Mo Tiwari issues

Results 86 issues of


                                            Mo Tiwari

Implement a `make clean`

User report: docker installation instructions in `README` are incomplete

"the docker instructions are incomplete and even if completed, they fail as the distribution has older libraries."

Fix platform-specific guide links in `README`

Idea: from first batch of ref points in each swap step, only keep track of best medoid to flip with (probably the one it's assigned to?)

Then this becomes a form of local search

Both `main` and `slowdown` branches appear to be nondeterministic unless number of threads is set to 1

Even though this shouldn't affect things. Why does it?

Stop SWAP steps after they improve loss by less than some threshold $t$ (e.g., 0.1%)

Allow for $\epsilon$-suboptimal arms

Need to ensure `k` elements broadcasting to for each arm are laid out contiguously in memory (after `slowdown` merge)

Cannot load checkpoints that script depends on

Running `python train_augment_net2.py --use_augment_net` as suggested by the README results in: ``` Could not open finetuned_checkpoints/dataset:mnist_datasize:1600_hyperparam:weightDecayGlobal_seed:1.pkl sailhome/motiwari/anaconda3/envs/ift-env/lib/python3.7/site-packages/numpy/core/fromnumeric.py:3420: RuntimeWarning: Mean of empty slice. out=out, **kwargs) /sailhome/motiwari/anaconda3/envs/ift-env/lib/python3.7/site-packages/numpy/core/_methods.py:188: RuntimeWarning: invalid value encountered in...

Cannot set attribute `train_data` etc. of datasets

When executing the code as-is, pytorch throws an error when trying to subsample the datasets according to the command line parameter. The attribute needs to be changed from `train_data` to...