Thesis/pytorch-dnc
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Update copy task defaults and readme

Browse Source
This commit is contained in:
ixaxaar 2017-11-12 15:56:33 +05:30
parent 256b323cdc
commit f9cc0529b1
2 changed files with 9 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
README.md
View File

@ -78,7 +78,14 @@ The copy task, as descibed in the original paper, is included in the repo.
From the project root:
```bash
python ./tasks/copy_task.py -cuda 0 -optim rmsprop -batch_size 100 -mem_slot 64
python ./tasks/copy_task.py -cuda 0 -optim rmsprop -batch_size 32 -mem_slot 64 # (original implementation)
python ./tasks/copy_task.py -cuda 0 -lr 0.001 -rnn_type lstm -nlayer 1 -nhlayer 2 -mem_slot 32 -batch_size 32 -optim adam # (faster convergence)
```
For the full set of options, see:
```
python ./tasks/copy_task.py --help
```
The copy task can be used to debug memory using [Visdom](https://github.com/facebookresearch/visdom).

2
tasks/copy_task.py
View File

@ -29,7 +29,7 @@ parser.add_argument('-rnn_type', type=str, default='lstm', help='type of recurre
parser.add_argument('-nhid', type=int, default=64, help='number of hidden units of the inner nn')
parser.add_argument('-dropout', type=float, default=0, help='controller dropout')
parser.add_argument('-nlayer', type=int, default=2, help='number of layers')
parser.add_argument('-nlayer', type=int, default=1, help='number of layers')
parser.add_argument('-nhlayer', type=int, default=2, help='number of hidden layers')
parser.add_argument('-lr', type=float, default=1e-4, help='initial learning rate')
parser.add_argument('-optim', type=str, default='adam', help='learning rule, supports adam|rmsprop')