-
- [ ] [Rvmmin](https://cran.r-project.org/web/packages/Rvmmin/index.html): for BFGS and L-BFGS-B
- [ ] [Rcgmin](https://cran.r-project.org/web/packages/Rcgmin/index.html): for Conjugate gradient
- […
-
- [ ] SGD
- [ ] Brent's Algorithm in One Dimension
- [x] Generating Set Search (compass search/pattern search)
- [ ] CMA-ES
- [ ] Conjugate Gradient Descent
- [x] BFGS
- [x] Linear Programming
…
-
Hi! Sorry for another question!
When executing the batch script, all iterations other than the first one only contains the last stimulus:
```
Stimulus name: stim_plain
[stim] configuration:
[…
-
when I run the test.py , the code get this , I think there is something wrong in the line 299.
Can u help ?
`Traceback (most recent call last):
File "test.py", line 152, in
example_MNIST…
-
Hello all,
@amartinhuertas and me have been working on a Geometric MultiGrid solver/preconditioner for Gridap. To solve the coarsest-level linear system in parallel, we decided to use MUMPS throug…
-
A good way to define `grad` in terms of `jvp` is as the vector that satisfies for all v
![image](https://user-images.githubusercontent.com/1458824/99009537-08544b80-24fd-11eb-9d78-502fec79530e.png)…
-
There's been quite a bit of interesting work recently looking at natural gradients for variational inference with exponential family q-distributions, with non-conjugate / non-exponential family likeli…
-
In the comment of `class RidgeSketch`, "ridgesketch" is listed as an option, but in fact it is not implemented.
```python
solver (str): defaults to 'auto'.
If dimensions small, uses direct so…
-
The speed of the matching and bundle adjustment are slow and if we use it in a moving robot, it can't work real time. And there are some new method to update such as :knn ,prosaic and conjugate gradie…
-
Describe other approaches than hill-climbing, in case that ever falls down (e.g. if it's too slow, search space is narrow in diagonal direction, or there are too many dimensions).
tinyKACTL tersely…