statepoint 0]
	kT = 2.465 ref_kT = 2.462
	Predicted entropy: 3.1596131324768066
	Predicted free_energy: -8603.34765625
	Predicted pressure: 117.39603424072266
[Step Size] Found optimal step size 0.04276621714234352 with residual 0.008692741394042969

[DiffTRe] Epoch 53
	Epoch loss = 0.01862
	Gradient norm: 0.006205528974533081
	Elapsed time = 1.771 min
[Statepoint 0]
	kT = 2.463 ref_kT = 2.462
	Predicted entropy: 3.230635166168213
	Predicted free_energy: -8245.4169921875
	Predicted pressure: 137.1189727783203
[Step Size] Found optimal step size 0.3002597689628601 with residual 0.00032806396484375

[DiffTRe] Epoch 54
	Epoch loss = 0.01819
	Gradient norm: 0.0021841914858669043
	Elapsed time = 1.772 min
[Statepoint 0]
	kT = 2.466 ref_kT = 2.462
	Predicted entropy: 3.2317585945129395
	Predicted free_energy: -8054.10546875
	Predicted pressure: 148.38095092773438
[Step Size] Found optimal step size 0.04302079230546951 with residual 0.0014622211456298828

[DiffTRe] Epoch 55
	Epoch loss = 0.01655
	Gradient norm: 0.007797648198902607
	Elapsed time = 1.769 min
[Statepoint 0]
	kT = 2.462 ref_kT = 2.462
	Predicted entropy: 2.911323070526123
	Predicted free_energy: -8654.2021484375
	Predicted pressure: 134.9159698486328
[Step Size] Found optimal step size 0.0760224312543869 with residual 0.0029044151306152344

[DiffTRe] Epoch 56
	Epoch loss = 0.01557
	Gradient norm: 0.002237932290881872
	Elapsed time = 1.777 min
[Statepoint 0]
	kT = 2.467 ref_kT = 2.462
	Predicted entropy: 2.780139446258545
	Predicted free_energy: -9145.6796875
	Predicted pressure: 109.39685821533203
[Step Size] Found optimal step size 0.02768418751657009 with residual 0.008258819580078125

[DiffTRe] Epoch 57
	Epoch loss = 0.01546
	Gradient norm: 0.010906880721449852
	Elapsed time = 1.773 min
[Statepoint 0]
	kT = 2.460 ref_kT = 2.462
	Predicted entropy: 2.688987970352173
	Predicted free_energy: -9821.7685546875
	Predicted pressure: 78.65141296386719
[Step Size] Found optimal step size 0.05929497629404068 with residual 0.0006308555603027344

[DiffTRe] Epoch 58
	Epoch loss = 0.01509
	Gradient norm: 0.0032003659289330244
	Elapsed time = 1.769 min
[Statepoint 0]
	kT = 2.458 ref_kT = 2.462
	Predicted entropy: 2.7652997970581055
	Predicted free_energy: -9426.9072265625
	Predicted pressure: 101.85169982910156
[Step Size] Found optimal step size 0.028007064014673233 with residual 0.0024328231811523438

[DiffTRe] Epoch 59
	Epoch loss = 0.01437
	Gradient norm: 0.011273509822785854
	Elapsed time = 1.771 min
[Statepoint 0]
	kT = 2.465 ref_kT = 2.462
	Predicted entropy: 2.77643084526062
	Predicted free_energy: -9010.1630859375
	Predicted pressure: 123.13855743408203
[Step Size] Found optimal step size 0.06415053457021713 with residual 0.0006976127624511719

[DiffTRe] Epoch 60
	Epoch loss = 0.01417
	Gradient norm: 0.00353602203540504
	Elapsed time = 1.771 min
[Statepoint 0]
	kT = 2.460 ref_kT = 2.462
	Predicted entropy: 2.714846134185791
	Predicted free_energy: -9381.2783203125
	Predicted pressure: 106.521240234375
[Step Size] Found optimal step size 0.07970446348190308 with residual 0.001405954360961914

[DiffTRe] Epoch 61
	Epoch loss = 0.01381
	Gradient norm: 0.0032279775477945805
	Elapsed time = 1.770 min
[Statepoint 0]
	kT = 2.467 ref_kT = 2.462
	Predicted entropy: 2.5572266578674316
	Predicted free_energy: -9668.9384765625
	Predicted pressure: 93.84137725830078
[Step Size] Found optimal step size 0.07493582367897034 with residual 0.0015723705291748047

[DiffTRe] Epoch 62
	Epoch loss = 0.01302
	Gradient norm: 0.00221348088234663
	Elapsed time = 1.769 min
[Statepoint 0]
	kT = 2.462 ref_kT = 2.462
	Predicted entropy: 2.5440609455108643
	Predicted free_energy: -9392.7724609375
	Predicted pressure: 108.91526794433594
[Step Size] Found optimal step size 0.10975674539804459 with residual 0.0003857612609863281

[DiffTRe] Epoch 63
	Epoch loss = 0.01280
	Gradient norm: 0.0032730766106396914
	Elapsed time = 1.772 min
[Statepoint 0]
	kT = 2.463 ref_kT = 2.462
	Predicted entropy: 2.504823923110962
	Predicted free_energy: -9594.3369140625
	Predicted pressure: 107.47467803955078
[Step Size] Found optimal step size 0.04200248792767525 with residual 0.0019233226776123047

[DiffTRe] Epoch 64
	Epoch loss = 0.01163
	Gradient norm: 0.00728271808475256
	Elapsed time = 1.772 min
[Statepoint 0]
	kT = 2.464 ref_kT = 2.462
	Predicted entropy: 2.529414415359497
	Predicted free_energy: -9858.0400390625
	Predicted pressure: 93.89028930664062
[Step Size] Found optimal step size 0.03361392393708229 with residual 0.0002562999725341797

[DiffTRe] Epoch 65
	Epoch loss = 0.01138
	Gradient norm: 0.00901087373495102
	Elapsed time = 1.771 min
[Statepoint 0]
	kT = 2.470 ref_kT = 2.462
	Predicted entropy: 2.396146774291992
	Predicted free_energy: -10475.2490234375
	Predicted pressure: 72.30216979980469
[Step Size] Found optimal step size 0.05376261845231056 with residual 0.0006253719329833984

[DiffTRe] Epoch 66
	Epoch loss = 0.01127
	Gradient norm: 0.0037447549402713776
	Elapsed time = 1.769 min
[Statepoint 0]
	kT = 2.457 ref_kT = 2.462
	Predicted entropy: 2.4331777095794678
	Predicted free_energy: -10000.9365234375
	Predicted pressure: 97.61521911621094
[Step Size] Found optimal step size 0.05396752059459686 with residual 0.0003514289855957031

[DiffTRe] Epoch 67
	Epoch loss = 0.01094
	Gradient norm: 0.004873099736869335
	Elapsed time = 1.773 min
[Statepoint 0]
	kT = 2.463 ref_kT = 2.462
	Predicted entropy: 2.2652106285095215
	Predicted free_energy: -10300.7119140625
	Predicted pressure: 88.89798736572266
[Step Size] Found optimal step size 0.03199955075979233 with residual 0.006774425506591797

[DiffTRe] Epoch 68
	Epoch loss = 0.01071
	Gradient norm: 0.019018474966287613
	Elapsed time = 1.773 min
[Statepoint 0]
	kT = 2.461 ref_kT = 2.462
	Predicted entropy: 2.204375982284546
	Predicted free_energy: -10675.3330078125
	Predicted pressure: 69.13204193115234
[Step Size] Found optimal step size 0.06865216791629791 with residual 0.000843048095703125

[DiffTRe] Epoch 69
	Epoch loss = 0.01032
	Gradient norm: 0.0038618596736341715
	Elapsed time = 1.768 min
[Statepoint 0]
	kT = 2.461 ref_kT = 2.462
	Predicted entropy: 2.264308214187622
	Predicted free_energy: -10342.8330078125
	Predicted pressure: 83.11007690429688
[Step Size] Found optimal step size 0.05997798219323158 with residual 0.0015430450439453125

[DiffTRe] Epoch 70
	Epoch loss = 0.00989
	Gradient norm: 0.005733059719204903
	Elapsed time = 1.770 min
[Statepoint 0]
	kT = 2.469 ref_kT = 2.462
	Predicted entropy: 2.230403184890747
	Predicted free_energy: -10280.5029296875
	Predicted pressure: 87.54673767089844
[Step Size] Found optimal step size 0.047379616647958755 with residual 0.004685640335083008

[DiffTRe] Epoch 71
	Epoch loss = 0.00964
	Gradient norm: 0.008825409226119518
	Elapsed time = 1.771 min
[Statepoint 0]
	kT = 2.462 ref_kT = 2.462
	Predicted entropy: 2.1868948936462402
	Predicted free_energy: -10845.8828125
	Predicted pressure: 65.38516998291016
[Step Size] Found optimal step size 0.058444324880838394 with residual 0.0016014575958251953

[DiffTRe] Epoch 72
	Epoch loss = 0.00958
	Gradient norm: 0.010151405818760395
	Elapsed time = 1.772 min
[Statepoint 0]
	kT = 2.459 ref_kT = 2.462
	Predicted entropy: 2.2830188274383545
	Predicted free_energy: -10433.2939453125
	Predicted pressure: 87.47785949707031
[Step Size] Found optimal step size 0.05984137952327728 with residual 0.002227306365966797

[DiffTRe] Epoch 73
	Epoch loss = 0.00904
	Gradient norm: 0.0061097778379917145
	Elapsed time = 1.771 min
[Statepoint 0]
	kT = 2.457 ref_kT = 2.462
	Predicted entropy: 2.0240530967712402
	Predicted free_energy: -10910.3916015625
	Predicted pressure: 68.04254150390625
[Step Size] Found optimal step size 0.0373704619705677 with residual 0.010038614273071289

[DiffTRe] Epoch 74
	Epoch loss = 0.00855
	Gradient norm: 0.011017464101314545
	Elapsed time = 1.770 min
[Statepoint 0]
	kT = 2.458 ref_kT = 2.462
	Predicted entropy: 2.0465686321258545
	Predicted free_energy: -10780.3173828125
	Predicted pressure: 71.87171936035156
[Step Size] Found optimal step size 0.055221762508153915 with residual 0.0011942386627197266

[DiffTRe] Epoch 75
	Epoch loss = 0.00831
	Gradient norm: 0.004701926838606596
	Elapsed time = 1.772 min
[Statepoint 0]
	kT = 2.459 ref_kT = 2.462
	Predicted entropy: 2.0780608654022217
	Predicted free_energy: -10614.900390625
	Predicted pressure: 80.10456848144531
[Step Size] Found optimal step size 0.17356199026107788 with residual 0.060565948486328125

[DiffTRe] Epoch 76
	Epoch loss = 0.00820
	Gradient norm: 0.0035037104971706867
	Elapsed time = 1.771 min
[Statepoint 0]
	kT = 2.462 ref_kT = 2.462
	Predicted entropy: 1.9392821788787842
	Predicted free_energy: -11193.203125
	Predicted pressure: 64.30824279785156
[Step Size] Found optimal step size 0.03320412337779999 with residual 0.024654388427734375

[DiffTRe] Epoch 77
	Epoch loss = 0.00741
	Gradient norm: 0.011143147945404053
	Elapsed time = 1.766 min
[Statepoint 0]
	kT = 2.466 ref_kT = 2.462
	Predicted entropy: 1.5679961442947388
	Predicted free_energy: -12255.2626953125
	Predicted pressure: 14.499754905700684
[Step Size] Found optimal step size 0.0915328860282898 with residual 0.003620147705078125

[DiffTRe] Epoch 78
	Epoch loss = 0.00682
	Gradient norm: 0.002394957933574915
	Elapsed time = 1.769 min
[Statepoint 0]
	kT = 2.465 ref_kT = 2.462
	Predicted entropy: 1.6865663528442383
	Predicted free_energy: -11729.025390625
	Predicted pressure: 40.93570327758789
[Step Size] Found optimal step size 0.0746750459074974 with residual 0.0007567405700683594

[DiffTRe] Epoch 79
	Epoch loss = 0.00634
	Gradient norm: 0.0027354941703379154
	Elapsed time = 1.771 min
[Statepoint 0]
	kT = 2.450 ref_kT = 2.462
	Predicted entropy: 1.762482762336731
	Predicted free_energy: -11434.109375
	Predicted pressure: 56.16904830932617
[Step Size] Found optimal step size 0.0413629487156868 with residual 0.0020532608032226562

[DiffTRe] Epoch 80
	Epoch loss = 0.00609
	Gradient norm: 0.012689990922808647
	Elapsed time = 1.769 min
[Statepoint 0]
	kT = 2.461 ref_kT = 2.462
	Predicted entropy: 1.6242798566818237
	Predicted free_energy: -11807.64453125
	Predicted pressure: 39.959495544433594
[Step Size] Found optimal step size 0.054855428636074066 with residual 0.005585908889770508

[DiffTRe] Epoch 81
	Epoch loss = 0.00575
	Gradient norm: 0.003840063000097871
	Elapsed time = 1.768 min
[Statepoint 0]
	kT = 2.461 ref_kT = 2.462
	Predicted entropy: 1.527915358543396
	Predicted free_energy: -11748.4599609375
	Predicted pressure: 47.61982727050781
[Step Size] Found optimal step size 0.027261966839432716 with residual 0.02236008644104004

[DiffTRe] Epoch 82
	Epoch loss = 0.00560
	Gradient norm: 0.011446226388216019
	Elapsed time = 1.766 min
[Statepoint 0]
	kT = 2.467 ref_kT = 2.462
	Predicted entropy: 1.565551996231079
	Predicted free_energy: -11462.107421875
	Predicted pressure: 59.21560287475586
[Step Size] Found optimal step size 0.0637345239520073 with residual 0.014830827713012695

[DiffTRe] Epoch 83
	Epoch loss = 0.00540
	Gradient norm: 0.0029903065878897905
	Elapsed time = 1.763 min
[Statepoint 0]
	kT = 2.456 ref_kT = 2.462
	Predicted entropy: 1.4084941148757935
	Predicted free_energy: -11835.341796875
	Predicted pressure: 45.845890045166016
[Step Size] Found optimal step size 0.04467863216996193 with residual 0.0011227130889892578

[DiffTRe] Epoch 84
	Epoch loss = 0.00524
	Gradient norm: 0.011893714778125286
	Elapsed time = 1.767 min
[Statepoint 0]
	kT = 2.459 ref_kT = 2.462
	Predicted entropy: 1.1590121984481812
	Predicted free_energy: -12303.0615234375
	Predicted pressure: 25.02168083190918
[Step Size] Found optimal step size 0.05652569234371185 with residual 0.0006198883056640625

[DiffTRe] Epoch 85
	Epoch loss = 0.00481
	Gradient norm: 0.0071122352965176105
	Elapsed time = 1.768 min
[Statepoint 0]
	kT = 2.469 ref_kT = 2.462
	Predicted entropy: 1.331125259399414
	Predicted free_energy: -11733.89453125
	Predicted pressure: 53.4546012878418
[Step Size] Found optimal step size 0.04641098901629448 with residual 0.0019674301147460938

[DiffTRe] Epoch 86
	Epoch loss = 0.00492
	Gradient norm: 0.008297842927277088
	Elapsed time = 1.764 min
[Statepoint 0]
	kT = 2.471 ref_kT = 2.462
	Predicted entropy: 1.1617870330810547
	Predicted free_energy: -12102.3251953125
	Predicted pressure: 39.1763916015625
[Step Size] Found optimal step size 0.13097962737083435 with residual 0.0004506111145019531

[DiffTRe] Epoch 87
	Epoch loss = 0.00459
	Gradient norm: 0.001578209688887
	Elapsed time = 1.764 min
[Statepoint 0]
	kT = 2.472 ref_kT = 2.462
	Predicted entropy: 1.2764118909835815
	Predicted free_energy: -11999.529296875
	Predicted pressure: 43.633358001708984
[Step Size] Found optimal step size 0.046454451978206635 with residual 0.008022785186767578

[DiffTRe] Epoch 88
	Epoch loss = 0.00411
	Gradient norm: 0.004014844074845314
	Elapsed time = 1.765 min
[Statepoint 0]
	kT = 2.457 ref_kT = 2.462
	Predicted entropy: 1.198604702949524
	Predicted free_energy: -11908.236328125
	Predicted pressure: 46.60331726074219
[Step Size] Found optimal step size 0.046280600130558014 with residual 0.005106925964355469

[DiffTRe] Epoch 89
	Epoch loss = 0.00393
	Gradient norm: 0.00346653675660491
	Elapsed time = 1.761 min
[Statepoint 0]
	kT = 2.462 ref_kT = 2.462
	Predicted entropy: 1.0015136003494263
	Predicted free_energy: -12414.6416015625
	Predicted pressure: 32.64585494995117
[Step Size] Found optimal step size 0.03122340328991413 with residual 0.0016050338745117188

[DiffTRe] Epoch 90
	Epoch loss = 0.00374
	Gradient norm: 0.009350153617560863
	Elapsed time = 1.762 min
[Statepoint 0]
	kT = 2.465 ref_kT = 2.462
	Predicted entropy: 1.1201905012130737
	Predicted free_energy: -12077.1337890625
	Predicted pressure: 39.58339309692383
[Step Size] Found optimal step size 0.10802438855171204 with residual 0.0003669261932373047

[DiffTRe] Epoch 91
	Epoch loss = 0.00367
	Gradient norm: 0.0016614142805337906
	Elapsed time = 1.762 min
[Statepoint 0]
	kT = 2.464 ref_kT = 2.462
	Predicted entropy: 0.919154703617096
	Predicted free_energy: -12486.99609375
	Predicted pressure: 26.438907623291016
[Step Size] Found optimal step size 0.032179608941078186 with residual 0.019947052001953125

[DiffTRe] Epoch 92
	Epoch loss = 0.00338
	Gradient norm: 0.009070460684597492
	Elapsed time = 1.763 min
[Statepoint 0]
	kT = 2.454 ref_kT = 2.462
	Predicted entropy: 1.0939300060272217
	Predicted free_energy: -12285.125
	Predicted pressure: 32.65285873413086
[Step Size] Found optimal step size 0.036439087241888046 with residual 0.0012977123260498047

[DiffTRe] Epoch 93
	Epoch loss = 0.00323
	Gradient norm: 0.004853527527302504
	Elapsed time = 1.761 min
[Statepoint 0]
	kT = 2.458 ref_kT = 2.462
	Predicted entropy: 0.863994836807251
	Predicted free_energy: -12637.169921875
	Predicted pressure: 22.98594856262207
[Step Size] Found optimal step size 0.04556654393672943 with residual 0.0008378028869628906

[DiffTRe] Epoch 94
	Epoch loss = 0.00324
	Gradient norm: 0.002976309973746538
	Elapsed time = 1.761 min
[Statepoint 0]
	kT = 2.451 ref_kT = 2.462
	Predicted entropy: 0.9951027035713196
	Predicted free_energy: -12253.5166015625
	Predicted pressure: 37.0708122253418
[Step Size] Found optimal step size 0.0642126202583313 with residual 0.006932735443115234

[DiffTRe] Epoch 95
	Epoch loss = 0.00308
	Gradient norm: 0.0018147379159927368
	Elapsed time = 1.763 min
[Statepoint 0]
	kT = 2.464 ref_kT = 2.462
	Predicted entropy: 0.8835887908935547
	Predicted free_energy: -12504.3046875
	Predicted pressure: 27.69625473022461
[Step Size] Found optimal step size 0.0345887616276741 with residual 0.009899616241455078

[DiffTRe] Epoch 96
	Epoch loss = 0.00311
	Gradient norm: 0.006083385553210974
	Elapsed time = 1.763 min
[Statepoint 0]
	kT = 2.465 ref_kT = 2.462
	Predicted entropy: 0.5424898266792297
	Predicted free_energy: -12611.263671875
	Predicted pressure: 26.641977310180664
[Step Size] Found optimal step size 0.060592688620090485 with residual 0.0012607574462890625

[DiffTRe] Epoch 97
	Epoch loss = 0.00280
	Gradient norm: 0.001903607277199626
	Elapsed time = 1.761 min
[Statepoint 0]
	kT = 2.472 ref_kT = 2.462
	Predicted entropy: 0.6397003531455994
	Predicted free_energy: -12563.3603515625
	Predicted pressure: 27.702320098876953
[Step Size] Found optimal step size 0.0364825502038002 with residual 0.006972312927246094

[DiffTRe] Epoch 98
	Epoch loss = 0.00282
	Gradient norm: 0.010619016364216805
	Elapsed time = 1.760 min
[Statepoint 0]
	kT = 2.461 ref_kT = 2.462
	Predicted entropy: 0.7220222353935242
	Predicted free_energy: -12147.0244140625
	Predicted pressure: 44.13990020751953
[Step Size] Found optimal step size 0.05803452059626579 with residual 0.0003066062927246094

[DiffTRe] Epoch 99
	Epoch loss = 0.00263
	Gradient norm: 0.0021411851048469543
	Elapsed time = 1.758 min
[Statepoint 0]
	kT = 2.461 ref_kT = 2.462
	Predicted entropy: 0.4640861749649048
	Predicted free_energy: -12653.62890625
	Predicted pressure: 25.95458221435547
[Step Size] Found optimal step size 0.034290723502635956 with residual 0.0006198883056640625

[DiffTRe] Epoch 100
	Epoch loss = 0.00300
	Gradient norm: 0.019285578280687332
	Elapsed time = 1.760 min
[Statepoint 0]
	kT = 2.453 ref_kT = 2.462
	Predicted entropy: 0.15646213293075562
	Predicted free_energy: -13082.91796875
	Predicted pressure: 4.7167863845825195
[Step Size] Found optimal step size 0.0629335343837738 with residual 0.0006275177001953125

[DiffTRe] Epoch 101
	Epoch loss = 0.00255
	Gradient norm: 0.0013339672004804015
	Elapsed time = 1.756 min
[Statepoint 0]
	kT = 2.458 ref_kT = 2.462
	Predicted entropy: 0.3822709023952484
	Predicted free_energy: -12752.7197265625
	Predicted pressure: 20.4913272857666
[Step Size] Found optimal step size 0.037935495376586914 with residual 0.0008206367492675781

[DiffTRe] Epoch 102
	Epoch loss = 0.00260
	Gradient norm: 0.015374801121652126
	Elapsed time = 1.766 min
[Statepoint 0]
	kT = 2.457 ref_kT = 2.462
	Predicted entropy: 0.5518105030059814
	Predicted free_energy: -12547.4140625
	Predicted pressure: 24.87945556640625
[Step Size] Found optimal step size 0.0653737336397171 with residual 0.018573760986328125

[DiffTRe] Epoch 103
	Epoch loss = 0.00244
	Gradient norm: 0.003675394458696246
	Elapsed time = 1.766 min
[Statepoint 0]
	kT = 2.467 ref_kT = 2.462
	Predicted entropy: 0.3156879246234894
	Predicted free_energy: -12759.1181640625
	Predicted pressure: 18.770078659057617
[Step Size] Found optimal step size 0.05080706626176834 with residual 0.0018422603607177734

[DiffTRe] Epoch 104
	Epoch loss = 0.00243
	Gradient norm: 0.0105065256357193
	Elapsed time = 1.761 min
[Statepoint 0]
	kT = 2.464 ref_kT = 2.462
	Predicted entropy: 0.4142893850803375
	Predicted free_energy: -12646.3427734375
	Predicted pressure: 21.30179786682129
[Step Size] Found optimal step size 0.08265380561351776 with residual 0.04776906967163086

[DiffTRe] Epoch 105
	Epoch loss = 0.00238
	Gradient norm: 0.003102404996752739
	Elapsed time = 1.763 min
[Statepoint 0]
	kT = 2.464 ref_kT = 2.462
	Predicted entropy: 0.2627021074295044
	Predicted free_energy: -12555.892578125
	Predicted pressure: 31.093366622924805
[Step Size] Found optimal step size 0.062051836401224136 with residual 0.0038661956787109375

[DiffTRe] Epoch 106
	Epoch loss = 0.00235
	Gradient norm: 0.004009361378848553
	Elapsed time = 1.760 min
[Statepoint 0]
	kT = 2.453 ref_kT = 2.462
	Predicted entropy: 0.23996761441230774
	Predicted free_energy: -12802.984375
	Predicted pressure: 22.84467315673828
[Step Size] Found optimal step size 0.04140020161867142 with residual 0.0032317638397216797

[DiffTRe] Epoch 107
	Epoch loss = 0.00225
	Gradient norm: 0.0077646891586482525
	Elapsed time = 1.758 min
[Statepoint 0]
	kT = 2.468 ref_kT = 2.462
	Predicted entropy: 0.1919274479150772
	Predicted free_energy: -12683.84765625
	Predicted pressure: 26.9686336517334
[Step Size] Found optimal step size 0.04413843899965286 with residual 0.003528118133544922

[DiffTRe] Epoch 108
	Epoch loss = 0.00217
	Gradient norm: 0.008992878720164299
	Elapsed time = 1.758 min
[Statepoint 0]
	kT = 2.462 ref_kT = 2.462
	Predicted entropy: 0.2857632040977478
	Predicted free_energy: -12731.67578125
	Predicted pressure: 21.255739212036133
[Step Size] Found optimal step size 0.05216687172651291 with residual 0.0004925727844238281

[DiffTRe] Epoch 109
	Epoch loss = 0.00216
	Gradient norm: 0.00625055655837059
	Elapsed time = 1.755 min
[Statepoint 0]
	kT = 2.465 ref_kT = 2.462
	Predicted entropy: 0.04382599517703056
	Predicted free_energy: -13021.6484375
	Predicted pressure: 15.643587112426758
[Step Size] Found optimal step size 0.048844970762729645 with residual 0.021888256072998047

[DiffTRe] Epoch 110
	Epoch loss = 0.00201
	Gradient norm: 0.005135493353009224
	Elapsed time = 1.759 min
[Statepoint 0]
	kT = 2.453 ref_kT = 2.462
	Predicted entropy: 0.2768315076828003
	Predicted free_energy: -12626.8681640625
	Predicted pressure: 28.051048278808594
[Step Size] Found optimal step size 0.04395216703414917 with residual 0.0032334327697753906

[DiffTRe] Epoch 111
	Epoch loss = 0.00202
	Gradient norm: 0.006862219423055649
	Elapsed time = 1.757 min
[Statepoint 0]
	kT = 2.462 ref_kT = 2.462
	Predicted entropy: -0.020100997760891914
	Predicted free_energy: -13065.7431640625
	Predicted pressure: 11.889209747314453
[Step Size] Found optimal step size 0.058481574058532715 with residual 0.0017867088317871094

[DiffTRe] Epoch 112
	Epoch loss = 0.00203
	Gradient norm: 0.004167914390563965
	Elapsed time = 1.763 min
[Statepoint 0]
	kT = 2.461 ref_kT = 2.462
	Predicted entropy: 0.16953548789024353
	Predicted free_energy: -12841.31640625
	Predicted pressure: 16.713973999023438
[Step Size] Found optimal step size 0.04058059677481651 with residual 0.01518869400024414

[DiffTRe] Epoch 113
	Epoch loss = 0.00199
	Gradient norm: 0.00914955697953701
	Elapsed time = 1.759 min
[Statepoint 0]
	kT = 2.458 ref_kT = 2.462
	Predicted entropy: 0.06808525323867798
	Predicted free_energy: -12792.935546875
	Predicted pressure: 24.328807830810547
[Step Size] Found optimal step size 0.048844970762729645 with residual 0.0005273818969726562

[DiffTRe] Epoch 114
	Epoch loss = 0.00196
	Gradient norm: 0.011135824024677277
	Elapsed time = 1.761 min
[Statepoint 0]
	kT = 2.447 ref_kT = 2.462
	Predicted entropy: 0.17946304380893707
	Predicted free_energy: -12679.353515625
	Predicted pressure: 25.831710815429688
[Step Size] Found optimal step size 0.04739203304052353 with residual 0.004068136215209961

[DiffTRe] Epoch 115
	Epoch loss = 0.00191
	Gradient norm: 0.008313554339110851
	Elapsed time = 1.757 min
[Statepoint 0]
	kT = 2.461 ref_kT = 2.462
	Predicted entropy: -0.1524038016796112
	Predicted free_energy: -13195.228515625
	Predicted pressure: 2.730201005935669
[Step Size] Found optimal step size 0.0653737336397171 with residual 0.023496627807617188

[DiffTRe] Epoch 116
	Epoch loss = 0.00189
	Gradient norm: 0.005585110746324062
	Elapsed time = 1.761 min
[Statepoint 0]
	kT = 2.465 ref_kT = 2.462
	Predicted entropy: 0.08519010245800018
	Predicted free_energy: -12838.634765625
	Predicted pressure: 22.12131690979004
[Step Size] Found optimal step size 0.043908700346946716 with residual 0.0039653778076171875

[DiffTRe] Epoch 117
	Epoch loss = 0.00198
	Gradient norm: 0.018981384113430977
	Elapsed time = 1.757 min
[Statepoint 0]
	kT = 2.469 ref_kT = 2.462
	Predicted entropy: -0.22986096143722534
	Predicted free_energy: -13119.2568359375
	Predicted pressure: 1.4093257188796997
[Step Size] Found optimal step size 0.06240576505661011 with residual 0.0008120536804199219

[DiffTRe] Epoch 118
	Epoch loss = 0.00186
	Gradient norm: 0.006229656282812357
	Elapsed time = 1.760 min
[Statepoint 0]
	kT = 2.470 ref_kT = 2.462
	Predicted entropy: 0.04791192337870598
	Predicted free_energy: -12846.580078125
	Predicted pressure: 21.69281005859375
[Step Size] Found optimal step size 0.04575282335281372 with residual 0.00033283233642578125

[DiffTRe] Epoch 119
	Epoch loss = 0.00188
	Gradient norm: 0.013808513060212135
	Elapsed time = 1.759 min
[Statepoint 0]
	kT = 2.471 ref_kT = 2.462
	Predicted entropy: -0.3215567171573639
	Predicted free_energy: -13118.599609375
	Predicted pressure: 12.794898986816406
[Step Size] Found optimal step size 0.09423387795686722 with residual 0.0013904571533203125

[DiffTRe] Epoch 120
	Epoch loss = 0.00177
	Gradient norm: 0.0031782384030520916
	Elapsed time = 1.765 min
[Statepoint 0]
	kT = 2.466 ref_kT = 2.462
	Predicted entropy: 0.0988541841506958
	Predicted free_energy: -12847.44921875
	Predicted pressure: 23.438608169555664
[Step Size] Found optimal step size 0.048776671290397644 with residual 0.001859426498413086

[DiffTRe] Epoch 121
	Epoch loss = 0.00183
	Gradient norm: 0.010145519860088825
	Elapsed time = 1.759 min
[Statepoint 0]
	kT = 2.472 ref_kT = 2.462
	Predicted entropy: -0.2906384766101837
	Predicted free_energy: -13131.3916015625
	Predicted pressure: 7.40873384475708
[Step Size] Found optimal step size 0.06954006850719452 with residual 0.007155418395996094

[DiffTRe] Epoch 122
	Epoch loss = 0.00174
	Gradient norm: 0.004111647140234709
	Elapsed time = 1.757 min
[Statepoint 0]
	kT = 2.471 ref_kT = 2.462
	Predicted entropy: -0.0450051985681057
	Predicted free_energy: -12976.017578125
	Predicted pressure: 10.634429931640625
[Step Size] Found optimal step size 0.06505706906318665 with residual 0.0004973411560058594

[DiffTRe] Epoch 123
	Epoch loss = 0.00173
	Gradient norm: 0.004065077751874924
	Elapsed time = 1.759 min
[Statepoint 0]
	kT = 2.466 ref_kT = 2.462
	Predicted entropy: -0.11143087595701218
	Predicted free_energy: -12752.6435546875
	Predicted pressure: 21.60564613342285
[Step Size] Found optimal step size 0.11591622233390808 with residual 0.0034637451171875

[DiffTRe] Epoch 124
	Epoch loss = 0.00164
	Gradient norm: 0.0015628798864781857
	Elapsed time = 1.761 min
[Statepoint 0]
	kT = 2.466 ref_kT = 2.462
	Predicted entropy: -0.22638368606567383
	Predicted free_energy: -13152.017578125
	Predicted pressure: 10.63815689086914
[Step Size] Found optimal step size 0.0646907240152359 with residual 0.002288818359375

[DiffTRe] Epoch 125
	Epoch loss = 0.00167
	Gradient norm: 0.006004170514643192
	Elapsed time = 1.760 min
[Statepoint 0]
	kT = 2.458 ref_kT = 2.462
	Predicted entropy: -0.009441063739359379
	Predicted free_energy: -12655.015625
	Predicted pressure: 27.928586959838867
[Step Size] Found optimal step size 0.3932914137840271 with residual 0.0016369819641113281

[DiffTRe] Epoch 126
	Epoch loss = 0.00154
	Gradient norm: 0.0002944466832559556
	Elapsed time = 1.762 min
[Statepoint 0]
	kT = 2.455 ref_kT = 2.462
	Predicted entropy: -0.2570746839046478
	Predicted free_energy: -13228.314453125
	Predicted pressure: 4.871341705322266
[Step Size] Found optimal step size 0.056972749531269073 with residual 0.0034639835357666016

[DiffTRe] Epoch 127
	Epoch loss = 0.00166
	Gradient norm: 0.008040644228458405
	Elapsed time = 1.759 min
[Statepoint 0]
	kT = 2.458 ref_kT = 2.462
	Predicted entropy: -0.06659235805273056
	Predicted free_energy: -13239.6845703125
	Predicted pressure: 3.624687671661377
[Step Size] Found optimal step size 0.04310771822929382 with residual 0.005718231201171875

[DiffTRe] Epoch 128
	Epoch loss = 0.00152
	Gradient norm: 0.012709123082458973
	Elapsed time = 1.761 min
[Statepoint 0]
	kT = 2.460 ref_kT = 2.462
	Predicted entropy: -0.22031450271606445
	Predicted free_energy: -13034.2490234375
	Predicted pressure: 12.001991271972656
[Step Size] Found optimal step size 0.05383092164993286 with residual 0.004390716552734375

[DiffTRe] Epoch 129
	Epoch loss = 0.00152
	Gradient norm: 0.01277815643697977
	Elapsed time = 1.759 min
[Statepoint 0]
	kT = 2.469 ref_kT = 2.462
	Predicted entropy: 0.059834953397512436
	Predicted free_energy: -12789.232421875
	Predicted pressure: 30.20612335205078
[Step Size] Found optimal step size 0.060729287564754486 with residual 0.0028505325317382812

[DiffTRe] Epoch 130
	Epoch loss = 0.00150
	Gradient norm: 0.012666875496506691
	Elapsed time = 1.757 min
[Statepoint 0]
	kT = 2.465 ref_kT = 2.462
	Predicted entropy: -0.29456645250320435
	Predicted free_energy: -13143.2275390625
	Predicted pressure: 9.574121475219727
[Step Size] Found optimal step size 0.07363811880350113 with residual 0.003725767135620117

[DiffTRe] Epoch 131
	Epoch loss = 0.00144
	Gradient norm: 0.0053233010694384575
	Elapsed time = 1.759 min
[Statepoint 0]
	kT = 2.467 ref_kT = 2.462
	Predicted entropy: -0.006262412294745445
	Predicted free_energy: -12893.3671875
	Predicted pressure: 19.383010864257812
[Step Size] Found optimal step size 0.05704105645418167 with residual 0.005639076232910156

[DiffTRe] Epoch 132
	Epoch loss = 0.00143
	Gradient norm: 0.009120093658566475
	Elapsed time = 1.759 min
[Statepoint 0]
	kT = 2.454 ref_kT = 2.462
	Predicted entropy: -0.19926005601882935
	Predicted free_energy: -12954.3798828125
	Predicted pressure: 18.217395782470703
[Step Size] Found optimal step size 0.07022307813167572 with residual 0.0013570785522460938

[DiffTRe] Epoch 133
	Epoch loss = 0.00146
	Gradient norm: 0.00855520274490118
	Elapsed time = 1.763 min
[Statepoint 0]
	kT = 2.469 ref_kT = 2.462
	Predicted entropy: -0.001468187547288835
	Predicted free_energy: -12881.1875
	Predicted pressure: 26.91283416748047
[Step Size] Found optimal step size 0.05419105291366577 with residual 0.0004119873046875

[DiffTRe] Epoch 134
	Epoch loss = 0.00141
	Gradient norm: 0.012492621317505836
	Elapsed time = 1.756 min
[Statepoint 0]
	kT = 2.479 ref_kT = 2.462
	Predicted entropy: -0.37764599919319153
	Predicted free_energy: -13303.470703125
	Predicted pressure: 5.007392406463623
[Step Size] Found optimal step size 0.09091818332672119 with residual 0.0024378299713134766

[DiffTRe] Epoch 135
	Epoch loss = 0.00134
	Gradient norm: 0.00745299831032753
	Elapsed time = 1.760 min
[Statepoint 0]
	kT = 2.472 ref_kT = 2.462
	Predicted entropy: -0.11115588247776031
	Predicted free_energy: -13048.5654296875
	Predicted pressure: 21.14873504638672
[Step Size] Found optimal step size 0.053036145865917206 with residual 0.032204627990722656

[DiffTRe] Epoch 136
	Epoch loss = 0.00159
	Gradient norm: 0.036902133375406265
	Elapsed time = 1.757 min
[Statepoint 0]
	kT = 2.458 ref_kT = 2.462
	Predicted entropy: -0.4891677796840668
	Predicted free_energy: -13293.62890625
	Predicted pressure: 6.186668395996094
[Step Size] Found optimal step size 0.4132290482521057 with residual 0.003261566162109375

[DiffTRe] Epoch 137
	Epoch loss = 0.00126
	Gradient norm: 0.0007596607320010662
	Elapsed time = 1.758 min
[Statepoint 0]
	kT = 2.464 ref_kT = 2.462
	Predicted entropy: -0.15870434045791626
	Predicted free_energy: -13064.90625
	Predicted pressure: 17.229324340820312
[Step Size] Found optimal step size 0.06250511109828949 with residual 0.004647254943847656

[DiffTRe] Epoch 138
	Epoch loss = 0.00144
	Gradient norm: 0.018387041985988617
	Elapsed time = 1.759 min
[Statepoint 0]
	kT = 2.466 ref_kT = 2.462
	Predicted entropy: -0.459107369184494
	Predicted free_energy: -13798.0986328125
	Predicted pressure: -12.345701217651367
[Step Size] Found optimal step size 0.11299170553684235 with residual 0.0046024322509765625

[DiffTRe] Epoch 139
	Epoch loss = 0.00126
	Gradient norm: 0.0017167078331112862
	Elapsed time = 1.757 min
[Statepoint 0]
	kT = 2.461 ref_kT = 2.462
	Predicted entropy: -0.16851244866847992
	Predicted free_energy: -13251.9921875
	Predicted pressure: 11.545320510864258
[Step Size] Found optimal step size 0.05895347148180008 with residual 0.004323482513427734

[DiffTRe] Epoch 140
	Epoch loss = 0.00137
	Gradient norm: 0.014939694665372372
	Elapsed time = 1.758 min
[Statepoint 0]
	kT = 2.462 ref_kT = 2.462
	Predicted entropy: -0.25694671273231506
	Predicted free_energy: -12972.556640625
	Predicted pressure: 14.9073486328125
[Step Size] Found optimal step size 0.08265381306409836 with residual 0.06426024436950684

[DiffTRe] Epoch 141
	Epoch loss = 0.00124
	Gradient norm: 0.00816381350159645
	Elapsed time = 1.759 min
[Statepoint 0]
	kT = 2.458 ref_kT = 2.462
	Predicted entropy: -0.07127826660871506
	Predicted free_energy: -13025.951171875
	Predicted pressure: 13.60842227935791
[Step Size] Found optimal step size 0.06250511109828949 with residual 0.0030884742736816406

[DiffTRe] Epoch 142
	Epoch loss = 0.00139
	Gradient norm: 0.028743578121066093
	Elapsed time = 1.758 min
[Statepoint 0]
	kT = 2.460 ref_kT = 2.462
	Predicted entropy: -0.5946515798568726
	Predicted free_energy: -13475.12890625
	Predicted pressure: 0.36087751388549805
[Step Size] Found optimal step size 0.11653093248605728 with residual 0.0006151199340820312

[DiffTRe] Epoch 143
	Epoch loss = 0.00126
	Gradient norm: 0.00586539413779974
	Elapsed time = 1.761 min
[Statepoint 0]
	kT = 2.463 ref_kT = 2.462
	Predicted entropy: -0.27059206366539
	Predicted free_energy: -13166.53125
	Predicted pressure: 17.46383285522461
[Step Size] Found optimal step size 0.07111099362373352 with residual 0.0012977123260498047

[DiffTRe] Epoch 144
	Epoch loss = 0.00133
	Gradient norm: 0.014640672132372856
	Elapsed time = 1.757 min
[Statepoint 0]
	kT = 2.460 ref_kT = 2.462
	Predicted entropy: -0.4450242817401886
	Predicted free_energy: -13171.5458984375
	Predicted pressure: 10.75906753540039
[Step Size] Found optimal step size 0.09806491434574127 with residual 0.002453327178955078

[DiffTRe] Epoch 145
	Epoch loss = 0.00124
	Gradient norm: 0.005843006540089846
	Elapsed time = 1.762 min
[Statepoint 0]
	kT = 2.463 ref_kT = 2.462
	Predicted entropy: -0.2267427295446396
	Predicted free_energy: -13128.65625
	Predicted pressure: 16.284404754638672
[Step Size] Found optimal step size 0.068087138235569 with residual 0.002670764923095703

[DiffTRe] Epoch 146
	Epoch loss = 0.00122
	Gradient norm: 0.013151883147656918
	Elapsed time = 1.765 min
[Statepoint 0]
	kT = 2.466 ref_kT = 2.462
	Predicted entropy: -0.530866265296936
	Predicted free_energy: -13320.0146484375
	Predicted pressure: 4.622373104095459
[Step Size] Found optimal step size 0.07363811880350113 with residual 0.004256248474121094

[DiffTRe] Epoch 147
	Epoch loss = 0.00123
	Gradient norm: 0.00858942698687315
	Elapsed time = 1.765 min
[Statepoint 0]
	kT = 2.456 ref_kT = 2.462
	Predicted entropy: -0.3198707401752472
	Predicted free_energy: -13093.0634765625
	Predicted pressure: 20.895618438720703
[Step Size] Found optimal step size 0.07917047291994095 with residual 0.01867842674255371

[DiffTRe] Epoch 148
	Epoch loss = 0.00123
	Gradient norm: 0.0054591624066233635
	Elapsed time = 1.755 min
[Statepoint 0]
	kT = 2.457 ref_kT = 2.462
	Predicted entropy: -0.5573383569717407
	Predicted free_energy: -13427.6416015625
	Predicted pressure: 1.977128028869629
[Step Size] Found optimal step size 0.11475510895252228 with residual 0.005603790283203125

[DiffTRe] Epoch 149
	Epoch loss = 0.00118
	Gradient norm: 0.0037443849723786116
	Elapsed time = 1.764 min
[Statepoint 0]
	kT = 2.462 ref_kT = 2.462
	Predicted entropy: -0.31681370735168457
	Predicted free_energy: -13261.244140625
	Predicted pressure: 7.99160623550415
[Step Size] Found optimal step size 0.057109348475933075 with residual 0.032433271408081055

[DiffTRe] Epoch 150
	Epoch loss = 0.00134
	Gradient norm: 0.02606062963604927
	Elapsed time = 1.760 min
[Statepoint 0]
	kT = 2.458 ref_kT = 2.462
	Predicted entropy: -0.47595158219337463
	Predicted free_energy: -13126.9365234375
	Predicted pressure: 13.706354141235352
[Step Size] Found optimal step size 0.079487144947052 with residual 0.002795696258544922

[DiffTRe] Epoch 151
	Epoch loss = 0.00114
	Gradient norm: 0.006159735843539238
	Elapsed time = 1.766 min
[Statepoint 0]
	kT = 2.453 ref_kT = 2.462
	Predicted entropy: -0.22984445095062256
	Predicted free_energy: -13010.2314453125
	Predicted pressure: 19.291500091552734
[Step Size] Found optimal step size 0.10177797079086304 with residual 3.4809112548828125e-05

[DiffTRe] Epoch 152
	Epoch loss = 0.00117
	Gradient norm: 0.0039109475910663605
	Elapsed time = 1.760 min
[Statepoint 0]
	kT = 2.465 ref_kT = 2.462
	Predicted entropy: -0.5120499730110168
	Predicted free_energy: -13381.166015625
	Predicted pressure: 8.104080200195312
[Step Size] Found optimal step size 0.06679562479257584 with residual 0.0009014606475830078

[DiffTRe] Epoch 153
	Epoch loss = 0.00124
	Gradient norm: 0.010437505319714546
	Elapsed time = 1.762 min
[Statepoint 0]
	kT = 2.457 ref_kT = 2.462
	Predicted entropy: -0.25648385286331177
	Predicted free_energy: -13146.48828125
	Predicted pressure: 13.300271987915039
[Step Size] Found optimal step size 0.11168158054351807 with residual 0.001280069351196289

[DiffTRe] Epoch 154
	Epoch loss = 0.00116
	Gradient norm: 0.005685101728886366
	Elapsed time = 1.759 min
[Statepoint 0]
	kT = 2.463 ref_kT = 2.462
	Predicted entropy: -0.42999058961868286
	Predicted free_energy: -13250.2314453125
	Predicted pressure: 14.120537757873535
[Step Size] Found optimal step size 0.07881034165620804 with residual 8.487701416015625e-05

[DiffTRe] Epoch 155
	Epoch loss = 0.00133
	Gradient norm: 0.02870955690741539
	Elapsed time = 1.762 min
[Statepoint 0]
	kT = 2.466 ref_kT = 2.462
	Predicted entropy: -0.17102426290512085
	Predicted free_energy: -13138.740234375
	Predicted pressure: 16.278947830200195
[Step Size] Found optimal step size 0.1120975986123085 with residual 0.00024199485778808594

[DiffTRe] Epoch 156
	Epoch loss = 0.00113
	Gradient norm: 0.006786848418414593
	Elapsed time = 1.760 min
[Statepoint 0]
	kT = 2.464 ref_kT = 2.462
	Predicted entropy: -0.6061277389526367
	Predicted free_energy: -13401.466796875
	Predicted pressure: 8.351635932922363
[Step Size] Found optimal step size 0.08265381306409836 with residual 0.13060760498046875

[DiffTRe] Epoch 157
	Epoch loss = 0.00115
	Gradient norm: 0.009203311987221241
	Elapsed time = 1.762 min
[Statepoint 0]
	kT = 2.470 ref_kT = 2.462
	Predicted entropy: -0.3211779296398163
	Predicted free_energy: -13181.7060546875
	Predicted pressure: 10.706298828125
[Step Size] Found optimal step size 0.08265381306409836 with residual 0.14641523361206055

[DiffTRe] Epoch 158
	Epoch loss = 0.00114
	Gradient norm: 0.009481983259320259
	Elapsed time = 1.760 min
[Statepoint 0]
	kT = 2.461 ref_kT = 2.462
	Predicted entropy: -0.5568732023239136
	Predicted free_energy: -13423.8984375
	Predicted pressure: 5.533432483673096
[Step Size] Found optimal step size 0.1002877801656723 with residual 0.0052204132080078125

[DiffTRe] Epoch 159
	Epoch loss = 0.00112
	Gradient norm: 0.012723380699753761
	Elapsed time = 1.762 min
[Statepoint 0]
	kT = 2.463 ref_kT = 2.462
	Predicted entropy: -0.32786446809768677
	Predicted free_energy: -13308.03125
	Predicted pressure: 3.633436679840088
[Step Size] Found optimal step size 0.10894954949617386 with residual 0.0030879974365234375

[DiffTRe] Epoch 160
	Epoch loss = 0.00125
	Gradient norm: 0.02252899669110775
	Elapsed time = 1.758 min
[Statepoint 0]
	kT = 2.453 ref_kT = 2.462
	Predicted entropy: -0.5758998990058899
	Predicted free_energy: -13330.2880859375
	Predicted pressure: 3.550553798675537
[Step Size] Found optimal step size 0.07910218089818954 with residual 0.005681514739990234

[DiffTRe] Epoch 161
	Epoch loss = 0.00123
	Gradient norm: 0.012591018341481686
	Elapsed time = 1.761 min
[Statepoint 0]
	kT = 2.462 ref_kT = 2.462
	Predicted entropy: -0.2725426256656647
	Predicted free_energy: -13084.2109375
	Predicted pressure: 20.908308029174805
[Step Size] Found optimal step size 0.5772188901901245 with residual 0.00301361083984375

[DiffTRe] Epoch 162
	Epoch loss = 0.00103
	Gradient norm: 0.00028245040448382497
	Elapsed time = 1.761 min
[Statepoint 0]
	kT = 2.458 ref_kT = 2.462
	Predicted entropy: -0.5157678127288818
	Predicted free_energy: -13379.3857421875
	Predicted pressure: 4.131623268127441
[Step Size] Found optimal step size 0.12233028560876846 with residual 0.003349781036376953

[DiffTRe] Epoch 163
	Epoch loss = 0.00116
	Gradient norm: 0.010493122972548008
	Elapsed time = 1.761 min
[Statepoint 0]
	kT = 2.473 ref_kT = 2.462
	Predicted entropy: -0.34306588768959045
	Predicted free_energy: -12841.966796875
	Predicted pressure: 26.502798080444336
[Step Size] Found optimal step size 0.07985348999500275 with residual 0.0029306411743164062

[DiffTRe] Epoch 164
	Epoch loss = 0.00108
	Gradient norm: 0.009944038465619087
	Elapsed time = 1.760 min
[Statepoint 0]
	kT = 2.459 ref_kT = 2.462
	Predicted entropy: -0.6691140532493591
	Predicted free_energy: -13550.0771484375
	Predicted pressure: 3.7016406059265137
[Step Size] Found optimal step size 0.1015854924917221 with residual 0.00016117095947265625

[DiffTRe] Epoch 165
	Epoch loss = 0.00109
	Gradient norm: 0.012493346817791462
	Elapsed time = 1.761 min
[Statepoint 0]
	kT = 2.462 ref_kT = 2.462
	Predicted entropy: -0.38962993025779724
	Predicted free_energy: -13323.244140625
	Predicted pressure: 10.351893424987793
[Step Size] Found optimal step size 0.05993451923131943 with residual 0.011365652084350586

[DiffTRe] Epoch 166
	Epoch loss = 0.00113
	Gradient norm: 0.02014072611927986
	Elapsed time = 1.762 min
[Statepoint 0]
	kT = 2.469 ref_kT = 2.462
	Predicted entropy: -0.6147303581237793
	Predicted free_energy: -13383.4794921875
	Predicted pressure: 7.2228617668151855
[Step Size] Found optimal step size 0.24958065152168274 with residual 0.0007834434509277344

[DiffTRe] Epoch 167
	Epoch loss = 0.00100
	Gradient norm: 0.0008527624886482954
	Elapsed time = 1.762 min
[Statepoint 0]
	kT = 2.464 ref_kT = 2.462
	Predicted entropy: -0.403968870639801
	Predicted free_energy: -13218.673828125
	Predicted pressure: 14.466536521911621
[Step Size] Found optimal step size 0.12964466214179993 with residual 0.013013839721679688

[DiffTRe] Epoch 168
	Epoch loss = 0.00119
	Gradient norm: 0.008082482032477856
	Elapsed time = 1.761 min
[Statepoint 0]
	kT = 2.463 ref_kT = 2.462
	Predicted entropy: -0.6800970435142517
	Predicted free_energy: -13727.6279296875
	Predicted pressure: -8.114789009094238
[Step Size] Found optimal step size 0.09411589056253433 with residual 0.0005931854248046875

[DiffTRe] Epoch 169
	Epoch loss = 0.00102
	Gradient norm: 0.006687653250992298
	Elapsed time = 1.763 min
[Statepoint 0]
	kT = 2.461 ref_kT = 2.462
	Predicted entropy: -0.576842725276947
	Predicted free_energy: -13265.2724609375
	Predicted pressure: 12.177997589111328
[Step Size] Found optimal step size 0.092420794069767 with residual 9.298324584960938e-06

[DiffTRe] Epoch 170
	Epoch loss = 0.00116
	Gradient norm: 0.02016681432723999
	Elapsed time = 1.764 min
[Statepoint 0]
	kT = 2.468 ref_kT = 2.462
	Predicted entropy: -0.24803099036216736
	Predicted free_energy: -13046.6455078125
	Predicted pressure: 24.972963333129883
[Step Size] Found optimal step size 0.19495871663093567 with residual 0.0004782676696777344

[DiffTRe] Epoch 171
	Epoch loss = 0.00100
	Gradient norm: 0.0036156177520751953
	Elapsed time = 1.764 min
[Statepoint 0]
	kT = 2.463 ref_kT = 2.462
	Predicted entropy: -0.6891888976097107
	Predicted free_energy: -13439.6044921875
	Predicted pressure: 6.421101093292236
[Step Size] Found optimal step size 0.08265380561351776 with residual 0.10960102081298828

[DiffTRe] Epoch 172
	Epoch loss = 0.00115
	Gradient norm: 0.01593577116727829
	Elapsed time = 1.762 min
[Statepoint 0]
	kT = 2.465 ref_kT = 2.462
	Predicted entropy: -0.4181995391845703
	Predicted free_energy: -13322.80078125
	Predicted pressure: 9.746463775634766
[Step Size] Found optimal step size 0.36801400780677795 with residual 0.0014023780822753906

[DiffTRe] Epoch 173
	Epoch loss = 0.00096
	Gradient norm: 0.0009385020239278674
	Elapsed time = 1.763 min
[Statepoint 0]
	kT = 2.451 ref_kT = 2.462
	Predicted entropy: -0.5771275162696838
	Predicted free_energy: -13404.224609375
	Predicted pressure: 8.738100051879883
[Step Size] Found optimal step size 0.07677994668483734 with residual 0.003634929656982422

[DiffTRe] Epoch 174
	Epoch loss = 0.00111
	Gradient norm: 0.024722937494516373
	Elapsed time = 1.763 min
[Statepoint 0]
	kT = 2.449 ref_kT = 2.462
	Predicted entropy: -0.258713960647583
	Predicted free_energy: -13114.298828125
	Predicted pressure: 19.93006706237793
[Step Size] Found optimal step size 0.3314048647880554 with residual 0.0031342506408691406

[DiffTRe] Epoch 175
	Epoch loss = 0.00095
	Gradient norm: 0.0016496634343639016
	Elapsed time = 1.762 min
[Statepoint 0]
	kT = 2.468 ref_kT = 2.462
	Predicted entropy: -0.6386738419532776
	Predicted free_energy: -13388.154296875
	Predicted pressure: 11.610526084899902
[Step Size] Found optimal step size 0.10842178016901016 with residual 0.0028400421142578125

[DiffTRe] Epoch 176
	Epoch loss = 0.00131
	Gradient norm: 0.027991842478513718
	Elapsed time = 1.763 min
[Statepoint 0]
	kT = 2.464 ref_kT = 2.462
	Predicted entropy: -0.54542076587677
	Predicted free_energy: -13682.1533203125
	Predicted pressure: -2.7731635570526123
[Step Size] Found optimal step size 0.2355789840221405 with residual 0.0014226436614990234

[DiffTRe] Epoch 177
	Epoch loss = 0.00096
	Gradient norm: 0.002313189208507538
	Elapsed time = 1.761 min
[Statepoint 0]
	kT = 2.464 ref_kT = 2.462
	Predicted entropy: -0.6714493036270142
	Predicted free_energy: -13604.4365234375
	Predicted pressure: 4.258894443511963
[Step Size] Found optimal step size 0.07855577021837234 with residual 0.0058269500732421875

[DiffTRe] Epoch 178
	Epoch loss = 0.00110
	Gradient norm: 0.02290099859237671
	Elapsed time = 1.761 min
[Statepoint 0]
	kT = 2.462 ref_kT = 2.462
	Predicted entropy: -0.3500922918319702
	Predicted free_energy: -13040.3935546875
	Predicted pressure: 27.059263229370117
[Step Size] Found optimal step size 0.3021721839904785 with residual 0.002565622329711914

[DiffTRe] Epoch 179
	Epoch loss = 0.00090
	Gradient norm: 0.0017167342593893409
	Elapsed time = 1.761 min
[Statepoint 0]
	kT = 2.457 ref_kT = 2.462
	Predicted entropy: -0.6751275658607483
	Predicted free_energy: -13329.98828125
	Predicted pressure: 12.870975494384766
[Step Size] Found optimal step size 0.11728223413228989 with residual 0.0002694129943847656

[DiffTRe] Epoch 180
	Epoch loss = 0.00111
	Gradient norm: 0.01710714027285576
	Elapsed time = 1.757 min
[Statepoint 0]
	kT = 2.461 ref_kT = 2.462
	Predicted entropy: -0.65134596824646
	Predicted free_energy: -13639.5361328125
	Predicted pressure: 1.6216219663619995
[Step Size] Found optimal step size 0.14050446450710297 with residual 0.003129243850708008

[DiffTRe] Epoch 181
	Epoch loss = 0.00098
	Gradient norm: 0.005657400004565716
	Elapsed time = 1.758 min
[Statepoint 0]
	kT = 2.457 ref_kT = 2.462
	Predicted entropy: -0.8598465323448181
	Predicted free_energy: -13600.0732421875
	Predicted pressure: 0.9818373918533325
[Step Size] Found optimal step size 0.10921654850244522 with residual 7.295608520507812e-05

[DiffTRe] Epoch 182
	Epoch loss = 0.00100
	Gradient norm: 0.011230857111513615
	Elapsed time = 1.761 min
[Statepoint 0]
	kT = 2.466 ref_kT = 2.462
	Predicted entropy: -0.5423478484153748
	Predicted free_energy: -13239.1845703125
	Predicted pressure: 16.079801559448242
[Step Size] Found optimal step size 0.16229237616062164 with residual 0.0005078315734863281

[DiffTRe] Epoch 183
	Epoch loss = 0.00097
	Gradient norm: 0.009291095659136772
	Elapsed time = 1.762 min
[Statepoint 0]
	kT = 2.466 ref_kT = 2.462
	Predicted entropy: -0.8656542897224426
	Predicted free_energy: -13399.0615234375
	Predicted pressure: 13.221773147583008
[Step Size] Found optimal step size 0.14174629747867584 with residual 0.0014235973358154297

[DiffTRe] Epoch 184
	Epoch loss = 0.00105
	Gradient norm: 0.014923816546797752
	Elapsed time = 1.757 min
[Statepoint 0]
	kT = 2.464 ref_kT = 2.462
	Predicted entropy: -0.6930996179580688
	Predicted free_energy: -13455.607421875
	Predicted pressure: 14.337471008300781
[Step Size] Found optimal step size 0.09984694421291351 with residual 0.0007338523864746094

[DiffTRe] Epoch 185
	Epoch loss = 0.00110
	Gradient norm: 0.023956522345542908
	Elapsed time = 1.762 min
[Statepoint 0]
	kT = 2.453 ref_kT = 2.462
	Predicted entropy: -1.0580991506576538
	Predicted free_energy: -13706.66796875
	Predicted pressure: -3.878673791885376
[Step Size] Found optimal step size 0.4936188757419586 with residual 0.0008592605590820312

[DiffTRe] Epoch 186
	Epoch loss = 0.00087
	Gradient norm: 0.001246549654752016
	Elapsed time = 1.758 min
[Statepoint 0]
	kT = 2.468 ref_kT = 2.462
	Predicted entropy: -0.7549477219581604
	Predicted free_energy: -13412.1552734375
	Predicted pressure: 12.052648544311523
[Step Size] Found optimal step size 0.10907994955778122 with residual 0.0011043548583984375

[DiffTRe] Epoch 187
	Epoch loss = 0.00132
	Gradient norm: 0.030545085668563843
	Elapsed time = 1.762 min
[Statepoint 0]
	kT = 2.459 ref_kT = 2.462
	Predicted entropy: -0.6535022854804993
	Predicted free_energy: -13148.1162109375
	Predicted pressure: 17.723224639892578
[Step Size] Found optimal step size 0.16884922981262207 with residual 0.0026535987854003906

[DiffTRe] Epoch 188
	Epoch loss = 0.00095
	Gradient norm: 0.0062306285835802555
	Elapsed time = 1.761 min
[Statepoint 0]
	kT = 2.452 ref_kT = 2.462
	Predicted entropy: -0.49812448024749756
	Predicted free_energy: -13209.78515625
	Predicted pressure: 14.843850135803223
[Step Size] Found optimal step size 0.16037994623184204 with residual 0.0001697540283203125

[DiffTRe] Epoch 189
	Epoch loss = 0.00094
	Gradient norm: 0.004702946171164513
	Elapsed time = 1.763 min
[Statepoint 0]
	kT = 2.461 ref_kT = 2.462
	Predicted entropy: -0.8218207359313965
	Predicted free_energy: -13723.3759765625
	Predicted pressure: -2.1909608840942383
[Step Size] Found optimal step size 0.1480175405740738 with residual 0.0030667781829833984

[DiffTRe] Epoch 190
	Epoch loss = 0.00094
	Gradient norm: 0.005732365883886814
	Elapsed time = 1.761 min
[Statepoint 0]
	kT = 2.455 ref_kT = 2.462
	Predicted entropy: -0.5883901119232178
	Predicted free_energy: -13348.2236328125
	Predicted pressure: 12.8528470993042
[Step Size] Found optimal step size 0.11646263301372528 with residual 0.0002243518829345703

[DiffTRe] Epoch 191
	Epoch loss = 0.00095
	Gradient norm: 0.013254457153379917
	Elapsed time = 1.761 min
[Statepoint 0]
	kT = 2.453 ref_kT = 2.462
	Predicted entropy: -0.8717754483222961
	Predicted free_energy: -13564.0712890625
	Predicted pressure: 6.592948913574219
[Step Size] Found optimal step size 0.15790870785713196 with residual 0.00441741943359375

[DiffTRe] Epoch 192
	Epoch loss = 0.00089
	Gradient norm: 0.0069028413854539394
	Elapsed time = 1.763 min
[Statepoint 0]
	kT = 2.457 ref_kT = 2.462
	Predicted entropy: -0.5631647109985352
	Predicted free_energy: -13421.6708984375
	Predicted pressure: 6.649895668029785
[Step Size] Found optimal step size 0.15559890866279602 with residual 0.000102996826171875

[DiffTRe] Epoch 193
	Epoch loss = 0.00095
	Gradient norm: 0.007805053144693375
	Elapsed time = 1.759 min
[Statepoint 0]
	kT = 2.463 ref_kT = 2.462
	Predicted entropy: -0.8289934992790222
	Predicted free_energy: -13465.5537109375
	Predicted pressure: 10.855778694152832
[Step Size] Found optimal step size 0.1944868117570877 with residual 0.0016460418701171875

[DiffTRe] Epoch 194
	Epoch loss = 0.00093
	Gradient norm: 0.005970433820039034
	Elapsed time = 1.763 min
[Statepoint 0]
	kT = 2.461 ref_kT = 2.462
	Predicted entropy: -0.6099807024002075
	Predicted free_energy: -13404.353515625
	Predicted pressure: 14.638969421386719
[Step Size] Found optimal step size 0.1114145889878273 with residual 0.0006690025329589844

[DiffTRe] Epoch 195
	Epoch loss = 0.00102
	Gradient norm: 0.016431672498583794
	Elapsed time = 1.763 min
[Statepoint 0]
	kT = 2.467 ref_kT = 2.462
	Predicted entropy: -0.7876816391944885
	Predicted free_energy: -13529.6025390625
	Predicted pressure: 7.173644065856934
[Step Size] Found optimal step size 0.31255385279655457 with residual 0.008148908615112305

[DiffTRe] Epoch 196
	Epoch loss = 0.00089
	Gradient norm: 0.002757064066827297
	Elapsed time = 1.761 min
[Statepoint 0]
	kT = 2.467 ref_kT = 2.462
	Predicted entropy: -0.6112387776374817
	Predicted free_energy: -13405.353515625
	Predicted pressure: 5.377559661865234
[Step Size] Found optimal step size 0.1268070638179779 with residual 0.010287284851074219

[DiffTRe] Epoch 197
	Epoch loss = 0.00114
	Gradient norm: 0.030014313757419586
	Elapsed time = 1.762 min
[Statepoint 0]
	kT = 2.463 ref_kT = 2.462
	Predicted entropy: -0.981273889541626
	Predicted free_energy: -13804.9462890625
	Predicted pressure: -5.090088367462158
[Step Size] Found optimal step size 0.5756356120109558 with residual 0.0008263587951660156

[DiffTRe] Epoch 198
	Epoch loss = 0.00086
	Gradient norm: 0.0008878438966348767
	Elapsed time = 1.763 min
[Statepoint 0]
	kT = 2.459 ref_kT = 2.462
	Predicted entropy: -0.5705790519714355
	Predicted free_energy: -13470.77734375
	Predicted pressure: 16.360149383544922
[Step Size] Found optimal step size 0.1508861631155014 with residual 0.0017282962799072266

[DiffTRe] Epoch 199
	Epoch loss = 0.00107
	Gradient norm: 0.010327127762138844
	Elapsed time = 1.764 min
[Statepoint 0]
	kT = 2.454 ref_kT = 2.462
	Predicted entropy: -0.4277544617652893
	Predicted free_energy: -13303.0244140625
	Predicted pressure: 12.82096004486084
[Step Size] Found optimal step size 0.19159334897994995 with residual 0.0010330677032470703

[DiffTRe] Epoch 200
	Epoch loss = 0.00088
	Gradient norm: 0.002098355209454894
	Elapsed time = 1.762 min
[Statepoint 0]
	kT = 2.456 ref_kT = 2.462
	Predicted entropy: -0.4322604537010193
	Predicted free_energy: -13557.6669921875
	Predicted pressure: 3.4293465614318848
[Step Size] Found optimal step size 0.1262296438217163 with residual 0.001789093017578125

[DiffTRe] Epoch 201
	Epoch loss = 0.00095
	Gradient norm: 0.013617577031254768
	Elapsed time = 1.759 min
[Statepoint 0]
	kT = 2.471 ref_kT = 2.462
	Predicted entropy: -0.8028342127799988
	Predicted free_energy: -13861.3701171875
	Predicted pressure: -4.612878322601318
[Step Size] Found optimal step size 0.2587329149246216 with residual 0.0008530616760253906

[DiffTRe] Epoch 202
	Epoch loss = 0.00087
	Gradient norm: 0.0038499110378324986
	Elapsed time = 1.762 min
[Statepoint 0]
	kT = 2.459 ref_kT = 2.462
	Predicted entropy: -0.5070548057556152
	Predicted free_energy: -13559.6767578125
	Predicted pressure: 6.921104907989502
[Step Size] Found optimal step size 0.15548093616962433 with residual 0.0009937286376953125

[DiffTRe] Epoch 203
	Epoch loss = 0.00106
	Gradient norm: 0.020939012989401817
	Elapsed time = 1.762 min
[Statepoint 0]
	kT = 2.467 ref_kT = 2.462
	Predicted entropy: -0.6505631804466248
	Predicted free_energy: -13439.1181640625
	Predicted pressure: 9.58065414428711
[Step Size] Found optimal step size 0.25294601917266846 with residual 0.002223968505859375

[DiffTRe] Epoch 204
	Epoch loss = 0.00086
	Gradient norm: 0.004642319865524769
	Elapsed time = 1.761 min
[Statepoint 0]
	kT = 2.458 ref_kT = 2.462
	Predicted entropy: -0.45198118686676025
	Predicted free_energy: -13410.982421875
	Predicted pressure: 11.645933151245117
[Step Size] Found optimal step size 0.1443479359149933 with residual 0.003912687301635742

[DiffTRe] Epoch 205
	Epoch loss = 0.00097
	Gradient norm: 0.019416332244873047
	Elapsed time = 1.760 min
[Statepoint 0]
	kT = 2.463 ref_kT = 2.462
	Predicted entropy: -0.8743810653686523
	Predicted free_energy: -13914.232421875
	Predicted pressure: -4.13382625579834
[Step Size] Found optimal step size 0.3245747685432434 with residual 0.000156402587890625

[DiffTRe] Epoch 206
	Epoch loss = 0.00086
	Gradient norm: 0.0017309188842773438
	Elapsed time = 1.760 min
[Statepoint 0]
	kT = 2.459 ref_kT = 2.462
	Predicted entropy: -0.47697171568870544
	Predicted free_energy: -13560.8076171875
	Predicted pressure: 9.242925643920898
[Step Size] Found optimal step size 0.24719011783599854 with residual 0.00793766975402832

[DiffTRe] Epoch 207
	Epoch loss = 0.00098
	Gradient norm: 0.01009127963334322
	Elapsed time = 1.763 min
[Statepoint 0]
	kT = 2.467 ref_kT = 2.462
	Predicted entropy: -0.42149582505226135
	Predicted free_energy: -13277.650390625
	Predicted pressure: 18.442445755004883
[Step Size] Found optimal step size 0.18455839157104492 with residual 0.001710653305053711

[DiffTRe] Epoch 208
	Epoch loss = 0.00090
	Gradient norm: 0.008919553831219673
	Elapsed time = 1.759 min
[Statepoint 0]
	kT = 2.464 ref_kT = 2.462
	Predicted entropy: -0.47109103202819824
	Predicted free_energy: -13647.2958984375
	Predicted pressure: 4.065258502960205
[Step Size] Found optimal step size 0.14637210965156555 with residual 0.007252216339111328

[DiffTRe] Epoch 209
	Epoch loss = 0.00094
	Gradient norm: 0.01642722077667713
	Elapsed time = 1.759 min
[Statepoint 0]
	kT = 2.459 ref_kT = 2.462
	Predicted entropy: -0.7698646783828735
	Predicted free_energy: -13722.6240234375
	Predicted pressure: -0.4107831120491028
[Step Size] Found optimal step size 0.284978985786438 with residual 0.0006921291351318359

[DiffTRe] Epoch 210
	Epoch loss = 0.00082
	Gradient norm: 0.00481268297880888
	Elapsed time = 1.761 min
[Statepoint 0]
	kT = 2.471 ref_kT = 2.462
	Predicted entropy: -0.4243020713329315
	Predicted free_energy: -13408.5283203125
	Predicted pressure: 12.424609184265137
[Step Size] Found optimal step size 0.16136100888252258 with residual 0.0023436546325683594

[DiffTRe] Epoch 211
	Epoch loss = 0.00100
	Gradient norm: 0.021675443276762962
	Elapsed time = 1.760 min
[Statepoint 0]
	kT = 2.463 ref_kT = 2.462
	Predicted entropy: -0.7666885852813721
	Predicted free_energy: -13711.58203125
	Predicted pressure: -0.47721877694129944
[Step Size] Found optimal step size 0.27860841155052185 with residual 0.0011363029479980469

[DiffTRe] Epoch 212
	Epoch loss = 0.00084
	Gradient norm: 0.0049429829232394695
	Elapsed time = 1.761 min
[Statepoint 0]
	kT = 2.463 ref_kT = 2.462
	Predicted entropy: -0.4619995951652527
	Predicted free_energy: -13517.6962890625
	Predicted pressure: 9.243338584899902
[Step Size] Found optimal step size 0.233343705534935 with residual 0.0014677047729492188

[DiffTRe] Epoch 213
	Epoch loss = 0.00096
	Gradient norm: 0.012968898750841618
	Elapsed time = 1.761 min
[Statepoint 0]
	kT = 2.455 ref_kT = 2.462
	Predicted entropy: -0.6351474523544312
	Predicted free_energy: -13573.8798828125
	Predicted pressure: 7.621545314788818
[Step Size] Found optimal step size 0.2092646360397339 with residual 0.0023915767669677734

[DiffTRe] Epoch 214
	Epoch loss = 0.00096
	Gradient norm: 0.016625214368104935
	Elapsed time = 1.761 min
[Statepoint 0]
	kT = 2.468 ref_kT = 2.462
	Predicted entropy: -0.41180428862571716
	Predicted free_energy: -13482.8525390625
	Predicted pressure: 12.209084510803223
[Step Size] Found optimal step size 0.25620579719543457 with residual 0.003364086151123047

[DiffTRe] Epoch 215
	Epoch loss = 0.00087
	Gradient norm: 0.007520496845245361
	Elapsed time = 1.764 min
[Statepoint 0]
	kT = 2.471 ref_kT = 2.462
	Predicted entropy: -0.6729598641395569
	Predicted free_energy: -13678.7021484375
	Predicted pressure: 4.3819475173950195
[Step Size] Found optimal step size 0.13920675218105316 with residual 0.0030889511108398438

[DiffTRe] Epoch 216
	Epoch loss = 0.00092
	Gradient norm: 0.014978437684476376
	Elapsed time = 1.761 min
[Statepoint 0]
	kT = 2.469 ref_kT = 2.462
	Predicted entropy: -0.2689206004142761
	Predicted free_energy: -13271.193359375
	Predicted pressure: 16.952167510986328
[Step Size] Found optimal step size 0.4072185754776001 with residual 0.0006008148193359375

[DiffTRe] Epoch 217
	Epoch loss = 0.00080
	Gradient norm: 0.002098860451951623
	Elapsed time = 1.763 min
[Statepoint 0]
	kT = 2.448 ref_kT = 2.462
	Predicted entropy: -0.5741846561431885
	Predicted free_energy: -13559.603515625
	Predicted pressure: 6.647152423858643
[Step Size] Found optimal step size 0.20974892377853394 with residual 0.0001347064971923828

[DiffTRe] Epoch 218
	Epoch loss = 0.00098
	Gradient norm: 0.015828287228941917
	Elapsed time = 1.758 min
[Statepoint 0]
	kT = 2.464 ref_kT = 2.462
	Predicted entropy: -0.3244549334049225
	Predicted free_energy: -13366.5380859375
	Predicted pressure: 18.91326332092285
[Step Size] Found optimal step size 0.3884793519973755 with residual 0.0003445148468017578

[DiffTRe] Epoch 219
	Epoch loss = 0.00084
	Gradient norm: 0.004547832999378443
	Elapsed time = 1.762 min
[Statepoint 0]
	kT = 2.466 ref_kT = 2.462
	Predicted entropy: -0.5439746975898743
	Predicted free_energy: -13591.9765625
	Predicted pressure: 9.286730766296387
[Step Size] Found optimal step size 0.21696394681930542 with residual 0.003752470016479492

[DiffTRe] Epoch 220
	Epoch loss = 0.00103
	Gradient norm: 0.016765404492616653
	Elapsed time = 1.760 min
[Statepoint 0]
	kT = 2.464 ref_kT = 2.462
	Predicted entropy: -0.500851571559906
	Predicted free_energy: -13762.619140625
	Predicted pressure: 6.635993957519531
[Step Size] Found optimal step size 0.27490153908729553 with residual 0.001787424087524414

[DiffTRe] Epoch 221
	Epoch loss = 0.00084
	Gradient norm: 0.0068414825946092606
	Elapsed time = 1.765 min
[Statepoint 0]
	kT = 2.456 ref_kT = 2.462
	Predicted entropy: -0.6419028043746948
	Predicted free_energy: -13684.021484375
	Predicted pressure: 5.81642484664917
[Step Size] Found optimal step size 0.2506299614906311 with residual 0.0005519390106201172

[DiffTRe] Epoch 222
	Epoch loss = 0.00094
	Gradient norm: 0.01829497516155243
	Elapsed time = 1.760 min
[Statepoint 0]
	kT = 2.465 ref_kT = 2.462
	Predicted entropy: -0.2654847800731659
	Predicted free_energy: -13232.65234375
	Predicted pressure: 19.518699645996094
[Step Size] Found optimal step size 0.2644701600074768 with residual 0.055776119232177734

[DiffTRe] Epoch 223
	Epoch loss = 0.00088
	Gradient norm: 0.011690497398376465
	Elapsed time = 1.761 min
[Statepoint 0]
	kT = 2.462 ref_kT = 2.462
	Predicted entropy: -0.6665109395980835
	Predicted free_energy: -13613.2744140625
	Predicted pressure: 3.593543767929077
[Step Size] Found optimal step size 0.20009368658065796 with residual 0.0185699462890625

[DiffTRe] Epoch 224
	Epoch loss = 0.00088
	Gradient norm: 0.00797909777611494
	Elapsed time = 1.763 min
[Statepoint 0]
	kT = 2.466 ref_kT = 2.462
	Predicted entropy: -0.4183037281036377
	Predicted free_energy: -13493.7548828125
	Predicted pressure: 14.050877571105957
[Step Size] Found optimal step size 0.49676692485809326 with residual 0.0007119178771972656

[DiffTRe] Epoch 225
	Epoch loss = 0.00078
	Gradient norm: 0.0016687816241756082
	Elapsed time = 1.760 min
[Statepoint 0]
	kT = 2.469 ref_kT = 2.462
	Predicted entropy: -0.5660961866378784
	Predicted free_energy: -13565.2158203125
	Predicted pressure: 12.961318969726562
[Step Size] Found optimal step size 0.23182246088981628 with residual 0.00521540641784668

[DiffTRe] Epoch 226
	Epoch loss = 0.00103
	Gradient norm: 0.0188386682420969
	Elapsed time = 1.761 min
[Statepoint 0]
	kT = 2.466 ref_kT = 2.462
	Predicted entropy: -0.3228333592414856
	Predicted free_energy: -13437.9755859375
	Predicted pressure: 19.656679153442383
[Step Size] Found optimal step size 0.2550446689128876 with residual 4.100799560546875e-05

[DiffTRe] Epoch 227
	Epoch loss = 0.00084
	Gradient norm: 0.009612246416509151
	Elapsed time = 1.760 min
[Statepoint 0]
	kT = 2.459 ref_kT = 2.462
	Predicted entropy: -0.5574268102645874
	Predicted free_energy: -13756.2568359375
	Predicted pressure: -0.6820961236953735
[Step Size] Found optimal step size 0.2997816801071167 with residual 0.004805088043212891

[DiffTRe] Epoch 228
	Epoch loss = 0.00084
	Gradient norm: 0.01147835049778223
	Elapsed time = 1.761 min
[Statepoint 0]
	kT = 2.461 ref_kT = 2.462
	Predicted entropy: -0.3438429534435272
	Predicted free_energy: -13552.23828125
	Predicted pressure: 12.984464645385742
[Step Size] Found optimal step size 0.16732178628444672 with residual 0.006402730941772461

[DiffTRe] Epoch 229
	Epoch loss = 0.00098
	Gradient norm: 0.02321508526802063
	Elapsed time = 1.758 min
[Statepoint 0]
	kT = 2.467 ref_kT = 2.462
	Predicted entropy: -0.7018745541572571
	Predicted free_energy: -13762.921875
	Predicted pressure: 5.062629222869873
[Step Size] Found optimal step size 0.8099192380905151 with residual 0.021932363510131836

[DiffTRe] Epoch 230
	Epoch loss = 0.00076
	Gradient norm: 0.00038735748967155814
	Elapsed time = 1.763 min
[Statepoint 0]
	kT = 2.461 ref_kT = 2.462
	Predicted entropy: -0.4871973693370819
	Predicted free_energy: -13656.9189453125
	Predicted pressure: 7.775619029998779
[Step Size] Found optimal step size 0.2319590449333191 with residual 0.0018267631530761719

[DiffTRe] Epoch 231
	Epoch loss = 0.00092
	Gradient norm: 0.012251690030097961
	Elapsed time = 1.764 min
[Statepoint 0]
	kT = 2.457 ref_kT = 2.462
	Predicted entropy: -0.09575682878494263
	Predicted free_energy: -13131.0595703125
	Predicted pressure: 23.158708572387695
[Step Size] Found optimal step size 0.3893176317214966 with residual 0.0001010894775390625

[DiffTRe] Epoch 232
	Epoch loss = 0.00077
	Gradient norm: 0.002982616890221834
	Elapsed time = 1.756 min
[Statepoint 0]
	kT = 2.452 ref_kT = 2.462
	Predicted entropy: -0.40538516640663147
	Predicted free_energy: -13606.5966796875
	Predicted pressure: 8.175992012023926
[Step Size] Found optimal step size 0.2786518633365631 with residual 0.0026273727416992188

[DiffTRe] Epoch 233
	Epoch loss = 0.00102
	Gradient norm: 0.019061878323554993
	Elapsed time = 1.762 min
[Statepoint 0]
	kT = 2.457 ref_kT = 2.462
	Predicted entropy: -0.31802523136138916
	Predicted free_energy: -13795.171875
	Predicted pressure: 1.1713640689849854
[Step Size] Found optimal step size 0.27498847246170044 with residual 0.000911712646484375

[DiffTRe] Epoch 234
	Epoch loss = 0.00080
	Gradient norm: 0.006083427928388119
	Elapsed time = 1.760 min
[Statepoint 0]
	kT = 2.459 ref_kT = 2.462
	Predicted entropy: -0.5919657945632935
	Predicted free_energy: -13832.93359375
	Predicted pressure: 1.8189667463302612
[Step Size] Found optimal step size 0.318663626909256 with residual 0.0030698776245117188

[DiffTRe] Epoch 235
	Epoch loss = 0.00081
	Gradient norm: 0.004472274333238602
	Elapsed time = 1.758 min
[Statepoint 0]
	kT = 2.464 ref_kT = 2.462
	Predicted entropy: -0.37188827991485596
	Predicted free_energy: -13516.201171875
	Predicted pressure: 12.918968200683594
[Step Size] Found optimal step size 0.16386328637599945 with residual 0.01461338996887207

[DiffTRe] Epoch 236
	Epoch loss = 0.00085
	Gradient norm: 0.014415902085602283
	Elapsed time = 1.756 min
[Statepoint 0]
	kT = 2.461 ref_kT = 2.462
	Predicted entropy: -0.6580810546875
	Predicted free_energy: -13768.4990234375
	Predicted pressure: 7.826974868774414
[Step Size] Found optimal step size 1.0 with residual 1.9564547538757324

[DiffTRe] Epoch 237
	Epoch loss = 0.00075
	Gradient norm: 0.00013546158152166754
	Elapsed time = 1.758 min
[Statepoint 0]
	kT = 2.468 ref_kT = 2.462
	Predicted entropy: -0.4509068727493286
	Predicted free_energy: -13626.9443359375
	Predicted pressure: 11.98849868774414
[Step Size] Found optimal step size 0.30845582485198975 with residual 0.0018401145935058594

[DiffTRe] Epoch 238
	Epoch loss = 0.00081
	Gradient norm: 0.004640280734747648
	Elapsed time = 1.761 min
[Statepoint 0]
	kT = 2.473 ref_kT = 2.462
	Predicted entropy: -0.4719264507293701
	Predicted free_energy: -13901.16796875
	Predicted pressure: 1.0696724653244019
[Step Size] Found optimal step size 0.256963312625885 with residual 0.00420832633972168

[DiffTRe] Epoch 239
	Epoch loss = 0.00098
	Gradient norm: 0.027450840920209885
	Elapsed time = 1.763 min
[Statepoint 0]
	kT = 2.475 ref_kT = 2.462
	Predicted entropy: -0.7248397469520569
	Predicted free_energy: -13886.27734375
	Predicted pressure: -4.115372657775879
[Step Size] Found optimal step size 0.43145909905433655 with residual 0.0011396408081054688

[DiffTRe] Epoch 240
	Epoch loss = 0.00078
	Gradient norm: 0.004702244885265827
	Elapsed time = 1.758 min
[Statepoint 0]
	kT = 2.461 ref_kT = 2.462
	Predicted entropy: -0.4012427031993866
	Predicted free_energy: -13635.056640625
	Predicted pressure: 9.302977561950684
[Step Size] Found optimal step size 0.211127370595932 with residual 0.008641958236694336

[DiffTRe] Epoch 241
	Epoch loss = 0.00083
	Gradient norm: 0.010596067644655704
	Elapsed time = 1.760 min
[Statepoint 0]
	kT = 2.463 ref_kT = 2.462
	Predicted entropy: -0.6871352791786194
	Predicted free_energy: -13883.5703125
	Predicted pressure: 2.6880476474761963
[Step Size] Found optimal step size 0.49969762563705444 with residual 0.025888681411743164

[DiffTRe] Epoch 242
	Epoch loss = 0.00075
	Gradient norm: 0.0023738136515021324
	Elapsed time = 1.761 min
[Statepoint 0]
	kT = 2.471 ref_kT = 2.462
	Predicted entropy: -0.4544769823551178
	Predicted free_energy: -13681.771484375
	Predicted pressure: 7.619271755218506
[Step Size] Found optimal step size 0.28359436988830566 with residual 0.0005469322204589844

[DiffTRe] Epoch 243
	Epoch loss = 0.00090
	Gradient norm: 0.022400256246328354
	Elapsed time = 1.763 min
[Statepoint 0]
	kT = 2.460 ref_kT = 2.462
	Predicted entropy: -0.734591007232666
	Predicted free_energy: -13894.4609375
	Predicted pressure: -5.560521602630615
[Step Size] Found optimal step size 0.5268129706382751 with residual 0.00018548965454101562

[DiffTRe] Epoch 244
	Epoch loss = 0.00078
	Gradient norm: 0.008294540457427502
	Elapsed time = 1.762 min
[Statepoint 0]
	kT = 2.462 ref_kT = 2.462
	Predicted entropy: -0.3479192554950714
	Predicted free_energy: -13656.919921875
	Predicted pressure: 7.652688026428223
[Step Size] Found optimal step size 0.2324371635913849 with residual 0.0006461143493652344

[DiffTRe] Epoch 245
	Epoch loss = 0.00105
	Gradient norm: 0.034134361892938614
	Elapsed time = 1.761 min
[Statepoint 0]
	kT = 2.460 ref_kT = 2.462
	Predicted entropy: -0.5829321146011353
	Predicted free_energy: -13553.0498046875
	Predicted pressure: 11.543190002441406
[Step Size] Found optimal step size 0.6363672614097595 with residual 0.0001430511474609375

[DiffTRe] Epoch 246
	Epoch loss = 0.00071
	Gradient norm: 0.0013186271535232663
	Elapsed time = 1.761 min
[Statepoint 0]
	kT = 2.472 ref_kT = 2.462
	Predicted entropy: -0.3566218614578247
	Predicted free_energy: -13497.166015625
	Predicted pressure: 13.254358291625977
[Step Size] Found optimal step size 0.5152081847190857 with residual 0.002469301223754883

[DiffTRe] Epoch 247
	Epoch loss = 0.00085
	Gradient norm: 0.008010191842913628
	Elapsed time = 1.763 min
[Statepoint 0]
	kT = 2.465 ref_kT = 2.462
	Predicted entropy: -0.5530474185943604
	Predicted free_energy: -14032.42578125
	Predicted pressure: -2.2074172496795654
[Step Size] Found optimal step size 0.24764341115951538 with residual 0.0004322528839111328

[DiffTRe] Epoch 248
	Epoch loss = 0.00095
	Gradient norm: 0.025548892095685005
	Elapsed time = 1.761 min
[Statepoint 0]
	kT = 2.463 ref_kT = 2.462
	Predicted entropy: -0.5645621418952942
	Predicted free_energy: -13684.3017578125
	Predicted pressure: 11.203063011169434
[Step Size] Found optimal step size 1.0 with residual 0.07136940956115723

[DiffTRe] Epoch 249
	Epoch loss = 0.00071
	Gradient norm: 0.0010438506724312901
	Elapsed time = 1.762 min
[Statepoint 0]
	kT = 2.462 ref_kT = 2.462
	Predicted entropy: -0.3621370196342468
	Predicted free_energy: -13583.3017578125
	Predicted pressure: 14.006451606750488
[Step Size] Found optimal step size 0.3275800347328186 with residual 0.001941680908203125

[DiffTRe] Epoch 250
	Epoch loss = 0.00094
	Gradient norm: 0.015199070796370506
	Elapsed time = 1.761 min
[Statepoint 0]
	kT = 2.461 ref_kT = 2.462
	Predicted entropy: -0.7869624495506287
	Predicted free_energy: -14269.853515625
	Predicted pressure: -13.619596481323242
[Step Size] Found optimal step size 0.49910151958465576 with residual 0.0006124973297119141

[DiffTRe] Epoch 251
	Epoch loss = 0.00076
	Gradient norm: 0.004831408150494099
	Elapsed time = 1.763 min
[Statepoint 0]
	kT = 2.473 ref_kT = 2.462
	Predicted entropy: -0.37312835454940796
	Predicted free_energy: -13795.921875
	Predicted pressure: -0.665032148361206
[Step Size] Found optimal step size 0.3967747688293457 with residual 0.0017504692077636719

[DiffTRe] Epoch 252
	Epoch loss = 0.00092
	Gradient norm: 0.01707988791167736
	Elapsed time = 1.758 min
[Statepoint 0]
	kT = 2.456 ref_kT = 2.462
	Predicted entropy: -0.5318353176116943
	Predicted free_energy: -13664.4150390625
	Predicted pressure: 5.685359954833984
[Step Size] Found optimal step size 0.3816741108894348 with residual 0.0002636909484863281

[DiffTRe] Epoch 253
	Epoch loss = 0.00081
	Gradient norm: 0.01240734476596117
	Elapsed time = 1.764 min
[Statepoint 0]
	kT = 2.458 ref_kT = 2.462
	Predicted entropy: -0.2478383332490921
	Predicted free_energy: -13537.970703125
	Predicted pressure: 17.465688705444336
[Step Size] Found optimal step size 0.47258228063583374 with residual 0.0002193450927734375

[DiffTRe] Epoch 254
	Epoch loss = 0.00076
	Gradient norm: 0.008069118484854698
	Elapsed time = 1.759 min
[Statepoint 0]
	kT = 2.477 ref_kT = 2.462
	Predicted entropy: -0.5724289417266846
	Predicted free_energy: -13824.4130859375
	Predicted pressure: 5.516286373138428
[Step Size] Found optimal step size 0.36494046449661255 with residual 0.0017056465148925781

[DiffTRe] Epoch 255
	Epoch loss = 0.00085
	Gradient norm: 0.012079142965376377
	Elapsed time = 1.760 min
[Statepoint 0]
	kT = 2.462 ref_kT = 2.462
	Predicted entropy: -0.28732988238334656
	Predicted free_energy: -13604.6572265625
	Predicted pressure: 13.299365997314453
[Step Size] Found optimal step size 0.4308505654335022 with residual 0.0010685920715332031

[DiffTRe] Epoch 256
	Epoch loss = 0.00072
	Gradient norm: 0.004869780037552118
	Elapsed time = 1.761 min
[Statepoint 0]
	kT = 2.456 ref_kT = 2.462
	Predicted entropy: -0.48849281668663025
	Predicted free_energy: -13765.6728515625
	Predicted pressure: 3.3695290088653564
[Step Size] Found optimal step size 0.3402218222618103 with residual 0.0011143684387207031

[DiffTRe] Epoch 257
	Epoch loss = 0.00078
	Gradient norm: 0.00898209773004055
	Elapsed time = 1.760 min
[Statepoint 0]
	kT = 2.456 ref_kT = 2.462
	Predicted entropy: -0.22982913255691528
	Predicted free_energy: -13513.806640625
	Predicted pressure: 13.19920539855957
[Step Size] Found optimal step size 0.47709015011787415 with residual 0.005435466766357422

[DiffTRe] Epoch 258
	Epoch loss = 0.00074
	Gradient norm: 0.005326222628355026
	Elapsed time = 1.758 min
[Statepoint 0]
	kT = 2.468 ref_kT = 2.462
	Predicted entropy: -0.5476616024971008
	Predicted free_energy: -13767.3232421875
	Predicted pressure: 7.95058536529541
[Step Size] Found optimal step size 0.4049646258354187 with residual 2.2411346435546875e-05

[DiffTRe] Epoch 259
	Epoch loss = 0.00078
	Gradient norm: 0.00391659140586853
	Elapsed time = 1.762 min
[Statepoint 0]
	kT = 2.454 ref_kT = 2.462
	Predicted entropy: -0.24580182135105133
	Predicted free_energy: -13450.2109375
	Predicted pressure: 16.497983932495117
[Step Size] Found optimal step size 0.4734329581260681 with residual 0.0009829998016357422

[DiffTRe] Epoch 260
	Epoch loss = 0.00071
	Gradient norm: 0.002744686556980014
	Elapsed time = 1.762 min
[Statepoint 0]
	kT = 2.471 ref_kT = 2.462
	Predicted entropy: -0.490560919046402
	Predicted free_energy: -13769.642578125
	Predicted pressure: 8.016983985900879
[Step Size] Found optimal step size 0.4655473232269287 with residual 0.0007920265197753906

[DiffTRe] Epoch 261
	Epoch loss = 0.00087
	Gradient norm: 0.016904881224036217
	Elapsed time = 1.758 min
[Statepoint 0]
	kT = 2.469 ref_kT = 2.462
	Predicted entropy: -0.3053056299686432
	Predicted free_energy: -13774.5888671875
	Predicted pressure: 2.490873098373413
[Step Size] Found optimal step size 0.35537832975387573 with residual 0.10056543350219727

[DiffTRe] Epoch 262
	Epoch loss = 0.00084
	Gradient norm: 0.017066210508346558
	Elapsed time = 1.763 min
[Statepoint 0]
	kT = 2.465 ref_kT = 2.462
	Predicted entropy: -0.5871603488922119
	Predicted free_energy: -13925.9267578125
	Predicted pressure: 1.6779022216796875
[Step Size] Found optimal step size 1.0 with residual 0.9009661674499512

[DiffTRe] Epoch 263
	Epoch loss = 0.00070
	Gradient norm: 0.0012726477580145001
	Elapsed time = 1.759 min
[Statepoint 0]
	kT = 2.455 ref_kT = 2.462
	Predicted entropy: -0.32889604568481445
	Predicted free_energy: -13695.7236328125
	Predicted pressure: 8.566021919250488
[Step Size] Found optimal step size 0.3150809705257416 with residual 6.389617919921875e-05

[DiffTRe] Epoch 264
	Epoch loss = 0.00077
	Gradient norm: 0.01206545066088438
	Elapsed time = 1.757 min
[Statepoint 0]
	kT = 2.463 ref_kT = 2.462
	Predicted entropy: -0.5272964835166931
	Predicted free_energy: -13939.3583984375
	Predicted pressure: -1.0292900800704956
[Step Size] Found optimal step size 0.6370129585266113 with residual 0.0013926029205322266

[DiffTRe] Epoch 265
	Epoch loss = 0.00072
	Gradient norm: 0.003812221111729741
	Elapsed time = 1.761 min
[Statepoint 0]
	kT = 2.454 ref_kT = 2.462
	Predicted entropy: -0.250454306602478
	Predicted free_energy: -13711.3349609375
	Predicted pressure: 12.75805377960205
[Step Size] Found optimal step size 0.3756636381149292 with residual 0.0003561973571777344

[DiffTRe] Epoch 266
	Epoch loss = 0.00085
	Gradient norm: 0.018855415284633636
	Elapsed time = 1.759 min
[Statepoint 0]
	kT = 2.463 ref_kT = 2.462
	Predicted entropy: -0.6544051170349121
	Predicted free_energy: -13841.86328125
	Predicted pressure: 9.36098861694336
[Step Size] Found optimal step size 1.0 with residual 0.25391721725463867

[DiffTRe] Epoch 267
	Epoch loss = 0.00067
	Gradient norm: 0.0027034946251660585
	Elapsed time = 1.759 min
[Statepoint 0]
	kT = 2.476 ref_kT = 2.462
	Predicted entropy: -0.3929193317890167
	Predicted free_energy: -13634.6640625
	Predicted pressure: 11.044511795043945
[Step Size] Found optimal step size 0.3501192331314087 with residual 0.00026416778564453125

[DiffTRe] Epoch 268
	Epoch loss = 0.00088
	Gradient norm: 0.021266553550958633
	Elapsed time = 1.761 min
[Statepoint 0]
	kT = 2.468 ref_kT = 2.462
	Predicted entropy: -0.7146888971328735
	Predicted free_energy: -14190.6298828125
	Predicted pressure: -6.023558139801025
[Step Size] Found optimal step size 1.0 with residual 0.6869020462036133

[DiffTRe] Epoch 269
	Epoch loss = 0.00067
	Gradient norm: 0.0009222234366461635
	Elapsed time = 1.759 min
[Statepoint 0]
	kT = 2.468 ref_kT = 2.462
	Predicted entropy: -0.33463358879089355
	Predicted free_energy: -13825.3134765625
	Predicted pressure: 7.037118911743164
[Step Size] Found optimal step size 0.4057159125804901 with residual 0.0008068084716796875

[DiffTRe] Epoch 270
	Epoch loss = 0.00081
	Gradient norm: 0.0168205164372921
	Elapsed time = 1.761 min
[Statepoint 0]
	kT = 2.456 ref_kT = 2.462
	Predicted entropy: -0.44601163268089294
	Predicted free_energy: -13590.53125
	Predicted pressure: 11.025237083435059
[Step Size] Found optimal step size 0.4936188757419586 with residual 0.0001068115234375

[DiffTRe] Epoch 271
	Epoch loss = 0.00075
	Gradient norm: 0.008387504145503044
	Elapsed time = 1.764 min
[Statepoint 0]
	kT = 2.471 ref_kT = 2.462
	Predicted entropy: -0.21470722556114197
	Predicted free_energy: -13574.0888671875
	Predicted pressure: 17.874404907226562
[Step Size] Found optimal step size 0.6739326119422913 with residual 0.0019981861114501953

[DiffTRe] Epoch 272
	Epoch loss = 0.00067
	Gradient norm: 0.003812673268839717
	Elapsed time = 1.762 min
[Statepoint 0]
	kT = 2.469 ref_kT = 2.462
	Predicted entropy: -0.5432747602462769
	Predicted free_energy: -13969.2666015625
	Predicted pressure: 3.0785696506500244
[Step Size] Found optimal step size 0.42884498834609985 with residual 0.0009860992431640625

[DiffTRe] Epoch 273
	Epoch loss = 0.00077
	Gradient norm: 0.009115997701883316
	Elapsed time = 1.764 min
[Statepoint 0]
	kT = 2.462 ref_kT = 2.462
	Predicted entropy: -0.2792910039424896
	Predicted free_energy: -13681.25
	Predicted pressure: 12.688334465026855
[Step Size] Found optimal step size 0.6714552044868469 with residual 0.0010733604431152344

[DiffTRe] Epoch 274
	Epoch loss = 0.00067
	Gradient norm: 0.005233466159552336
	Elapsed time = 1.760 min
[Statepoint 0]
	kT = 2.455 ref_kT = 2.462
	Predicted entropy: -0.5208068490028381
	Predicted free_energy: -13813.638671875
	Predicted pressure: 4.605494976043701
[Step Size] Found optimal step size 0.4107018709182739 with residual 0.005812883377075195

[DiffTRe] Epoch 275
	Epoch loss = 0.00085
	Gradient norm: 0.01829446107149124
	Elapsed time = 1.764 min
[Statepoint 0]
	kT = 2.458 ref_kT = 2.462
	Predicted entropy: -0.21199533343315125
	Predicted free_energy: -13628.6806640625
	Predicted pressure: 13.187896728515625
[Step Size] Found optimal step size 1.0 with residual 1.774580478668213

[DiffTRe] Epoch 276
	Epoch loss = 0.00065
	Gradient norm: 0.0007239797851070762
	Elapsed time = 1.758 min
[Statepoint 0]
	kT = 2.467 ref_kT = 2.462
	Predicted entropy: -0.44919049739837646
	Predicted free_energy: -13817.154296875
	Predicted pressure: 1.9278284311294556
[Step Size] Found optimal step size 1.0 with residual 0.6851344108581543

[DiffTRe] Epoch 277
	Epoch loss = 0.00066
	Gradient norm: 0.0005113395163789392
	Elapsed time = 1.761 min
[Statepoint 0]
	kT = 2.460 ref_kT = 2.462
	Predicted entropy: -0.3471807539463043
	Predicted free_energy: -13695.8232421875
	Predicted pressure: 7.549846172332764
[Step Size] Found optimal step size 0.3507836163043976 with residual 0.0019783973693847656

[DiffTRe] Epoch 278
	Epoch loss = 0.00077
	Gradient norm: 0.013715056702494621
	Elapsed time = 1.759 min
[Statepoint 0]
	kT = 2.459 ref_kT = 2.462
	Predicted entropy: -0.5781703591346741
	Predicted free_energy: -13763.2607421875
	Predicted pressure: 6.163106918334961
[Step Size] Found optimal step size 1.0 with residual 0.08103084564208984

[DiffTRe] Epoch 279
	Epoch loss = 0.00065
	Gradient norm: 0.0015004201559349895
	Elapsed time = 1.762 min
[Statepoint 0]
	kT = 2.457 ref_kT = 2.462
	Predicted entropy: -0.34780824184417725
	Predicted free_energy: -13629.58984375
	Predicted pressure: 17.36839485168457
[Step Size] Found optimal step size 0.5364433526992798 with residual 0.0007619857788085938

[DiffTRe] Epoch 280
	Epoch loss = 0.00086
	Gradient norm: 0.010341592133045197
	Elapsed time = 1.760 min
[Statepoint 0]
	kT = 2.464 ref_kT = 2.462
	Predicted entropy: -0.5713717341423035
	Predicted free_energy: -14341.5673828125
	Predicted pressure: -12.00585651397705
[Step Size] Found optimal step size 1.0 with residual 1.0993256568908691

[DiffTRe] Epoch 281
	Epoch loss = 0.00063
	Gradient norm: 0.00020760866755153984
	Elapsed time = 1.762 min
[Statepoint 0]
	kT = 2.459 ref_kT = 2.462
	Predicted entropy: -0.39104947447776794
	Predicted free_energy: -13842.4248046875
	Predicted pressure: 4.350360870361328
[Step Size] Found optimal step size 0.44628655910491943 with residual 0.0548090934753418

[DiffTRe] Epoch 282
	Epoch loss = 0.00074
	Gradient norm: 0.01393510028719902
	Elapsed time = 1.760 min
[Statepoint 0]
	kT = 2.467 ref_kT = 2.462
	Predicted entropy: -0.19984468817710876
	Predicted free_energy: -13591.4599609375
	Predicted pressure: 18.087907791137695
[Step Size] Found optimal step size 0.670312762260437 with residual 0.0018048286437988281

[DiffTRe] Epoch 283
	Epoch loss = 0.00067
	Gradient norm: 0.0056343115866184235
	Elapsed time = 1.762 min
[Statepoint 0]
	kT = 2.458 ref_kT = 2.462
	Predicted entropy: -0.5312609672546387
	Predicted free_energy: -13844.3486328125
	Predicted pressure: 6.258537769317627
[Step Size] Found optimal step size 0.6506420373916626 with residual 0.0007810592651367188

[DiffTRe] Epoch 284
	Epoch loss = 0.00070
	Gradient norm: 0.0066388570703566074
	Elapsed time = 1.763 min
[Statepoint 0]
	kT = 2.444 ref_kT = 2.462
	Predicted entropy: -0.3184623420238495
	Predicted free_energy: -13703.5908203125
	Predicted pressure: 12.492599487304688
[Step Size] Found optimal step size 0.622837483882904 with residual 6.9141387939453125e-06

[DiffTRe] Epoch 285
	Epoch loss = 0.00070
	Gradient norm: 0.00894918292760849
	Elapsed time = 1.763 min
[Statepoint 0]
	kT = 2.465 ref_kT = 2.462
	Predicted entropy: -0.6264825463294983
	Predicted free_energy: -13959.3525390625
	Predicted pressure: 2.7580044269561768
[Step Size] Found optimal step size 0.6980428099632263 with residual 0.00051116943359375

[DiffTRe] Epoch 286
	Epoch loss = 0.00071
	Gradient norm: 0.007530572824180126
	Elapsed time = 1.762 min
[Statepoint 0]
	kT = 2.461 ref_kT = 2.462
	Predicted entropy: -0.31681889295578003
	Predicted free_energy: -13779.5732421875
	Predicted pressure: 11.552346229553223
[Step Size] Found optimal step size 0.5330966711044312 with residual 0.00551915168762207

[DiffTRe] Epoch 287
	Epoch loss = 0.00075
	Gradient norm: 0.014269266277551651
	Elapsed time = 1.763 min
[Statepoint 0]
	kT = 2.462 ref_kT = 2.462
	Predicted entropy: -0.6468425989151001
	Predicted free_energy: -14038.134765625
	Predicted pressure: -5.400629997253418
[Step Size] Found optimal step size 0.9090917706489563 with residual 0.011655092239379883

[DiffTRe] Epoch 288
	Epoch loss = 0.00067
	Gradient norm: 0.0040963757783174515
	Elapsed time = 1.762 min
[Statepoint 0]
	kT = 2.470 ref_kT = 2.462
	Predicted entropy: -0.3053446412086487
	Predicted free_energy: -13681.048828125
	Predicted pressure: 11.110568046569824
[Step Size] Found optimal step size 0.39355844259262085 with residual 0.0023622512817382812

[DiffTRe] Epoch 289
	Epoch loss = 0.00079
	Gradient norm: 0.02197999507188797
	Elapsed time = 1.762 min
[Statepoint 0]
	kT = 2.467 ref_kT = 2.462
	Predicted entropy: -0.5847234725952148
	Predicted free_energy: -13912.2529296875
	Predicted pressure: 4.470338821411133
[Step Size] Found optimal step size 1.0 with residual 2.269263744354248

[DiffTRe] Epoch 290
	Epoch loss = 0.00059
	Gradient norm: 0.00014032617036718875
	Elapsed time = 1.763 min
[Statepoint 0]
	kT = 2.472 ref_kT = 2.462
	Predicted entropy: -0.3915296494960785
	Predicted free_energy: -13822.7978515625
	Predicted pressure: 6.6984734535217285
[Step Size] Found optimal step size 1.0 with residual 1.4583687782287598

[DiffTRe] Epoch 291
	Epoch loss = 0.00061
	Gradient norm: 0.0003177525941282511
	Elapsed time = 1.762 min
[Statepoint 0]
	kT = 2.461 ref_kT = 2.462
	Predicted entropy: -0.368774950504303
	Predicted free_energy: -13808.76953125
	Predicted pressure: 4.906651020050049
[Step Size] Found optimal step size 0.49625781178474426 with residual 0.011878490447998047

[DiffTRe] Epoch 292
	Epoch loss = 0.00066
	Gradient norm: 0.00462500611320138
	Elapsed time = 1.760 min
[Statepoint 0]
	kT = 2.462 ref_kT = 2.462
	Predicted entropy: -0.5578121542930603
	Predicted free_energy: -14002.408203125
	Predicted pressure: 0.45346102118492126
[Step Size] Found optimal step size 0.7490634918212891 with residual 0.004949808120727539

[DiffTRe] Epoch 293
	Epoch loss = 0.00062
	Gradient norm: 0.004185026045888662
	Elapsed time = 1.763 min
[Statepoint 0]
	kT = 2.461 ref_kT = 2.462
	Predicted entropy: -0.26345306634902954
	Predicted free_energy: -13676.1220703125
	Predicted pressure: 10.673123359680176
[Step Size] Found optimal step size 0.5965480208396912 with residual 8.797645568847656e-05

[DiffTRe] Epoch 294
	Epoch loss = 0.00075
	Gradient norm: 0.014619100838899612
	Elapsed time = 1.760 min
[Statepoint 0]
	kT = 2.460 ref_kT = 2.462
	Predicted entropy: -0.6218453645706177
	Predicted free_energy: -14006.36328125
	Predicted pressure: 3.062483072280884
[Step Size] Found optimal step size 0.9008274078369141 with residual 0.0044324398040771484

[DiffTRe] Epoch 295
	Epoch loss = 0.00066
	Gradient norm: 0.006504797842353582
	Elapsed time = 1.765 min
[Statepoint 0]
	kT = 2.477 ref_kT = 2.462
	Predicted entropy: -0.35592448711395264
	Predicted free_energy: -13808.10546875
	Predicted pressure: 10.370424270629883
[Step Size] Found optimal step size 0.7145032286643982 with residual 0.0008113384246826172

[DiffTRe] Epoch 296
	Epoch loss = 0.00078
	Gradient norm: 0.01357301790267229
	Elapsed time = 1.759 min
[Statepoint 0]
	kT = 2.457 ref_kT = 2.462
	Predicted entropy: -0.584895133972168
	Predicted free_energy: -13839.78515625
	Predicted pressure: 7.514540672302246
[Step Size] Found optimal step size 0.8646964430809021 with residual 0.0008141994476318359

[DiffTRe] Epoch 297
	Epoch loss = 0.00066
	Gradient norm: 0.00481678219512105
	Elapsed time = 1.763 min
[Statepoint 0]
	kT = 2.467 ref_kT = 2.462
	Predicted entropy: -0.3300211429595947
	Predicted free_energy: -13786.205078125
	Predicted pressure: 10.507013320922852
[Step Size] Found optimal step size 0.9090917706489563 with residual 0.03108501434326172

[DiffTRe] Epoch 298
	Epoch loss = 0.00068
	Gradient norm: 0.004295304883271456
	Elapsed time = 1.760 min
[Statepoint 0]
	kT = 2.464 ref_kT = 2.462
	Predicted entropy: -0.4787406027317047
	Predicted free_energy: -13984.955078125
	Predicted pressure: -1.2944719791412354
[Step Size] Found optimal step size 0.7302807569503784 with residual 0.0009570121765136719

[DiffTRe] Epoch 299
	Epoch loss = 0.00069
	Gradient norm: 0.009580373764038086
	Elapsed time = 1.764 min
[Statepoint 0]
	kT = 2.456 ref_kT = 2.462
	Predicted entropy: -0.22430558502674103
	Predicted free_energy: -13703.98046875
	Predicted pressure: 16.0960636138916
Total training time:  8.8 hours

Results

predicted_quantities = results["predictions"][0]
loss_history = results["epoch_losses"]
num_updates = list(predicted_quantities.keys())[-1]

Loss History

Show code cell source

Hide code cell source

fig, ax = plt.subplots(1, 1, layout="constrained")
color = 'tab:red'
ax.set_xlabel('Update Step')
ax.set_ylabel('MSE Loss')
ax.plot(loss_history, color=color, label='Loss')

[<matplotlib.lines.Line2D at 0x7ef47dd7b850>]

RDF

Show code cell source

Hide code cell source

fig, ax = plt.subplots(1, 1, layout="constrained")
r = onp.linspace(0, 1, 300)
ax.plot(r, predicted_quantities[0]["rdf"], label='Initial')
ax.plot(r, predicted_quantities[num_updates]["rdf"], label='Predicted')
ax.plot(r, targets["rdf"]["target"], label='Reference')
ax.set_xlabel("r [nm]")
ax.set_ylabel("RDF")
ax.legend()

<matplotlib.legend.Legend at 0x7efa542fdcf0>

ADF

Show code cell source

Hide code cell source

fig, ax = plt.subplots(1, 1, layout="constrained")
alpha = onp.linspace(0, onp.pi, 150)
ax.plot(alpha, predicted_quantities[0]["adf"], label='Initial')
ax.plot(alpha, predicted_quantities[num_updates]["adf"], label='Predicted')
ax.plot(alpha, targets["adf"]["target"], label='Reference')
ax.set_xlabel("Angle [rad]")
ax.set_ylabel("ADF")
ax.legend()

<matplotlib.legend.Legend at 0x7ef47dc11390>

Pressure

Show code cell source

Hide code cell source

fig, ax = plt.subplots(1, 1, layout="constrained")
ax.set_xlabel('Update Step')
ax.set_ylabel('Pressure [kJ / (mol nm^3)]')
ax.plot([preds["pressure"] for preds in predicted_quantities.values()], label='Predicted')
ax.axhline(y=pressure_target, linestyle='--', label='Reference', color='k')
ax.legend()

<matplotlib.legend.Legend at 0x7ef47c253a90>

References

[1]

Reith, D.; Pütz, M.; Müller-Plathe, F. Deriving Effective Mesoscale Potentials from Atomistic Simulations: Mesoscale Potentials from Atomistic Simulations. J. Comput. Chem. 2003, 24 (13), 1624–1636. https://doi.org/10.1002/jcc.10307.

[2]

Lyubartsev, A. P.; Laaksonen, A. Calculation of Effective Interaction Potentials from Radial Distribution Functions: A Reverse Monte Carlo Approach. Phys. Rev. E 1995, 52 (4), 3730–3737. https://doi.org/10.1103/PhysRevE.52.3730.

[3]

Dhabal, D.; Singh, M.; Wikfeldt, K. T.; Chakravarty, C. Triplet Correlation Functions in Liquid Water. The Journal of Chemical Physics 2014, 141 (17), 174504. https://doi.org/10.1063/1.4898755.

[4]

Thaler, S.; Zavadlav, J. Learning Neural Network Potentials from Experimental Data via Differentiable Trajectory Reweighting. Nat Commun 2021, 12 (1), 6884. https://doi.org/10.1038/s41467-021-27241-4.

[5]

Torrie, G. M.; Valleau, J. P. Nonphysical Sampling Distributions in Monte Carlo Free-Energy Estimation: Umbrella Sampling. Journal of Computational Physics 1977, 23 (2), 187–199. https://doi.org/10.1016/0021-9991(77)90121-8.

[6]

Gasteiger, J.; Giri, S.; Margraf, J. T.; Günnemann, S. Fast and Uncertainty-Aware Directional Message Passing for Non-Equilibrium Molecules. http://arxiv.org/abs/2011.14115 (accessed 2023-06-29).

[7]

Soper, A. K.; Benmore, C. J. Quantum Differences between Heavy and Light Water. Phys. Rev. Lett. 2008, 101 (6), 065502. https://doi.org/10.1103/PhysRevLett.101.065502.