issues
search
kostaleonard
/
great-model-theory
A deep learning library for Scala.
MIT License
2
stars
0
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
Adds working MNIST example
#98
kostaleonard
closed
1 year ago
0
Use of java.util.IdentityHashMap is error-prone
#97
kostaleonard
opened
1 year ago
0
Implement DifferentiableFunction.getInputs from DifferentiableFunction.getParents
#96
kostaleonard
opened
1 year ago
0
Users should not need to know about the autodiff module to fit a model
#95
kostaleonard
opened
1 year ago
0
Dense layer initialization is hand-tuned for MNIST
#94
kostaleonard
opened
1 year ago
0
Adds MNIST dataset, NDArray refactor, and README updates
#93
kostaleonard
closed
1 year ago
0
Implement Model.predict
#92
kostaleonard
closed
1 year ago
0
Implement Model.evaluate
#91
kostaleonard
opened
1 year ago
0
Don't ignore MNIST training test in MNISTSpec
#90
kostaleonard
opened
1 year ago
0
Make NDArray broadcast faster
#89
kostaleonard
opened
1 year ago
0
Scaliest failAfter does not terminate the test after exceeding the time limit
#88
kostaleonard
opened
1 year ago
0
Input case class relies on Option argument's memory address in comparisons
#87
kostaleonard
opened
1 year ago
0
Refactor DifferentiableFunction, Layer, and Model interfaces
#86
kostaleonard
opened
1 year ago
0
Adds model fit method
#85
kostaleonard
closed
1 year ago
0
Removes Try/Success/Failure
#84
kostaleonard
closed
1 year ago
0
Removes TODO items
#83
kostaleonard
closed
1 year ago
0
Adds backpropagation and proof of concept gradient descent
#82
kostaleonard
closed
1 year ago
0
Refactor DifferentiableFunction interfaces to be more user-friendly and easier for performing gradient descent
#81
kostaleonard
opened
1 year ago
0
Remove symbolic gradient or redefine in terms of backpropagation
#80
kostaleonard
closed
1 year ago
0
Updates to ModelParameters may erroneously overwrite multiple copies
#79
kostaleonard
closed
1 year ago
0
Add autodiff and backpropagation docs for laymen
#78
kostaleonard
opened
1 year ago
0
Add nonlinearity through Sigmoid layer
#77
kostaleonard
closed
1 year ago
0
Consider ways to improve the numerical stability of DifferentiableFunctions like Reciprocal
#76
kostaleonard
opened
1 year ago
0
NDArray arrayApproximatelyEquals fails if arrays contain positive or negative infinity
#75
kostaleonard
opened
1 year ago
0
Add nonlinearity through activation function
#74
kostaleonard
closed
1 year ago
0
Tag first release when it works on MNIST
#73
kostaleonard
opened
1 year ago
0
Add links in main readme to autodiff resources for those who would like to read further
#72
kostaleonard
opened
1 year ago
0
Make NDArray toString hide all elements but first 10 and last 10 by default
#71
kostaleonard
opened
1 year ago
0
Adds basic automatic differentiation components for Dense layer gradient descent
#70
kostaleonard
closed
1 year ago
0
Consider making a MatMul DifferentiableFunction and having DotProduct call it
#69
kostaleonard
opened
1 year ago
0
DifferentiableFunction DotProduct has repeated code segments
#68
kostaleonard
opened
1 year ago
0
Check private method docstrings for information that may be relevant in public docstrings
#67
kostaleonard
opened
1 year ago
0
Add's getOutputShape (and possibly others) fails when supplied with two of the same input with shape placeholders
#66
kostaleonard
opened
1 year ago
0
Adds DifferentiableFunction tests and getOutputShape
#65
kostaleonard
closed
1 year ago
0
Add better error messages for DifferentiableFunction's DotProduct getOutputShape
#64
kostaleonard
opened
1 year ago
0
Add test coverage report and metrics
#63
kostaleonard
opened
1 year ago
0
If possible, make DifferentiableFunction's getOutputShape concrete
#62
kostaleonard
opened
1 year ago
0
NDArray arange and random may create invalid objects when shape is an empty array
#61
kostaleonard
opened
1 year ago
0
Use arrays instead of lists as arguments to NDArray apply (indexing) and slice
#60
kostaleonard
closed
1 year ago
0
Changes NDArray indexing and slicing to use array arguments
#59
kostaleonard
closed
1 year ago
0
Adds consistent shape argument types
#58
kostaleonard
closed
1 year ago
0
Adds broadcasting for == and ~=
#57
kostaleonard
closed
1 year ago
0
Add broadcasting for NDArray operations
#56
kostaleonard
closed
1 year ago
1
Make NDArray, Model, and Layer shape arguments use consistent type scheme
#55
kostaleonard
closed
1 year ago
0
Updated readme
#54
kostaleonard
closed
1 year ago
0
Adds computation graph to neural networks and broadcasting for NDArrays
#53
kostaleonard
closed
1 year ago
0
Add DifferentiableFunction tests
#52
kostaleonard
opened
1 year ago
0
Consider throwing exception in Input's compute method if the input shape does not match the object's field
#51
kostaleonard
opened
1 year ago
0
Consider making Variable a trait
#50
kostaleonard
closed
1 year ago
1
DifferentiableFunction's getInputs should not be abstract
#49
kostaleonard
opened
1 year ago
1
Next