Skip to content

How to snapshot and resume a simulation

This guide shows how to save a simulation checkpoint and resume from it in a new segment. For background on why snapshotting works this way and what is preserved, see Snapshotting.

Save an ABM snapshot

Run the first segment to completion, then save:

1
2
3
4
5
6
7
8
import laser.measles as lm

params1 = lm.ABMParams(num_ticks=3650, seed=42, start_time="2000-01")
model1 = lm.ABMModel(scenario, params1)
model1.components = [lm.VitalDynamicsProcess, lm.InfectionProcess]
model1.run()

lm.save_snapshot(model1, "checkpoint.h5")

save_snapshot prints a summary:

1
2
3
4
5
Squashing 12,483 recovered agents
Snapshot saved → checkpoint.h5
  t_snap=3650  snap_date=2009-12-28  people.count=486,712
  cbr_per_1000=32.0
  Resume with: params.start_time='2009-12'

Note the Resume with: line — you need this value for the next step.

Do not use the model after saving

save_snapshot mutates the ABM model (squashes R agents, normalizes vaccination dates). Treat it as a terminal operation on model1.

Resume an ABM from a snapshot

Create new params with start_time matching the snapshot date, then load:

1
2
3
4
params2 = lm.ABMParams(num_ticks=1825, seed=42, start_time="2009-12")
model2 = lm.load_snapshot("checkpoint.h5", params2,
                           components=[lm.VitalDynamicsProcess, lm.InfectionProcess])
model2.run()

You can also use the classmethod alias:

1
2
model2 = lm.ABMModel.from_snapshot("checkpoint.h5", params2,
                                    components=[lm.VitalDynamicsProcess, lm.InfectionProcess])

start_time must match the snapshot date

Setting the wrong start_time causes date-dependent components (SIA calendars, importation, age-based vaccination) to fire at incorrect calendar times.

Save and resume a compartmental model

The procedure is the same, using compartmental classes:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
import laser.measles as lm
from laser.measles.compartmental.components import InfectionSeedingProcess, InfectionProcess

# Segment 1
params1 = lm.CompartmentalParams(num_ticks=365, seed=42, start_time="2000-01")
model1 = lm.CompartmentalModel(scenario, params1)
model1.components = [InfectionSeedingProcess, InfectionProcess]
model1.run()

lm.save_snapshot(model1, "checkpoint.h5")

1
2
3
4
# Segment 2
params2 = lm.CompartmentalParams(num_ticks=365, seed=42, start_time="2001-01")
model2 = lm.load_snapshot("checkpoint.h5", params2, components=[InfectionProcess])
model2.run()

Or using the classmethod alias:

1
2
model2 = lm.CompartmentalModel.from_snapshot("checkpoint.h5", params2,
                                              components=[InfectionProcess])

Exclude seeding from the second segment

Pass the same component list you used in segment 1 — except for InfectionSeedingProcess. Infections are already present in the restored state; re-seeding would create spurious infections.

1
2
3
4
5
# Segment 1 — seeding needed to start the epidemic
model1.components = [InfectionSeedingProcess, InfectionProcess]

# Segment 2 — epidemic is already running; seeding not needed
model2 = lm.load_snapshot(snap, params2, components=[InfectionProcess])

Similarly, do not include InitializeEquilibriumStatesProcess in the second segment — it would overwrite the loaded state.

Inspect a snapshot file

Snapshots are standard HDF5 files, readable with any HDF5-compatible tool:

1
2
3
4
5
6
7
8
import h5py

with h5py.File("checkpoint.h5", "r") as f:
    print("snap_date:", f["snap_date"][()].decode())
    print("t_snap:   ", int(f["t_snap"][()]))
    print("people.count:", int(f["people/count"][()]))
    print("patch_states shape:", f["patch_states"].shape)
    print("keys:", list(f.keys()))

Or from the command line:

1
h5ls -r checkpoint.h5

See also