This package is to JLD2 what PyCallJLD is to JLD, implementing a serializer for saving and loading PyCall PyObjects with JLD2.
Please see the official documentation for usage and contribution guidelines.
| Documentation | Coverage | CI Status | Releases |
|---|---|---|---|
 |
 |
 |
 |
 |
 |
 |
 |
| Dependents | Issues | JuliaHub Status | Downloads |
 |
 |
 |
Please see the documentation for full usage.
To save and load JLD2, load PyCall, JLD2, and PyCallJLD2 in the same scope as where you intend to use the JLD2.save and JLD2.load functions.
If you are coming from PyCallJLD, simply replace JLD with JLD2 everywhere in your usage.
The following example is take from PyCallJLD for direct comparison:
using PyCall, JLD2, PyCallJLD2
@pyimport sklearn.linear_model as lm
# Create some Python objects
m1 = lm.LinearRegression()
m2 = lm.ARDRegression()
# Save them to models.jld2
JLD2.save("models.jld2", "mods", [m1, m2])
# Load them back
models = JLD2.load("models.jld2", "mods")
# Do a dance🕺Just as in PyCallJLD, these objects are saved with pickle.dumps.
The following authors are responsible for authoring this package:
- Sasha Petrenko [email protected] @AP6YC
This package is heavily based upon the PyCallJLD.jl package; the funky-monkey-wrenching of PyCall ccalls, pointers, and other low-level tomfoolery would have been arcane and indecipherable without this prior work.
This package merely modifies its internal usage to match the modified JLD2 API for custom serialization.
This package uses the MIT License.