Differentiable Trajectory Reweighting (DiffTRe)#

Concepts#

Molecular dynamics is an efficient method to sample positions $\mathbf r$ from a molecular systems. However, back-propagating gradients through the simulation is costly and prone to exploding gradients [1]. Instead, DiffTRe [2] estimates gradients of ensemble averages by employing a probabilistic perspective like in the Umbrella Sampling method [3].

In the umbrella sampling method, biasing a potential enables an efficient computation of ensemble averages of an unbiased potential. A postprocessing step then corrects the effect of the bias on the ensemble average. This correction re-weights the collected samples by accounting for the change in relative probability between the biased reference and unbiased target distribution.

For a trainable target potential $U_\theta$ and samples $x^{(i)}$ from the biased reference potential $\tilde U$, the discretized reweighting reads

\[\langle a \rangle_{U_\theta} \approx \sum_{i=1}^N w^{(i)}a\left(x^{(i)}\right), \quad w^{(i)} = \frac{e^{-\beta\left(U_\theta(x^{(i)}) - \tilde U(x^{(i)})\right)}}{\sum_{j=1}^Ne^{-\beta\left(U_\theta(x^{(j)}) - \tilde U(x^{(j)})\right)}}\]

Since the reference and target potential are independent, DiffTRe assumes that the reference potential $\tilde U$ no longer depend on the learnable parameters. Therefore, also the samples $x^{(i)}$ are independent of the target potential. Hence, the only contribution to the gradients arises through the weights $w^{(i)}$ and the instantaneous states $a(x)$ of the observables. Thus, by employing this umbrella sampling procedure, DiffTRe can compute gradients of the loss function without differentiating through the costly molecular dynamics simulation.

Unfortunately, the statistical error of the approximation grows exponentially fast with the difference between the target and the reference potential [4]. If the number of effective samples[5]

\[N_{eff} = e^{-\sum_{i=1}^N w^{(i)}\log w^{(i)}}\]

decreases below a threshold, DiffTRe replaces the reference potential by the current potential $\tilde U \leftarrow U_\theta$ and resamples the conformations $x^{(i)}$.

To accelerate the resampling and use the full computational capabilities of a GPU, DiffTRe enables to sample from multiple simulations in parallel, using the vectorization capabilities of JAX. Additionally, DiffTRe provides the option to re-seed these parallel simulations by choosing initial states from the simulated trajectory, with probability corresponding to their weight. In principle, this resampling of initial states enables a faster convergence of the simulated trajectories to the new equilibrium distribution.

Toy Example#

For a canonical system of two-particles connected by a spring, the Boltzmann factor is

\[ \rho(\mathbf{r}) \propto e^{-\frac{1}{2}b(||\mathbf{r}_1 - \mathbf{r}_2|| - r_0)^2}, \quad b = \beta\cdot b_S,\]

with temperature dependent effective spring constant $b$.

Hence, the probability of finding the two identical particles in a distance of $r$ is

\[ p(r) = \sqrt{\frac{2b}{\pi}}\left(\frac{b}{1 + br_0^2}\right)r^{2}e^{-\frac{1}{2}b(r - r_0)^2}.\]

The term $r^2$ in front of the exponential factor emerges in the transformation from cartesian to spherical coordinates. With this probability distribution, we can directly compute the radial distribution function.

\[ g(r) = \frac{V}{4 \pi r^2 N^2}p(r).\]

box = 1.0

def radial_distribution(r, r_0=0.35, b=250.0, kbt=2.56):
    b = b / kbt
    norm = onp.sqrt(onp.pi / (2 * b)) * (1 + b * r_0 ** 2) / b
    g_r = box ** 3 / (16 * onp.pi) * onp.exp(-0.5 * b * (r - r_0) ** 2) / norm
    return g_r

We now want to learn the parameters of this harmonic bond based on a reference radial distribution function.

r = onp.linspace(0.0, box, 100)
target = onp.vstack((r, radial_distribution(r))).T

Although we could find an analytic relation to the potential parameters, this is not possible for more complex systems. Thus, we need to simulate this relation and set up a model of the system.

Thus, we first need to define an appropriate potential model.

from jax_md import energy, space, simulate, partition

def energy_fn_template(params):
    energy_fn = energy.simple_spring_bond(
        displacement_fn,
        jnp.asarray([[0, 1]]),
        length=params["r_0"],
        epsilon=100 * params["scaled_b"],
        alpha=2.0
    )
    return energy_fn

init_params = {"r_0": 0.3, "scaled_b": 1.5}

Secondly, we need a routine to simulate the positions of the particles.

r_init = jnp.asarray([[0.0, 0.0, 0.0], [0.11, 0.09, 0.12]])
displacement_fn, shift_fn = space.periodic_general(box)

dt = 0.01
timings = ensemble.sampling.process_printouts(dt, 1100, 100, 1.0)

simulator_template = partial(
    simulate.nvt_langevin, shift_fn=shift_fn,
    dt=dt, kT=2.56, gamma=0.5, mass=10.0)

neighbor_fn = partition.neighbor_list(displacement_fn, box, 0.5)

simulator_init, _ = simulator_template(energy_fn_template(init_params))
simulator_init_state = simulator_init(jax.random.PRNGKey(0), r_init)
nbrs_init = neighbor_fn.allocate(r_init)

reference_state = ensemble.sampling.SimulatorState(
    sim_state=simulator_init_state, nbrs=nbrs_init)

system = {
    'displacement_fn': displacement_fn,
    'reference_box': box
}

/home/docs/checkouts/readthedocs.org/user_builds/chemtrain/envs/latest/lib/python3.11/site-packages/jax/_src/numpy/reductions.py:230: 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/jax-ml/jax#current-gotchas for more.
  return _reduction(a, "sum", lax.add, 0, preproc=_cast_to_numeric,

There are multiple classical approaches that enable the inversion of a radial distribution function into a pair-potential. However, they are not applicable to general models, e.g., neural networks. Thus, DiffTRe enables gradient based training, which we are going to set up in the next step.

import optax

lr_schedule = optax.exponential_decay(-0.05, 300, 0.1)
optimizer = optax.chain(
    optax.scale_by_rms(0.9),
    optax.scale_by_schedule(lr_schedule)
)

Finally, we have to specify the training targets, which is in our case the radial distribution function. Since we only have two particles in a box, we approximate the distribution with slightly coarser bins.

target_builder = quantity.targets.TargetBuilder()

target_builder['rdf'] = quantity.targets.init_radial_distribution_target(
    target, rdf_start=0.00, rdf_cut=1.0, nbins=50)

targets, compute_fns = target_builder.build(system)

We now created a numerical representation of the system and can run the trainer.

trainer = trainers.Difftre(
    init_params, optimizer, reweight_ratio=0.99
)

trainer.add_statepoint(
  energy_fn_template, simulator_template, neighbor_fn, timings, 
  {'kT': 2.56}, compute_fns, reference_state, targets=targets)

/home/docs/checkouts/readthedocs.org/user_builds/chemtrain/envs/latest/lib/python3.11/site-packages/chemtrain/ensemble/reweighting.py:798: 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(

[Propagation] Time for trajectory compilation 0: 0.023458929856618245 mins

[Propagation] Time for trajectory simulation 0: 1.0329484939575196e-05 mins

trainer.train(300)

Show code cell output

Hide code cell output

[Propagate] Effective sample size: 1000.0015258789062 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.555 ref_kT = 2.560

	Predicted entropy: 1.1946637812343397e-07

	Predicted free_energy: 0.0

[DiffTRe] Epoch 0
	Epoch loss = 0.04675
	Gradient norm: 3.0267252922058105
	Elapsed time = 0.045 min

[Propagate] Effective sample size: 574.6226196289062 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.552 ref_kT = 2.560

	Predicted entropy: 0.0025615664198994637

	Predicted free_energy: -1.302438735961914

[DiffTRe] Epoch 1
	Epoch loss = 0.07730
	Gradient norm: 0.6168447136878967
	Elapsed time = 0.046 min

[Propagate] Effective sample size: 942.693359375 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.544 ref_kT = 2.560

	Predicted entropy: 0.002350736176595092

	Predicted free_energy: -1.0858967304229736

[DiffTRe] Epoch 2
	Epoch loss = 0.02244
	Gradient norm: 0.2206382304430008
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 975.7899780273438 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.665 ref_kT = 2.560

	Predicted entropy: 0.0017072916962206364

	Predicted free_energy: -0.7166634798049927

[DiffTRe] Epoch 3
	Epoch loss = 0.01341
	Gradient norm: 0.0071304962038993835
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 996.8170166015625 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.665 ref_kT = 2.560

	Predicted entropy: -0.1098317801952362

	Predicted free_energy: -0.5528524518013

[DiffTRe] Epoch 4
	Epoch loss = 0.01095
	Gradient norm: 0.0063537489622831345
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 997.1207885742188 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.665 ref_kT = 2.560

	Predicted entropy: -0.11276046931743622

	Predicted free_energy: -0.564949631690979

[DiffTRe] Epoch 5
	Epoch loss = 0.00937
	Gradient norm: 0.003944072872400284
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 991.7974853515625 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.665 ref_kT = 2.560

	Predicted entropy: -0.19238394498825073

	Predicted free_energy: -0.4502132833003998

[DiffTRe] Epoch 6
	Epoch loss = 0.00822
	Gradient norm: 0.007096436806023121
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 993.7431640625 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.665 ref_kT = 2.560

	Predicted entropy: -0.16906613111495972

	Predicted free_energy: -0.49813565611839294

[DiffTRe] Epoch 7
	Epoch loss = 0.00742
	Gradient norm: 0.010227860882878304
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 984.2299194335938 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.680 ref_kT = 2.560

	Predicted entropy: 0.0008471879409626126

	Predicted free_energy: -0.3460797965526581

[DiffTRe] Epoch 8
	Epoch loss = 0.00768
	Gradient norm: 0.04156037047505379
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 990.369384765625 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.680 ref_kT = 2.560

	Predicted entropy: 0.11925072222948074

	Predicted free_energy: -0.5497382283210754

[DiffTRe] Epoch 9
	Epoch loss = 0.00867
	Gradient norm: 0.10055117309093475
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 993.7327270507812 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.680 ref_kT = 2.560

	Predicted entropy: -0.1448879837989807

	Predicted free_energy: -0.12953561544418335

[DiffTRe] Epoch 10
	Epoch loss = 0.01238
	Gradient norm: 0.44592517614364624
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 940.8705444335938 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.596 ref_kT = 2.560

	Predicted entropy: 0.0016361231682822108

	Predicted free_energy: -0.8031996488571167

[DiffTRe] Epoch 11
	Epoch loss = 0.03036
	Gradient norm: 0.6291072368621826
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 881.1871337890625 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.666 ref_kT = 2.560

	Predicted entropy: 0.0002903615531977266

	Predicted free_energy: -0.08003389835357666

[DiffTRe] Epoch 12
	Epoch loss = 0.04070
	Gradient norm: 3.371819496154785
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 749.9911499023438 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.504 ref_kT = 2.560

	Predicted entropy: 0.001884909812361002

	Predicted free_energy: -1.0086530447006226

[DiffTRe] Epoch 13
	Epoch loss = 0.07177
	Gradient norm: 0.9690219759941101
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 931.703125 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.622 ref_kT = 2.560

	Predicted entropy: 0.001313022687099874

	Predicted free_energy: -0.6317083239555359

[DiffTRe] Epoch 14
	Epoch loss = 0.01275
	Gradient norm: 0.12587369978427887
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 986.425048828125 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.605 ref_kT = 2.560

	Predicted entropy: 0.0008622339810244739

	Predicted free_energy: -0.40608155727386475

[DiffTRe] Epoch 15
	Epoch loss = 0.00537
	Gradient norm: 0.0002004766429308802
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 999.892822265625 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.605 ref_kT = 2.560

	Predicted entropy: -0.02099117450416088

	Predicted free_energy: -0.3771728575229645

[DiffTRe] Epoch 16
	Epoch loss = 0.00487
	Gradient norm: 0.00014944502618163824
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.6864013671875 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.605 ref_kT = 2.560

	Predicted entropy: -0.03707002103328705

	Predicted free_energy: -0.3571118414402008

[DiffTRe] Epoch 17
	Epoch loss = 0.00444
	Gradient norm: 0.0001369121455354616
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.312255859375 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.605 ref_kT = 2.560

	Predicted entropy: -0.05550415441393852

	Predicted free_energy: -0.3329687714576721

[DiffTRe] Epoch 18
	Epoch loss = 0.00407
	Gradient norm: 0.00011938781244680285
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 998.9663696289062 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.605 ref_kT = 2.560

	Predicted entropy: -0.06863317638635635

	Predicted free_energy: -0.31694093346595764

[DiffTRe] Epoch 19
	Epoch loss = 0.00375
	Gradient norm: 0.00011763078509829938
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 998.4063720703125 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.605 ref_kT = 2.560

	Predicted entropy: -0.0856819674372673

	Predicted free_energy: -0.2944970726966858

[DiffTRe] Epoch 20
	Epoch loss = 0.00346
	Gradient norm: 0.00013343342288862914
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 998.0565185546875 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.605 ref_kT = 2.560

	Predicted entropy: -0.09509997069835663

	Predicted free_energy: -0.283806174993515

[DiffTRe] Epoch 21
	Epoch loss = 0.00321
	Gradient norm: 0.0001877536706160754
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 997.2396850585938 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.605 ref_kT = 2.560

	Predicted entropy: -0.11381367594003677

	Predicted free_energy: -0.25841066241264343

[DiffTRe] Epoch 22
	Epoch loss = 0.00299
	Gradient norm: 0.00040229083970189095
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 997.1849975585938 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.605 ref_kT = 2.560

	Predicted entropy: -0.11509336531162262

	Predicted free_energy: -0.2599780559539795

[DiffTRe] Epoch 23
	Epoch loss = 0.00281
	Gradient norm: 0.0010877206223085523
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 995.478515625 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.605 ref_kT = 2.560

	Predicted entropy: -0.14604270458221436

	Predicted free_energy: -0.21540285646915436

[DiffTRe] Epoch 24
	Epoch loss = 0.00267
	Gradient norm: 0.004139352589845657
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 996.857421875 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.605 ref_kT = 2.560

	Predicted entropy: -0.11631663143634796

	Predicted free_energy: -0.2645837664604187

[DiffTRe] Epoch 25
	Epoch loss = 0.00273
	Gradient norm: 0.01669508032500744
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 989.3724975585938 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.663 ref_kT = 2.560

	Predicted entropy: 0.00011963657016167417

	Predicted free_energy: -0.10676842927932739

[DiffTRe] Epoch 26
	Epoch loss = 0.00338
	Gradient norm: 0.0835823193192482
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 972.6830444335938 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.581 ref_kT = 2.560

	Predicted entropy: 0.0008411420858465135

	Predicted free_energy: -0.40584319829940796

[DiffTRe] Epoch 27
	Epoch loss = 0.00503
	Gradient norm: 0.22138555347919464
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 928.9136962890625 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.643 ref_kT = 2.560

	Predicted entropy: -0.0004550320445559919

	Predicted free_energy: 0.11697715520858765

[DiffTRe] Epoch 28
	Epoch loss = 0.01502
	Gradient norm: 1.2771906852722168
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 780.719482421875 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.451 ref_kT = 2.560

	Predicted entropy: 0.0013940442586317658

	Predicted free_energy: -0.7062948346138

[DiffTRe] Epoch 29
	Epoch loss = 0.04070
	Gradient norm: 0.9780896306037903
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 877.9375610351562 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.562 ref_kT = 2.560

	Predicted entropy: 0.00022351962979882956

	Predicted free_energy: -0.12358295917510986

[DiffTRe] Epoch 30
	Epoch loss = 0.00222
	Gradient norm: 0.002429415937513113
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 999.7049560546875 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.562 ref_kT = 2.560

	Predicted entropy: 0.016487905755639076

	Predicted free_energy: -0.1509498953819275

[DiffTRe] Epoch 31
	Epoch loss = 0.00212
	Gradient norm: 0.0035836747847497463
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.822265625 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.562 ref_kT = 2.560

	Predicted entropy: -0.029248781502246857

	Predicted free_energy: -0.08645527809858322

[DiffTRe] Epoch 32
	Epoch loss = 0.00207
	Gradient norm: 0.007771903648972511
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.2498168945312 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.562 ref_kT = 2.560

	Predicted entropy: 0.015040313825011253

	Predicted free_energy: -0.15467670559883118

[DiffTRe] Epoch 33
	Epoch loss = 0.00214
	Gradient norm: 0.016704298555850983
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 998.4649047851562 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.562 ref_kT = 2.560

	Predicted entropy: -0.07845155149698257

	Predicted free_energy: -0.020106399431824684

[DiffTRe] Epoch 34
	Epoch loss = 0.00249
	Gradient norm: 0.05101644620299339
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 995.0091552734375 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.562 ref_kT = 2.560

	Predicted entropy: 0.061938658356666565

	Predicted free_energy: -0.2295253425836563

[DiffTRe] Epoch 35
	Epoch loss = 0.00407
	Gradient norm: 0.13097934424877167
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 987.9935302734375 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.610 ref_kT = 2.560

	Predicted entropy: -0.0003382210270501673

	Predicted free_energy: 0.1293480396270752

[DiffTRe] Epoch 36
	Epoch loss = 0.01828
	Gradient norm: 1.7987767457962036
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 804.3521728515625 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.455 ref_kT = 2.560

	Predicted entropy: 0.0010280682472512126

	Predicted free_energy: -0.6366381645202637

[DiffTRe] Epoch 37
	Epoch loss = 0.05080
	Gradient norm: 1.2871023416519165
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 879.4862060546875 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.641 ref_kT = 2.560

	Predicted entropy: -7.769190415274352e-05

	Predicted free_energy: -0.05433845520019531

[DiffTRe] Epoch 38
	Epoch loss = 0.00123
	Gradient norm: 0.005107658915221691
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 999.3294067382812 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.641 ref_kT = 2.560

	Predicted entropy: -0.037748370319604874

	Predicted free_energy: 0.002416914561763406

[DiffTRe] Epoch 39
	Epoch loss = 0.00118
	Gradient norm: 0.004807199351489544
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.9767456054688 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.641 ref_kT = 2.560

	Predicted entropy: -0.010236669331789017

	Predicted free_energy: -0.04071906954050064

[DiffTRe] Epoch 40
	Epoch loss = 0.00116
	Gradient norm: 0.004814951680600643
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.0697631835938 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.641 ref_kT = 2.560

	Predicted entropy: -0.04986882582306862

	Predicted free_energy: 0.018742598593235016

[DiffTRe] Epoch 41
	Epoch loss = 0.00115
	Gradient norm: 0.006699834950268269
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.94189453125 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.641 ref_kT = 2.560

	Predicted entropy: -0.012247757986187935

	Predicted free_energy: -0.039558183401823044

[DiffTRe] Epoch 42
	Epoch loss = 0.00118
	Gradient norm: 0.009746949188411236
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 998.2335205078125 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.641 ref_kT = 2.560

	Predicted entropy: -0.07109180837869644

	Predicted free_energy: 0.04873771592974663

[DiffTRe] Epoch 43
	Epoch loss = 0.00127
	Gradient norm: 0.01914048008620739
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.6024780273438 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.641 ref_kT = 2.560

	Predicted entropy: 0.002744082361459732

	Predicted free_energy: -0.06449958682060242

[DiffTRe] Epoch 44
	Epoch loss = 0.00158
	Gradient norm: 0.03770803287625313
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 994.895751953125 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.641 ref_kT = 2.560

	Predicted entropy: -0.11799370497465134

	Predicted free_energy: 0.11634757369756699

[DiffTRe] Epoch 45
	Epoch loss = 0.00230
	Gradient norm: 0.10254605859518051
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 994.4688720703125 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.641 ref_kT = 2.560

	Predicted entropy: 0.06704913824796677

	Predicted free_energy: -0.1679029017686844

[DiffTRe] Epoch 46
	Epoch loss = 0.00488
	Gradient norm: 0.21772775053977966
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 981.2869873046875 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.467 ref_kT = 2.560

	Predicted entropy: -0.0007759998552501202

	Predicted free_energy: 0.257595956325531

[DiffTRe] Epoch 47
	Epoch loss = 0.01252
	Gradient norm: 1.148505449295044
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 851.3020629882812 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.630 ref_kT = 2.560

	Predicted entropy: 0.0006329991156235337

	Predicted free_energy: -0.43150144815444946

[DiffTRe] Epoch 48
	Epoch loss = 0.02127
	Gradient norm: 0.9290533065795898
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 892.1986083984375 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.584 ref_kT = 2.560

	Predicted entropy: -0.0004859749460592866

	Predicted free_energy: 0.06910556554794312

[DiffTRe] Epoch 49
	Epoch loss = 0.01238
	Gradient norm: 1.1244304180145264
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 889.9742431640625 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.566 ref_kT = 2.560

	Predicted entropy: 0.0006321467808447778

	Predicted free_energy: -0.412415087223053

[DiffTRe] Epoch 50
	Epoch loss = 0.02756
	Gradient norm: 0.8233168721199036
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 938.435791015625 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.643 ref_kT = 2.560

	Predicted entropy: -0.00023181094729807228

	Predicted free_energy: -0.002568572759628296

[DiffTRe] Epoch 51
	Epoch loss = 0.00050
	Gradient norm: 0.00021025640307925642
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 999.98486328125 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.643 ref_kT = 2.560

	Predicted entropy: -0.0034191906452178955

	Predicted free_energy: 0.002505890093743801

[DiffTRe] Epoch 52
	Epoch loss = 0.00049
	Gradient norm: 6.482613389380276e-05
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.9967651367188 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.643 ref_kT = 2.560

	Predicted entropy: 0.0024470267817378044

	Predicted free_energy: -0.005807098001241684

[DiffTRe] Epoch 53
	Epoch loss = 0.00049
	Gradient norm: 5.220895036472939e-05
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.9876708984375 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.643 ref_kT = 2.560

	Predicted entropy: 0.0017068276647478342

	Predicted free_energy: -0.00437155133113265

[DiffTRe] Epoch 54
	Epoch loss = 0.00048
	Gradient norm: 2.0184408640488982e-05
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.9867553710938 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.643 ref_kT = 2.560

	Predicted entropy: 0.006024559028446674

	Predicted free_energy: -0.010443328879773617

[DiffTRe] Epoch 55
	Epoch loss = 0.00048
	Gradient norm: 2.6890405933954753e-05
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.9781494140625 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.643 ref_kT = 2.560

	Predicted entropy: 0.005740659777075052

	Predicted free_energy: -0.009713348001241684

[DiffTRe] Epoch 56
	Epoch loss = 0.00047
	Gradient norm: 1.448838520445861e-05
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.9714965820312 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.643 ref_kT = 2.560

	Predicted entropy: 0.009665940888226032

	Predicted free_energy: -0.01525045745074749

[DiffTRe] Epoch 57
	Epoch loss = 0.00047
	Gradient norm: 2.580379441496916e-05
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.9638671875 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.643 ref_kT = 2.560

	Predicted entropy: 0.008961542509496212

	Predicted free_energy: -0.013928435742855072

[DiffTRe] Epoch 58
	Epoch loss = 0.00047
	Gradient norm: 2.229074016213417e-05
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.952392578125 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.643 ref_kT = 2.560

	Predicted entropy: 0.013315456919372082

	Predicted free_energy: -0.020150359719991684

[DiffTRe] Epoch 59
	Epoch loss = 0.00047
	Gradient norm: 4.5828997826902196e-05
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.9457397460938 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.643 ref_kT = 2.560

	Predicted entropy: 0.01114768348634243

	Predicted free_energy: -0.016683563590049744

[DiffTRe] Epoch 60
	Epoch loss = 0.00047
	Gradient norm: 6.545397627633065e-05
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.9266357421875 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.643 ref_kT = 2.560

	Predicted entropy: 0.01748388633131981

	Predicted free_energy: -0.025909636169672012

[DiffTRe] Epoch 61
	Epoch loss = 0.00047
	Gradient norm: 0.0001537134958198294
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.9247436523438 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.643 ref_kT = 2.560

	Predicted entropy: 0.011212676763534546

	Predicted free_energy: -0.016353972256183624

[DiffTRe] Epoch 62
	Epoch loss = 0.00047
	Gradient norm: 0.00033778988290578127
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.8623046875 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.643 ref_kT = 2.560

	Predicted entropy: 0.024388164281845093

	Predicted free_energy: -0.035854704678058624

[DiffTRe] Epoch 63
	Epoch loss = 0.00048
	Gradient norm: 0.0009556197910569608
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.8451538085938 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.643 ref_kT = 2.560

	Predicted entropy: 0.004586532711982727

	Predicted free_energy: -0.0061452328227460384

[DiffTRe] Epoch 64
	Epoch loss = 0.00051
	Gradient norm: 0.0029878204222768545
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.42041015625 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.643 ref_kT = 2.560

	Predicted entropy: 0.04400096833705902

	Predicted free_energy: -0.06510152667760849

[DiffTRe] Epoch 65
	Epoch loss = 0.00063
	Gradient norm: 0.010344978421926498
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 998.6172485351562 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.643 ref_kT = 2.560

	Predicted entropy: -0.030591662973165512

	Predicted free_energy: 0.04596414044499397

[DiffTRe] Epoch 66
	Epoch loss = 0.00113
	Gradient norm: 0.04490908607840538
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 992.188232421875 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.643 ref_kT = 2.560

	Predicted entropy: 0.12357451766729355

	Predicted free_energy: -0.18850606679916382

[DiffTRe] Epoch 67
	Epoch loss = 0.00343
	Gradient norm: 0.15572743117809296
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 982.6060180664062 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.000708793115336448

	Predicted free_energy: 0.22724926471710205

[DiffTRe] Epoch 68
	Epoch loss = 0.01345
	Gradient norm: 1.069933295249939
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 851.9806518554688 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.561 ref_kT = 2.560

	Predicted entropy: 0.0005983008886687458

	Predicted free_energy: -0.4753161668777466

[DiffTRe] Epoch 69
	Epoch loss = 0.03162
	Gradient norm: 0.9131786227226257
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 905.2389526367188 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.386 ref_kT = 2.560

	Predicted entropy: -0.0004977317294105887

	Predicted free_energy: 0.07365459203720093

[DiffTRe] Epoch 70
	Epoch loss = 0.00270
	Gradient norm: 0.006978568620979786
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 999.4137573242188 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.386 ref_kT = 2.560

	Predicted entropy: 0.01924058049917221

	Predicted free_energy: 0.039653126150369644

[DiffTRe] Epoch 71
	Epoch loss = 0.00248
	Gradient norm: 0.0069139874540269375
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.591552734375 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.386 ref_kT = 2.560

	Predicted entropy: -0.05013740807771683

	Predicted free_energy: 0.13077738881111145

[DiffTRe] Epoch 72
	Epoch loss = 0.00237
	Gradient norm: 0.0132637619972229
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 998.8587036132812 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.386 ref_kT = 2.560

	Predicted entropy: 0.002707958687096834

	Predicted free_energy: 0.053364064544439316

[DiffTRe] Epoch 73
	Epoch loss = 0.00241
	Gradient norm: 0.022630654275417328
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 997.9556274414062 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.386 ref_kT = 2.560

	Predicted entropy: -0.10783547163009644

	Predicted free_energy: 0.20228251814842224

[DiffTRe] Epoch 74
	Epoch loss = 0.00269
	Gradient norm: 0.058984752744436264
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 995.4827880859375 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.386 ref_kT = 2.560

	Predicted entropy: 0.03716640919446945

	Predicted free_energy: -0.0014833457535132766

[DiffTRe] Epoch 75
	Epoch loss = 0.00392
	Gradient norm: 0.1290608048439026
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 990.4382934570312 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.386 ref_kT = 2.560

	Predicted entropy: -0.21151745319366455

	Predicted free_energy: 0.3371519148349762

[DiffTRe] Epoch 76
	Epoch loss = 0.00626
	Gradient norm: 0.4127722680568695
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 974.835693359375 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.655 ref_kT = 2.560

	Predicted entropy: 2.5361316602356965e-06

	Predicted free_energy: -0.1660287082195282

[DiffTRe] Epoch 77
	Epoch loss = 0.00673
	Gradient norm: 0.2840656340122223
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 958.5482177734375 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.627 ref_kT = 2.560

	Predicted entropy: -0.0007890472188591957

	Predicted free_energy: 0.18986856937408447

[DiffTRe] Epoch 78
	Epoch loss = 0.01304
	Gradient norm: 1.1984845399856567
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 897.6869506835938 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.593 ref_kT = 2.560

	Predicted entropy: 0.00023620552383363247

	Predicted free_energy: -0.3025163412094116

[DiffTRe] Epoch 79
	Epoch loss = 0.02211
	Gradient norm: 0.7880833148956299
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 940.1749877929688 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.629 ref_kT = 2.560

	Predicted entropy: -0.00044712002272717655

	Predicted free_energy: 0.07185512781143188

[DiffTRe] Epoch 80
	Epoch loss = 0.00372
	Gradient norm: 0.0022368214558809996
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 999.830810546875 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.629 ref_kT = 2.560

	Predicted entropy: 0.0027388674207031727

	Predicted free_energy: 0.061418116092681885

[DiffTRe] Epoch 81
	Epoch loss = 0.00351
	Gradient norm: 0.00397969176992774
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.8384399414062 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.629 ref_kT = 2.560

	Predicted entropy: -0.025698846206068993

	Predicted free_energy: 0.10935024172067642

[DiffTRe] Epoch 82
	Epoch loss = 0.00335
	Gradient norm: 0.0070315636694431305
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.5834350585938 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.629 ref_kT = 2.560

	Predicted entropy: -0.007896918803453445

	Predicted free_energy: 0.07290004938840866

[DiffTRe] Epoch 83
	Epoch loss = 0.00327
	Gradient norm: 0.016036804765462875
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 998.8839721679688 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.629 ref_kT = 2.560

	Predicted entropy: -0.059344951063394547

	Predicted free_energy: 0.16193203628063202

[DiffTRe] Epoch 84
	Epoch loss = 0.00338
	Gradient norm: 0.04403276741504669
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 997.6021118164062 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.629 ref_kT = 2.560

	Predicted entropy: 0.005380378104746342

	Predicted free_energy: 0.040251124650239944

[DiffTRe] Epoch 85
	Epoch loss = 0.00419
	Gradient norm: 0.11617767810821533
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 992.541259765625 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.629 ref_kT = 2.560

	Predicted entropy: -0.12846893072128296

	Predicted free_energy: 0.27276697754859924

[DiffTRe] Epoch 86
	Epoch loss = 0.00646
	Gradient norm: 0.45049411058425903
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 977.866943359375 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.653 ref_kT = 2.560

	Predicted entropy: -9.087033686228096e-05

	Predicted free_energy: -0.12617039680480957

[DiffTRe] Epoch 87
	Epoch loss = 0.01756
	Gradient norm: 0.6274001002311707
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 942.55224609375 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.567 ref_kT = 2.560

	Predicted entropy: -0.0011716231238096952

	Predicted free_energy: 0.3283880054950714

[DiffTRe] Epoch 88
	Epoch loss = 0.00493
	Gradient norm: 0.3859088122844696
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 961.9190673828125 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.534 ref_kT = 2.560

	Predicted entropy: -0.0003509364614728838

	Predicted free_energy: -0.0441569983959198

[DiffTRe] Epoch 89
	Epoch loss = 0.01495
	Gradient norm: 0.5593915581703186
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 957.136474609375 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.534 ref_kT = 2.560

	Predicted entropy: -0.0012761804973706603

	Predicted free_energy: 0.3797266185283661

[DiffTRe] Epoch 90
	Epoch loss = 0.00516
	Gradient norm: 0.41053280234336853
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 965.5310668945312 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.667 ref_kT = 2.560

	Predicted entropy: -0.0005038752569817007

	Predicted free_energy: 0.04445579648017883

[DiffTRe] Epoch 91
	Epoch loss = 0.00354
	Gradient norm: 0.10976412892341614
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 991.0543212890625 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.667 ref_kT = 2.560

	Predicted entropy: -0.09674376249313354

	Predicted free_energy: 0.19781029224395752

[DiffTRe] Epoch 92
	Epoch loss = 0.00220
	Gradient norm: 0.035113152116537094
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 998.5315551757812 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.667 ref_kT = 2.560

	Predicted entropy: -0.03581651672720909

	Predicted free_energy: 0.1022963747382164

[DiffTRe] Epoch 93
	Epoch loss = 0.00204
	Gradient norm: 0.019718797877430916
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 993.34423828125 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.667 ref_kT = 2.560

	Predicted entropy: -0.07982731610536575

	Predicted free_energy: 0.17262107133865356

[DiffTRe] Epoch 94
	Epoch loss = 0.00190
	Gradient norm: 0.011989914812147617
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 997.89990234375 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.667 ref_kT = 2.560

	Predicted entropy: -0.04084757715463638

	Predicted free_energy: 0.11159202456474304

[DiffTRe] Epoch 95
	Epoch loss = 0.00188
	Gradient norm: 0.009304608218371868
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 993.90380859375 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.667 ref_kT = 2.560

	Predicted entropy: -0.07348418980836868

	Predicted free_energy: 0.16389304399490356

[DiffTRe] Epoch 96
	Epoch loss = 0.00185
	Gradient norm: 0.008514118380844593
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 997.9432373046875 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.667 ref_kT = 2.560

	Predicted entropy: -0.03786601126194

	Predicted free_energy: 0.10804588347673416

[DiffTRe] Epoch 97
	Epoch loss = 0.00187
	Gradient norm: 0.009126314893364906
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 993.5830078125 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.667 ref_kT = 2.560

	Predicted entropy: -0.07326561957597733

	Predicted free_energy: 0.16471946239471436

[DiffTRe] Epoch 98
	Epoch loss = 0.00189
	Gradient norm: 0.011863499879837036
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 998.5301513671875 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.667 ref_kT = 2.560

	Predicted entropy: -0.028348784893751144

	Predicted free_energy: 0.09398093819618225

[DiffTRe] Epoch 99
	Epoch loss = 0.00198
	Gradient norm: 0.016941577196121216
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 991.9592895507812 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.667 ref_kT = 2.560

	Predicted entropy: -0.08136757463216782

	Predicted free_energy: 0.17858542501926422

[DiffTRe] Epoch 100
	Epoch loss = 0.00211
	Gradient norm: 0.030180132016539574
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.6425170898438 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.667 ref_kT = 2.560

	Predicted entropy: -0.006113278213888407

	Predicted free_energy: 0.059430159628391266

[DiffTRe] Epoch 101
	Epoch loss = 0.00254
	Gradient norm: 0.052465252578258514
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 987.0103149414062 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.523 ref_kT = 2.560

	Predicted entropy: -0.000854903890285641

	Predicted free_energy: 0.21242593228816986

[DiffTRe] Epoch 102
	Epoch loss = 0.00597
	Gradient norm: 0.4833478033542633
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 934.8814697265625 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.800 ref_kT = 2.560

	Predicted entropy: 1.5202981558104511e-05

	Predicted free_energy: -0.20237572491168976

[DiffTRe] Epoch 103
	Epoch loss = 0.02687
	Gradient norm: 0.9303874373435974
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 927.0706787109375 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.729 ref_kT = 2.560

	Predicted entropy: -0.001077306573279202

	Predicted free_energy: 0.29630619287490845

[DiffTRe] Epoch 104
	Epoch loss = 0.00738
	Gradient norm: 0.6481943726539612
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 950.4762573242188 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.578 ref_kT = 2.560

	Predicted entropy: -0.00021094133262522519

	Predicted free_energy: -0.07202541828155518

[DiffTRe] Epoch 105
	Epoch loss = 0.00670
	Gradient norm: 0.3189665377140045
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 976.646484375 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.649 ref_kT = 2.560

	Predicted entropy: -0.0008176940609700978

	Predicted free_energy: 0.16878333687782288

[DiffTRe] Epoch 106
	Epoch loss = 0.00416
	Gradient norm: 0.15410971641540527
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 987.5771484375 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.500 ref_kT = 2.560

	Predicted entropy: -0.0003917607828043401

	Predicted free_energy: -0.00771719217300415

[DiffTRe] Epoch 107
	Epoch loss = 0.00355
	Gradient norm: 0.17004017531871796
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 986.6273193359375 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.621 ref_kT = 2.560

	Predicted entropy: -0.0008471727487631142

	Predicted free_energy: 0.18955807387828827

[DiffTRe] Epoch 108
	Epoch loss = 0.00075
	Gradient norm: 0.021247977390885353
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 998.1574096679688 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.621 ref_kT = 2.560

	Predicted entropy: 0.05703233927488327

	Predicted free_energy: 0.11516354233026505

[DiffTRe] Epoch 109
	Epoch loss = 0.00067
	Gradient norm: 0.013275409117341042
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.9519653320312 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.621 ref_kT = 2.560

	Predicted entropy: 0.005904989317059517

	Predicted free_energy: 0.18009640276432037

[DiffTRe] Epoch 110
	Epoch loss = 0.00060
	Gradient norm: 0.009649951010942459
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 998.5382080078125 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.621 ref_kT = 2.560

	Predicted entropy: 0.047957707196474075

	Predicted free_energy: 0.12596432864665985

[DiffTRe] Epoch 111
	Epoch loss = 0.00060
	Gradient norm: 0.008517008274793625
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.9505004882812 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.621 ref_kT = 2.560

	Predicted entropy: 0.0032609698828309774

	Predicted free_energy: 0.18270382285118103

[DiffTRe] Epoch 112
	Epoch loss = 0.00059
	Gradient norm: 0.009078269824385643
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 998.400634765625 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.621 ref_kT = 2.560

	Predicted entropy: 0.04802253469824791

	Predicted free_energy: 0.12517696619033813

[DiffTRe] Epoch 113
	Epoch loss = 0.00063
	Gradient norm: 0.011099657043814659
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.9881591796875 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.621 ref_kT = 2.560

	Predicted entropy: -0.006807331927120686

	Predicted free_energy: 0.19490598142147064

[DiffTRe] Epoch 114
	Epoch loss = 0.00068
	Gradient norm: 0.01656450890004635
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 997.5103149414062 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.621 ref_kT = 2.560

	Predicted entropy: 0.059673011302948

	Predicted free_energy: 0.10970821976661682

[DiffTRe] Epoch 115
	Epoch loss = 0.00084
	Gradient norm: 0.026600630953907967
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.7607421875 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.621 ref_kT = 2.560

	Predicted entropy: -0.030425719916820526

	Predicted free_energy: 0.2245202362537384

[DiffTRe] Epoch 116
	Epoch loss = 0.00110
	Gradient norm: 0.052932411432266235
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 993.9421997070312 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.621 ref_kT = 2.560

	Predicted entropy: 0.09637924283742905

	Predicted free_energy: 0.06236935779452324

[DiffTRe] Epoch 117
	Epoch loss = 0.00187
	Gradient norm: 0.0968426764011383
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 997.7095947265625 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.621 ref_kT = 2.560

	Predicted entropy: -0.07742184400558472

	Predicted free_energy: 0.28411129117012024

[DiffTRe] Epoch 118
	Epoch loss = 0.00281
	Gradient norm: 0.21387746930122375
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 983.3189086914062 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.574 ref_kT = 2.560

	Predicted entropy: -0.00030420804978348315

	Predicted free_energy: -0.008842960000038147

[DiffTRe] Epoch 119
	Epoch loss = 0.00147
	Gradient norm: 0.04154118523001671
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 993.6275024414062 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.574 ref_kT = 2.560

	Predicted entropy: -0.09607196599245071

	Predicted free_energy: 0.09480083733797073

[DiffTRe] Epoch 120
	Epoch loss = 0.00165
	Gradient norm: 0.06469057500362396
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.390380859375 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.574 ref_kT = 2.560

	Predicted entropy: 0.03943949192762375

	Predicted free_energy: -0.050735048949718475

[DiffTRe] Epoch 121
	Epoch loss = 0.00227
	Gradient norm: 0.09470164030790329
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 990.24658203125 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.574 ref_kT = 2.560

	Predicted entropy: -0.11161117255687714

	Predicted free_energy: 0.11264019459486008

[DiffTRe] Epoch 122
	Epoch loss = 0.00265
	Gradient norm: 0.1536547690629959
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 996.795654296875 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.574 ref_kT = 2.560

	Predicted entropy: 0.08611160516738892

	Predicted free_energy: -0.1002601906657219

[DiffTRe] Epoch 123
	Epoch loss = 0.00390
	Gradient norm: 0.19200237095355988
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 988.4293823242188 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.529 ref_kT = 2.560

	Predicted entropy: -0.000660488847643137

	Predicted free_energy: 0.1423467993736267

[DiffTRe] Epoch 124
	Epoch loss = 0.00287
	Gradient norm: 0.06291582435369492
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 991.2367553710938 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.529 ref_kT = 2.560

	Predicted entropy: -0.1339389681816101

	Predicted free_energy: 0.3424298167228699

[DiffTRe] Epoch 125
	Epoch loss = 0.00309
	Gradient norm: 0.10912136733531952
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.4190063476562 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.529 ref_kT = 2.560

	Predicted entropy: 0.011486236937344074

	Predicted free_energy: 0.11440734565258026

[DiffTRe] Epoch 126
	Epoch loss = 0.00418
	Gradient norm: 0.15457680821418762
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 986.2265625 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.599 ref_kT = 2.560

	Predicted entropy: -0.0012881926959380507

	Predicted free_energy: 0.404424250125885

[DiffTRe] Epoch 127
	Epoch loss = 0.00596
	Gradient norm: 0.5780408382415771
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 945.673828125 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.654 ref_kT = 2.560

	Predicted entropy: -0.00026100155082531273

	Predicted free_energy: -0.025200754404067993

[DiffTRe] Epoch 128
	Epoch loss = 0.00795
	Gradient norm: 0.3816920518875122
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 969.9471435546875 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.574 ref_kT = 2.560

	Predicted entropy: -0.0009975760476663709

	Predicted free_energy: 0.28181976079940796

[DiffTRe] Epoch 129
	Epoch loss = 0.00272
	Gradient norm: 0.168165385723114
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 985.9646606445312 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.670 ref_kT = 2.560

	Predicted entropy: -0.0004271599173080176

	Predicted free_energy: 0.051613032817840576

[DiffTRe] Epoch 130
	Epoch loss = 0.00409
	Gradient norm: 0.09030678123235703
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 991.533203125 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.670 ref_kT = 2.560

	Predicted entropy: -0.1342727690935135

	Predicted free_energy: 0.23116014897823334

[DiffTRe] Epoch 131
	Epoch loss = 0.00363
	Gradient norm: 0.10060285776853561
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.84130859375 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.670 ref_kT = 2.560

	Predicted entropy: -0.028655074536800385

	Predicted free_energy: 0.08149950951337814

[DiffTRe] Epoch 132
	Epoch loss = 0.00394
	Gradient norm: 0.12228792905807495
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 988.062744140625 (990.0) -> Recompute is True

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.0010103689273819327

	Predicted free_energy: 0.29210013151168823

[DiffTRe] Epoch 133
	Epoch loss = 0.00161
	Gradient norm: 0.013772635720670223
	Elapsed time = 0.003 min

[Propagate] Effective sample size: 998.591552734375 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.06006990000605583

	Predicted free_energy: 0.37363943457603455

[DiffTRe] Epoch 134
	Epoch loss = 0.00159
	Gradient norm: 0.014867790974676609
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.993896484375 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.003046935424208641

	Predicted free_energy: 0.2931426167488098

[DiffTRe] Epoch 135
	Epoch loss = 0.00164
	Gradient norm: 0.017770439386367798
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 998.0189208984375 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.07479299604892731

	Predicted free_energy: 0.39228200912475586

[DiffTRe] Epoch 136
	Epoch loss = 0.00170
	Gradient norm: 0.026485273614525795
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.8451538085938 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: 0.00853012129664421

	Predicted free_energy: 0.27512502670288086

[DiffTRe] Epoch 137
	Epoch loss = 0.00191
	Gradient norm: 0.04054659232497215
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 996.2354125976562 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.10462861508131027

	Predicted free_energy: 0.4317583441734314

[DiffTRe] Epoch 138
	Epoch loss = 0.00220
	Gradient norm: 0.07709997147321701
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 998.4839477539062 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: 0.044426608830690384

	Predicted free_energy: 0.22270315885543823

[DiffTRe] Epoch 139
	Epoch loss = 0.00306
	Gradient norm: 0.12559017539024353
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 992.2677001953125 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.14975132048130035

	Predicted free_energy: 0.4919780492782593

[DiffTRe] Epoch 140
	Epoch loss = 0.00364
	Gradient norm: 0.23000741004943848
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 994.8473510742188 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: 0.09009312838315964

	Predicted free_energy: 0.15636038780212402

[DiffTRe] Epoch 141
	Epoch loss = 0.00528
	Gradient norm: 0.27070412039756775
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 991.8187866210938 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.1549019068479538

	Predicted free_energy: 0.4984746277332306

[DiffTRe] Epoch 142
	Epoch loss = 0.00379
	Gradient norm: 0.2477351874113083
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 996.6925048828125 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: 0.06919284909963608

	Predicted free_energy: 0.18629935383796692

[DiffTRe] Epoch 143
	Epoch loss = 0.00422
	Gradient norm: 0.2045501321554184
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 994.564697265625 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.12898997962474823

	Predicted free_energy: 0.4629265367984772

[DiffTRe] Epoch 144
	Epoch loss = 0.00270
	Gradient norm: 0.1304692178964615
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.0144653320312 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: 0.029103998094797134

	Predicted free_energy: 0.24292288720607758

[DiffTRe] Epoch 145
	Epoch loss = 0.00269
	Gradient norm: 0.10080568492412567
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 996.0473022460938 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.11353899538516998

	Predicted free_energy: 0.44101858139038086

[DiffTRe] Epoch 146
	Epoch loss = 0.00215
	Gradient norm: 0.07460283488035202
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.5953369140625 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: 0.00872277282178402

	Predicted free_energy: 0.270801305770874

[DiffTRe] Epoch 147
	Epoch loss = 0.00218
	Gradient norm: 0.06423304229974747
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 996.4454345703125 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.11049360781908035

	Predicted free_energy: 0.43586844205856323

[DiffTRe] Epoch 148
	Epoch loss = 0.00198
	Gradient norm: 0.05907338485121727
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.6563720703125 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: 0.0027441363781690598

	Predicted free_energy: 0.2782244086265564

[DiffTRe] Epoch 149
	Epoch loss = 0.00211
	Gradient norm: 0.05925573781132698
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 996.1342163085938 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.11658642441034317

	Predicted free_energy: 0.44325003027915955

[DiffTRe] Epoch 150
	Epoch loss = 0.00205
	Gradient norm: 0.06586708128452301
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.469970703125 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: 0.007502784952521324

	Predicted free_energy: 0.270618200302124

[DiffTRe] Epoch 151
	Epoch loss = 0.00230
	Gradient norm: 0.07432467490434647
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 995.25634765625 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.1290406882762909

	Predicted free_energy: 0.4594341218471527

[DiffTRe] Epoch 152
	Epoch loss = 0.00230
	Gradient norm: 0.09178338199853897
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 998.998779296875 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: 0.019524360075592995

	Predicted free_energy: 0.2529521882534027

[DiffTRe] Epoch 153
	Epoch loss = 0.00273
	Gradient norm: 0.10584801435470581
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 994.251708984375 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.1414545625448227

	Predicted free_energy: 0.47574636340141296

[DiffTRe] Epoch 154
	Epoch loss = 0.00262
	Gradient norm: 0.12512829899787903
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 998.5386962890625 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: 0.028766421601176262

	Predicted free_energy: 0.2394682914018631

[DiffTRe] Epoch 155
	Epoch loss = 0.00309
	Gradient norm: 0.1319637894630432
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 994.06591796875 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.14408273994922638

	Predicted free_energy: 0.4790007472038269

[DiffTRe] Epoch 156
	Epoch loss = 0.00267
	Gradient norm: 0.13059812784194946
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 998.7139282226562 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: 0.024477796629071236

	Predicted free_energy: 0.24534597992897034

[DiffTRe] Epoch 157
	Epoch loss = 0.00296
	Gradient norm: 0.12273899465799332
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 994.7145385742188 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.13758839666843414

	Predicted free_energy: 0.46994924545288086

[DiffTRe] Epoch 158
	Epoch loss = 0.00243
	Gradient norm: 0.10621116310358047
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.1793212890625 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: 0.012145462445914745

	Predicted free_energy: 0.262543261051178

[DiffTRe] Epoch 159
	Epoch loss = 0.00258
	Gradient norm: 0.09523043036460876
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 995.3787841796875 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.13062992691993713

	Predicted free_energy: 0.4601689875125885

[DiffTRe] Epoch 160
	Epoch loss = 0.00220
	Gradient norm: 0.08255302906036377
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.4490356445312 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: 0.0021159606985747814

	Predicted free_energy: 0.2763298749923706

[DiffTRe] Epoch 161
	Epoch loss = 0.00232
	Gradient norm: 0.07612398266792297
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 995.6887817382812 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.12766364216804504

	Predicted free_energy: 0.4557439386844635

[DiffTRe] Epoch 162
	Epoch loss = 0.00209
	Gradient norm: 0.07128884643316269
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.5223999023438 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.0022599452640861273

	Predicted free_energy: 0.2820964753627777

[DiffTRe] Epoch 163
	Epoch loss = 0.00223
	Gradient norm: 0.06993239372968674
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 995.64794921875 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.1289641112089157

	Predicted free_energy: 0.45713308453559875

[DiffTRe] Epoch 164
	Epoch loss = 0.00208
	Gradient norm: 0.07128601521253586
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.4752197265625 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.0015614884905517101

	Predicted free_energy: 0.2807634472846985

[DiffTRe] Epoch 165
	Epoch loss = 0.00228
	Gradient norm: 0.0736842229962349
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 995.37451171875 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.13292697072029114

	Predicted free_energy: 0.4622087776660919

[DiffTRe] Epoch 166
	Epoch loss = 0.00216
	Gradient norm: 0.0787559300661087
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.3632202148438 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: 0.0018163876375183463

	Predicted free_energy: 0.2757500410079956

[DiffTRe] Epoch 167
	Epoch loss = 0.00240
	Gradient norm: 0.08250429481267929
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 995.0926513671875 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.1367053985595703

	Predicted free_energy: 0.4671257734298706

[DiffTRe] Epoch 168
	Epoch loss = 0.00224
	Gradient norm: 0.08704885095357895
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.2798461914062 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: 0.004121461417526007

	Predicted free_energy: 0.27234548330307007

[DiffTRe] Epoch 169
	Epoch loss = 0.00248
	Gradient norm: 0.08862358331680298
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 995.0409545898438 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.1375756710767746

	Predicted free_energy: 0.4681755602359772

[DiffTRe] Epoch 170
	Epoch loss = 0.00225
	Gradient norm: 0.08822562545537949
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.3156127929688 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: 0.0025900225155055523

	Predicted free_energy: 0.2744036018848419

[DiffTRe] Epoch 171
	Epoch loss = 0.00244
	Gradient norm: 0.08601081371307373
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 995.2459106445312 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.1353892683982849

	Predicted free_energy: 0.46510428190231323

[DiffTRe] Epoch 172
	Epoch loss = 0.00218
	Gradient norm: 0.08121155947446823
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.4275512695312 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.001898693386465311

	Predicted free_energy: 0.28061941266059875

[DiffTRe] Epoch 173
	Epoch loss = 0.00232
	Gradient norm: 0.07723169773817062
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 995.5226440429688 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.13231050968170166

	Predicted free_energy: 0.46079885959625244

[DiffTRe] Epoch 174
	Epoch loss = 0.00209
	Gradient norm: 0.07207505404949188
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.5223999023438 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.006480006501078606

	Predicted free_energy: 0.2869255840778351

[DiffTRe] Epoch 175
	Epoch loss = 0.00221
	Gradient norm: 0.06889086216688156
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 995.7096557617188 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.13030262291431427

	Predicted free_energy: 0.45792287588119507

[DiffTRe] Epoch 176
	Epoch loss = 0.00203
	Gradient norm: 0.06588432937860489
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.5667724609375 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.00925377756357193

	Predicted free_energy: 0.2906694710254669

[DiffTRe] Epoch 177
	Epoch loss = 0.00215
	Gradient norm: 0.06449592113494873
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 995.7581176757812 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.12999287247657776

	Predicted free_energy: 0.45735159516334534

[DiffTRe] Epoch 178
	Epoch loss = 0.00201
	Gradient norm: 0.0639176070690155
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.570556640625 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.009955634362995625

	Predicted free_energy: 0.29151174426078796

[DiffTRe] Epoch 179
	Epoch loss = 0.00215
	Gradient norm: 0.06407351791858673
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 995.708740234375 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.1308962106704712

	Predicted free_energy: 0.45844900608062744

[DiffTRe] Epoch 180
	Epoch loss = 0.00202
	Gradient norm: 0.06499619036912918
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.5519409179688 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.009471110999584198

	Predicted free_energy: 0.2907158434391022

[DiffTRe] Epoch 181
	Epoch loss = 0.00217
	Gradient norm: 0.0657447800040245
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 995.6427612304688 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.13191229104995728

	Predicted free_energy: 0.45973318815231323

[DiffTRe] Epoch 182
	Epoch loss = 0.00203
	Gradient norm: 0.06663332879543304
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.5391235351562 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.009192532859742641

	Predicted free_energy: 0.2902434468269348

[DiffTRe] Epoch 183
	Epoch loss = 0.00218
	Gradient norm: 0.06681258976459503
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 995.6431884765625 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.13203644752502441

	Predicted free_energy: 0.4598320722579956

[DiffTRe] Epoch 184
	Epoch loss = 0.00203
	Gradient norm: 0.06635323911905289
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.5543212890625 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.010136627592146397

	Predicted free_energy: 0.2915080785751343

[DiffTRe] Epoch 185
	Epoch loss = 0.00216
	Gradient norm: 0.06538275629281998
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 995.7329711914062 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.13103412091732025

	Predicted free_energy: 0.45841237902641296

[DiffTRe] Epoch 186
	Epoch loss = 0.00200
	Gradient norm: 0.06349285691976547
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.5867919921875 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.012236570939421654

	Predicted free_energy: 0.2943950593471527

[DiffTRe] Epoch 187
	Epoch loss = 0.00211
	Gradient norm: 0.061770785599946976
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 995.864501953125 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.1294952929019928

	Predicted free_energy: 0.4562578499317169

[DiffTRe] Epoch 188
	Epoch loss = 0.00196
	Gradient norm: 0.05944215506315231
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.62109375 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.014624198898673058

	Predicted free_energy: 0.29767507314682007

[DiffTRe] Epoch 189
	Epoch loss = 0.00206
	Gradient norm: 0.05777949094772339
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 995.97607421875 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.12819154560565948

	Predicted free_energy: 0.45441335439682007

[DiffTRe] Epoch 190
	Epoch loss = 0.00193
	Gradient norm: 0.0559820681810379
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.6439819335938 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.016479847952723503

	Predicted free_energy: 0.3001995086669922

[DiffTRe] Epoch 191
	Epoch loss = 0.00202
	Gradient norm: 0.05489754304289818
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 996.0387573242188 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.12752492725849152

	Predicted free_energy: 0.4534319043159485

[DiffTRe] Epoch 192
	Epoch loss = 0.00191
	Gradient norm: 0.053971681743860245
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.6525268554688 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.017519673332571983

	Predicted free_energy: 0.30158132314682007

[DiffTRe] Epoch 193
	Epoch loss = 0.00201
	Gradient norm: 0.05350435525178909
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 996.0582275390625 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.1274145245552063

	Predicted free_energy: 0.4532122015953064

[DiffTRe] Epoch 194
	Epoch loss = 0.00190
	Gradient norm: 0.053220320492982864
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.6553955078125 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.01799958571791649

	Predicted free_energy: 0.302185595035553

[DiffTRe] Epoch 195
	Epoch loss = 0.00200
	Gradient norm: 0.05305977538228035
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 996.06298828125 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.1274636834859848

	Predicted free_energy: 0.4532195031642914

[DiffTRe] Epoch 196
	Epoch loss = 0.00190
	Gradient norm: 0.05288906395435333
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.6572875976562 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.018439486622810364

	Predicted free_energy: 0.30274587869644165

[DiffTRe] Epoch 197
	Epoch loss = 0.00199
	Gradient norm: 0.05261729285120964
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 996.0872192382812 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.12725217640399933

	Predicted free_energy: 0.4528825879096985

[DiffTRe] Epoch 198
	Epoch loss = 0.00189
	Gradient norm: 0.05208341032266617
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.6640014648438 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.019268721342086792

	Predicted free_energy: 0.3038603961467743

[DiffTRe] Epoch 199
	Epoch loss = 0.00198
	Gradient norm: 0.051450081169605255
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 996.141357421875 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.1266137808561325

	Predicted free_energy: 0.4519707262516022

[DiffTRe] Epoch 200
	Epoch loss = 0.00187
	Gradient norm: 0.05043390765786171
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.6759033203125 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.020556610077619553

	Predicted free_energy: 0.30561819672584534

[DiffTRe] Epoch 201
	Epoch loss = 0.00195
	Gradient norm: 0.049498699605464935
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 996.2169189453125 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.12569758296012878

	Predicted free_energy: 0.4506804645061493

[DiffTRe] Epoch 202
	Epoch loss = 0.00185
	Gradient norm: 0.04824952408671379
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.6878051757812 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.022039173170924187

	Predicted free_energy: 0.3076457977294922

[DiffTRe] Epoch 203
	Epoch loss = 0.00193
	Gradient norm: 0.047267839312553406
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 996.291015625 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.12480000406503677

	Predicted free_energy: 0.44941213726997375

[DiffTRe] Epoch 204
	Epoch loss = 0.00183
	Gradient norm: 0.046131931245326996
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.6978149414062 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.023380788043141365

	Predicted free_energy: 0.3094731867313385

[DiffTRe] Epoch 205
	Epoch loss = 0.00190
	Gradient norm: 0.04532783105969429
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 996.3470458984375 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.12412836402654648

	Predicted free_energy: 0.4484502375125885

[DiffTRe] Epoch 206
	Epoch loss = 0.00181
	Gradient norm: 0.04449610039591789
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.7054443359375 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.02440648339688778

	Predicted free_energy: 0.3108574450016022

[DiffTRe] Epoch 207
	Epoch loss = 0.00188
	Gradient norm: 0.04393539950251579
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 996.3836669921875 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.12370893359184265

	Predicted free_energy: 0.44783255457878113

[DiffTRe] Epoch 208
	Epoch loss = 0.00180
	Gradient norm: 0.043379418551921844
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.7078247070312 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.025161970406770706

	Predicted free_energy: 0.31186696887016296

[DiffTRe] Epoch 209
	Epoch loss = 0.00187
	Gradient norm: 0.04297623783349991
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 996.4130859375 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.12340706586837769

	Predicted free_energy: 0.4473784565925598

[DiffTRe] Epoch 210
	Epoch loss = 0.00179
	Gradient norm: 0.04252403974533081
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.710205078125 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.025839418172836304

	Predicted free_energy: 0.3127715289592743

[DiffTRe] Epoch 211
	Epoch loss = 0.00186
	Gradient norm: 0.04212641716003418
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 996.4444580078125 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.12304447591304779

	Predicted free_energy: 0.44684621691703796

[DiffTRe] Epoch 212
	Epoch loss = 0.00178
	Gradient norm: 0.041593968868255615
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.712158203125 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.026629094034433365

	Predicted free_energy: 0.3138359785079956

[DiffTRe] Epoch 213
	Epoch loss = 0.00185
	Gradient norm: 0.04109013453125954
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 996.4853515625 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.12252261489629745

	Predicted free_energy: 0.4461003839969635

[DiffTRe] Epoch 214
	Epoch loss = 0.00177
	Gradient norm: 0.04040052369236946
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.7164306640625 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.027593348175287247

	Predicted free_energy: 0.3151470124721527

[DiffTRe] Epoch 215
	Epoch loss = 0.00183
	Gradient norm: 0.03978240489959717
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 996.53759765625 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.12186888605356216

	Predicted free_energy: 0.44517508149147034

[DiffTRe] Epoch 216
	Epoch loss = 0.00175
	Gradient norm: 0.03899119794368744
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.720703125 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.02866331860423088

	Predicted free_energy: 0.31660452485084534

[DiffTRe] Epoch 217
	Epoch loss = 0.00181
	Gradient norm: 0.038336195051670074
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 996.58984375 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.12119050323963165

	Predicted free_energy: 0.4442155957221985

[DiffTRe] Epoch 218
	Epoch loss = 0.00174
	Gradient norm: 0.03756209835410118
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.7230834960938 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.02970162220299244

	Predicted free_energy: 0.3180181086063385

[DiffTRe] Epoch 219
	Epoch loss = 0.00179
	Gradient norm: 0.036960307508707047
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 996.6373901367188 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.12058760970830917

	Predicted free_energy: 0.4433586597442627

[DiffTRe] Epoch 220
	Epoch loss = 0.00173
	Gradient norm: 0.03629249706864357
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.7254638671875 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.03061837889254093

	Predicted free_energy: 0.31926077604293823

[DiffTRe] Epoch 221
	Epoch loss = 0.00178
	Gradient norm: 0.035787299275398254
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 996.6759033203125 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.12009628862142563

	Predicted free_energy: 0.4426543414592743

[DiffTRe] Epoch 222
	Epoch loss = 0.00172
	Gradient norm: 0.035239361226558685
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.7259521484375 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.031404539942741394

	Predicted free_energy: 0.32032278180122375

[DiffTRe] Epoch 223
	Epoch loss = 0.00177
	Gradient norm: 0.03481350839138031
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 996.7105712890625 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.11967642605304718

	Predicted free_energy: 0.4420488476753235

[DiffTRe] Epoch 224
	Epoch loss = 0.00171
	Gradient norm: 0.03433310240507126
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.7249755859375 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.03212727606296539

	Predicted free_energy: 0.32129812240600586

[DiffTRe] Epoch 225
	Epoch loss = 0.00176
	Gradient norm: 0.03393297642469406
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 996.742919921875 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.11925564706325531

	Predicted free_energy: 0.44144460558891296

[DiffTRe] Epoch 226
	Epoch loss = 0.00170
	Gradient norm: 0.03345019370317459
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.7249755859375 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.03286939114332199

	Predicted free_energy: 0.3223015367984772

[DiffTRe] Epoch 227
	Epoch loss = 0.00174
	Gradient norm: 0.03302714601159096
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 996.7808837890625 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.1187814325094223

	Predicted free_energy: 0.4407683312892914

[DiffTRe] Epoch 228
	Epoch loss = 0.00169
	Gradient norm: 0.032500579953193665
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.7254638671875 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.033672064542770386

	Predicted free_energy: 0.3233916461467743

[DiffTRe] Epoch 229
	Epoch loss = 0.00173
	Gradient norm: 0.032039444893598557
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 996.8198852539062 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.11824668943881989

	Predicted free_energy: 0.44001150131225586

[DiffTRe] Epoch 230
	Epoch loss = 0.00168
	Gradient norm: 0.03147547319531441
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.7240600585938 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.03452304005622864

	Predicted free_energy: 0.3245488703250885

[DiffTRe] Epoch 231
	Epoch loss = 0.00172
	Gradient norm: 0.030996764078736305
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 996.8612670898438 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.11769098788499832

	Predicted free_energy: 0.43922537565231323

[DiffTRe] Epoch 232
	Epoch loss = 0.00167
	Gradient norm: 0.030434345826506615
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.7230834960938 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.03537020832300186

	Predicted free_energy: 0.32570120692253113

[DiffTRe] Epoch 233
	Epoch loss = 0.00171
	Gradient norm: 0.02997283823788166
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 996.8997192382812 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.11715710163116455

	Predicted free_energy: 0.4384697675704956

[DiffTRe] Epoch 234
	Epoch loss = 0.00166
	Gradient norm: 0.029450107365846634
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.72119140625 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.036168210208415985

	Predicted free_energy: 0.3267851769924164

[DiffTRe] Epoch 235
	Epoch loss = 0.00169
	Gradient norm: 0.02902863919734955
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 996.9368286132812 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.11667381227016449

	Predicted free_energy: 0.43778249621391296

[DiffTRe] Epoch 236
	Epoch loss = 0.00165
	Gradient norm: 0.028559286147356033
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.7178344726562 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.03690410032868385

	Predicted free_energy: 0.3277825117111206

[DiffTRe] Epoch 237
	Epoch loss = 0.00168
	Gradient norm: 0.02817797288298607
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 996.9686889648438 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.11622802168130875

	Predicted free_energy: 0.43714895844459534

[DiffTRe] Epoch 238
	Epoch loss = 0.00164
	Gradient norm: 0.027749083936214447
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.7164306640625 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.03759602829813957

	Predicted free_energy: 0.32872000336647034

[DiffTRe] Epoch 239
	Epoch loss = 0.00167
	Gradient norm: 0.02739047072827816
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 997.0005493164062 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.1157974973320961

	Predicted free_energy: 0.43653616309165955

[DiffTRe] Epoch 240
	Epoch loss = 0.00163
	Gradient norm: 0.0269757192581892
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.7135620117188 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.03827947378158569

	Predicted free_energy: 0.3296465277671814

[DiffTRe] Epoch 241
	Epoch loss = 0.00166
	Gradient norm: 0.026619311422109604
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 997.0338134765625 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.1153503954410553

	Predicted free_energy: 0.4359026253223419

[DiffTRe] Epoch 242
	Epoch loss = 0.00162
	Gradient norm: 0.02619890123605728
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.710693359375 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.038977738469839096

	Predicted free_energy: 0.3305950164794922

[DiffTRe] Epoch 243
	Epoch loss = 0.00165
	Gradient norm: 0.02583378739655018
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 997.066162109375 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.11488210409879684

	Predicted free_energy: 0.4352397620677948

[DiffTRe] Epoch 244
	Epoch loss = 0.00161
	Gradient norm: 0.025404104962944984
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.7073364257812 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.039694249629974365

	Predicted free_energy: 0.3315691351890564

[DiffTRe] Epoch 245
	Epoch loss = 0.00164
	Gradient norm: 0.02503449283540249
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 997.100341796875 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.11439895629882812

	Predicted free_energy: 0.434558629989624

[DiffTRe] Epoch 246
	Epoch loss = 0.00161
	Gradient norm: 0.024607257917523384
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.7025756835938 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.04040977731347084

	Predicted free_energy: 0.3325420320034027

[DiffTRe] Epoch 247
	Epoch loss = 0.00163
	Gradient norm: 0.024245919659733772
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 997.1341552734375 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.11392441391944885

	Predicted free_energy: 0.4338884651660919

[DiffTRe] Epoch 248
	Epoch loss = 0.00160
	Gradient norm: 0.023836011067032814
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.6983032226562 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.041103996336460114

	Predicted free_energy: 0.3334856331348419

[DiffTRe] Epoch 249
	Epoch loss = 0.00162
	Gradient norm: 0.023493144661188126
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 997.16455078125 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.11347094178199768

	Predicted free_energy: 0.4332475960254669

[DiffTRe] Epoch 250
	Epoch loss = 0.00159
	Gradient norm: 0.02310953661799431
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.6954345703125 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.04176544398069382

	Predicted free_energy: 0.3343840539455414

[DiffTRe] Epoch 251
	Epoch loss = 0.00162
	Gradient norm: 0.022787710651755333
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 997.1954345703125 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.11304014921188354

	Predicted free_energy: 0.4326384663581848

[DiffTRe] Epoch 252
	Epoch loss = 0.00158
	Gradient norm: 0.022427435964345932
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.691650390625 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.04239906743168831

	Predicted free_energy: 0.33524465560913086

[DiffTRe] Epoch 253
	Epoch loss = 0.00161
	Gradient norm: 0.022121943533420563
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 997.2254028320312 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.11262167990207672

	Predicted free_energy: 0.43204641342163086

[DiffTRe] Epoch 254
	Epoch loss = 0.00158
	Gradient norm: 0.0217763502150774
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.6873168945312 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.04301804304122925

	Predicted free_energy: 0.3360857367515564

[DiffTRe] Epoch 255
	Epoch loss = 0.00160
	Gradient norm: 0.02147875353693962
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 997.2525024414062 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.11220299452543259

	Predicted free_energy: 0.4314555823802948

[DiffTRe] Epoch 256
	Epoch loss = 0.00157
	Gradient norm: 0.021138885989785194
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.6825561523438 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.043636396527290344

	Predicted free_energy: 0.33692556619644165

[DiffTRe] Epoch 257
	Epoch loss = 0.00159
	Gradient norm: 0.02084331586956978
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 997.281982421875 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.11177729070186615

	Predicted free_energy: 0.43085500597953796

[DiffTRe] Epoch 258
	Epoch loss = 0.00156
	Gradient norm: 0.020505057647824287
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.6782836914062 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.044256504625082016

	Predicted free_energy: 0.3377690613269806

[DiffTRe] Epoch 259
	Epoch loss = 0.00158
	Gradient norm: 0.020211519673466682
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 997.31005859375 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.11134753376245499

	Predicted free_energy: 0.43024954199790955

[DiffTRe] Epoch 260
	Epoch loss = 0.00156
	Gradient norm: 0.01987781934440136
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.6725463867188 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.04487500339746475

	Predicted free_energy: 0.33861014246940613

[DiffTRe] Epoch 261
	Epoch loss = 0.00158
	Gradient norm: 0.019589783623814583
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 997.337646484375 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.11091926693916321

	Predicted free_energy: 0.4296465218067169

[DiffTRe] Epoch 262
	Epoch loss = 0.00155
	Gradient norm: 0.019266264513134956
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.6658935546875 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.045479368418455124

	Predicted free_energy: 0.33943289518356323

[DiffTRe] Epoch 263
	Epoch loss = 0.00157
	Gradient norm: 0.01898890733718872
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 997.3662109375 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.11050308495759964

	Predicted free_energy: 0.42906057834625244

[DiffTRe] Epoch 264
	Epoch loss = 0.00154
	Gradient norm: 0.018679890781641006
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.6611328125 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.04606745019555092

	Predicted free_energy: 0.3402324616909027

[DiffTRe] Epoch 265
	Epoch loss = 0.00156
	Gradient norm: 0.018414771184325218
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 997.391845703125 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.11009776592254639

	Predicted free_energy: 0.4284905195236206

[DiffTRe] Epoch 266
	Epoch loss = 0.00154
	Gradient norm: 0.018120476976037025
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.6553955078125 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.046635352075099945

	Predicted free_energy: 0.3410051465034485

[DiffTRe] Epoch 267
	Epoch loss = 0.00156
	Gradient norm: 0.01786637306213379
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 997.4170532226562 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.10970351099967957

	Predicted free_energy: 0.42793509364128113

[DiffTRe] Epoch 268
	Epoch loss = 0.00153
	Gradient norm: 0.017583433538675308
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.650146484375 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.0471920371055603

	Predicted free_energy: 0.3417620062828064

[DiffTRe] Epoch 269
	Epoch loss = 0.00155
	Gradient norm: 0.01733742654323578
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 997.4432373046875 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.1093129813671112

	Predicted free_energy: 0.42738577723503113

[DiffTRe] Epoch 270
	Epoch loss = 0.00153
	Gradient norm: 0.01706208661198616
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.6453857421875 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.047740355134010315

	Predicted free_energy: 0.34250783920288086

[DiffTRe] Epoch 271
	Epoch loss = 0.00154
	Gradient norm: 0.016821404919028282
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 997.4689331054688 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.10892331600189209

	Predicted free_energy: 0.426837682723999

[DiffTRe] Epoch 272
	Epoch loss = 0.00152
	Gradient norm: 0.01655171625316143
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.6392211914062 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.048284389078617096

	Predicted free_energy: 0.34324881434440613

[DiffTRe] Epoch 273
	Epoch loss = 0.00154
	Gradient norm: 0.016314944252371788
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 997.4931640625 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.10853317379951477

	Predicted free_energy: 0.4262895882129669

[DiffTRe] Epoch 274
	Epoch loss = 0.00152
	Gradient norm: 0.016050705686211586
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.6325073242188 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.048825446516275406

	Predicted free_energy: 0.3439849019050598

[DiffTRe] Epoch 275
	Epoch loss = 0.00153
	Gradient norm: 0.01581910066306591
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 997.5183715820312 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.10814514011144638

	Predicted free_energy: 0.4257451593875885

[DiffTRe] Epoch 276
	Epoch loss = 0.00151
	Gradient norm: 0.015562040731310844
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.6263427734375 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.04935704171657562

	Predicted free_energy: 0.34470877051353455

[DiffTRe] Epoch 277
	Epoch loss = 0.00153
	Gradient norm: 0.015337575227022171
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 997.5435791015625 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.10776461660861969

	Predicted free_energy: 0.42521047592163086

[DiffTRe] Epoch 278
	Epoch loss = 0.00151
	Gradient norm: 0.015089784748852253
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.6187133789062 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.04987607151269913

	Predicted free_energy: 0.34541556239128113

[DiffTRe] Epoch 279
	Epoch loss = 0.00152
	Gradient norm: 0.014873301610350609
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 997.56640625 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.1073911041021347

	Predicted free_energy: 0.4246855676174164

[DiffTRe] Epoch 280
	Epoch loss = 0.00150
	Gradient norm: 0.014634912833571434
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.6134643554688 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.050382304936647415

	Predicted free_energy: 0.3461052477359772

[DiffTRe] Epoch 281
	Epoch loss = 0.00151
	Gradient norm: 0.014426208101212978
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 997.5878295898438 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.10702453553676605

	Predicted free_energy: 0.42417165637016296

[DiffTRe] Epoch 282
	Epoch loss = 0.00150
	Gradient norm: 0.014196699485182762
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.6068115234375 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.05087818205356598

	Predicted free_energy: 0.34678030014038086

[DiffTRe] Epoch 283
	Epoch loss = 0.00151
	Gradient norm: 0.013994988054037094
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 997.6106567382812 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.10666337609291077

	Predicted free_energy: 0.4236650764942169

[DiffTRe] Epoch 284
	Epoch loss = 0.00149
	Gradient norm: 0.013772637583315372
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.6005859375 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.05136691778898239

	Predicted free_energy: 0.3474455773830414

[DiffTRe] Epoch 285
	Epoch loss = 0.00150
	Gradient norm: 0.013576410710811615
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 997.632568359375 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.10630543529987335

	Predicted free_energy: 0.42316335439682007

[DiffTRe] Epoch 286
	Epoch loss = 0.00149
	Gradient norm: 0.013360058888792992
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.5929565429688 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.05184826999902725

	Predicted free_energy: 0.34810110926628113

[DiffTRe] Epoch 287
	Epoch loss = 0.00150
	Gradient norm: 0.013168377801775932
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 997.6549072265625 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.10595142096281052

	Predicted free_energy: 0.4226665496826172

[DiffTRe] Epoch 288
	Epoch loss = 0.00148
	Gradient norm: 0.012957380153238773
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.5863037109375 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.05232396721839905

	Predicted free_energy: 0.3487493097782135

[DiffTRe] Epoch 289
	Epoch loss = 0.00149
	Gradient norm: 0.012770150788128376
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 997.6758422851562 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.10559830069541931

	Predicted free_energy: 0.4221721589565277

[DiffTRe] Epoch 290
	Epoch loss = 0.00148
	Gradient norm: 0.012564953416585922
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.57958984375 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.05279240384697914

	Predicted free_energy: 0.3493877351284027

[DiffTRe] Epoch 291
	Epoch loss = 0.00149
	Gradient norm: 0.01238304190337658
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 997.69677734375 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.10525020956993103

	Predicted free_energy: 0.4216838777065277

[DiffTRe] Epoch 292
	Epoch loss = 0.00148
	Gradient norm: 0.012184222228825092
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.5734252929688 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.053251851350069046

	Predicted free_energy: 0.35001394152641296

[DiffTRe] Epoch 293
	Epoch loss = 0.00148
	Gradient norm: 0.012008155696094036
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 997.7172241210938 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.10490732640028

	Predicted free_energy: 0.4212041199207306

[DiffTRe] Epoch 294
	Epoch loss = 0.00147
	Gradient norm: 0.011816306971013546
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.5671997070312 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.05370230972766876

	Predicted free_energy: 0.35062795877456665

[DiffTRe] Epoch 295
	Epoch loss = 0.00148
	Gradient norm: 0.011645985767245293
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 997.7371826171875 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.10457095503807068

	Predicted free_energy: 0.42073294520378113

[DiffTRe] Epoch 296
	Epoch loss = 0.00147
	Gradient norm: 0.011460814625024796
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.56005859375 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.05414401739835739

	Predicted free_energy: 0.3512297570705414

[DiffTRe] Epoch 297
	Epoch loss = 0.00148
	Gradient norm: 0.011296131648123264
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 997.7576904296875 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.10423905402421951

	Predicted free_energy: 0.42026907205581665

[DiffTRe] Epoch 298
	Epoch loss = 0.00146
	Gradient norm: 0.011117134243249893
	Elapsed time = 0.000 min

[Propagate] Effective sample size: 999.5538940429688 (990.0) -> Recompute is False

[Statepoint 0]

	kT = 2.563 ref_kT = 2.560

	Predicted entropy: -0.05457602068781853

	Predicted free_energy: 0.35181936621665955

[DiffTRe] Epoch 299
	Epoch loss = 0.00147
	Gradient norm: 0.010957307182252407
	Elapsed time = 0.000 min

Results#

plt.plot(trainer.epoch_losses)
plt.title("Loss History")
plt.xlabel("Iterations")
plt.ylabel("Loss")

Text(0, 0.5, 'Loss')

../_images/b5e40e0b24ca370c0b5a4acd6034c531a8165cbbfbab20dfb0a13d9a573d5958.png

The plot shows, that DiffTRe is able to learn the correct ensemble average.

last_epoch = len(trainer.predictions[0]) - 1

plt.plot(onp.linspace(0.00, 1.0, 50), trainer.predictions[0][0]['rdf'], label="Initial")
plt.plot(onp.linspace(0.00, 1.0, 50), trainer.predictions[0][last_epoch]['rdf'], label="Final")
plt.plot(r, radial_distribution(r), label="Reference")
plt.legend()
plt.title("Radial Distribution Function")
plt.show()

../_images/c114272c02e68075d3fa4b8ba78d0602aa544295897230738617933910d97b0f.png

Let’s also take a look at the inferred parameters.

print(trainer.params)

{'r_0': Array(0.3475688, dtype=float32), 'scaled_b': Array(2.873409, dtype=float32)}

Differentiable Trajectory Reweighting (DiffTRe)

Contents

Differentiable Trajectory Reweighting (DiffTRe)#

Concepts#

Toy Example#

Results#

Further Reading#

Examples#

Publications#

References#