iicd/xautodl
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Update NATS-Bench README

Browse Source
This commit is contained in:
D-X-Y
2020-10-29 12:34:04 -07:00
parent 45ce442058
commit e2daffa0ee
2 changed files with 6 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
docs/NATS-Bench.md
View File

@@ -35,6 +35,8 @@ To merge the chunks into the original full archive, you can use `cat file_name*
|:-----------|:---------------------:|:-------------:|:------------------:|:-------------------------------:|:--------------------------:|:------------------:|
| 2020.08.31 | [NATS-tss-v1_0-3ffb9.pickle.pbz2](https://drive.google.com/file/d/1vzyK0UVH2D3fTpa1_dSWnp1gvGpAxRul/view?usp=sharing) | [NATS-tss-v1_0-3ffb9-simple.tar](https://drive.google.com/file/d/17_saCsj_krKjlCBLOJEpNtzPXArMCqxU/view?usp=sharing) | [NATS-tss-v1_0-3ffb9-full](https://drive.google.com/drive/folders/17S2Xg_rVkUul4KuJdq0WaWoUuDbo8ZKB?usp=sharing) | [NATS-sss-v1_0-50262.pickle.pbz2](https://drive.google.com/file/d/1IabIvzWeDdDAWICBzFtTCMXxYWPIOIOX/view?usp=sharing) | [NATS-sss-v1_0-50262-simple.tar](https://drive.google.com/file/d/1scOMTUwcQhAMa_IMedp9lTzwmgqHLGgA/view?usp=sharing) | [NATS-sss-v1_0-50262-full](api.reload(index=12)) |
These benchmark files (without pretrained weights) can also be downloaded from [Dropbox](https://www.dropbox.com/sh/gm0s1wrgcq10lhb/AAANoryEyR2Q9xHtSJJVE_bha?dl=0).
1, create the benchmark instance:
```

4
lib/models/cell_searchs/search_model_darts_nasnet.py
View File

@@ -80,6 +80,10 @@ class NASNetworkDARTS(nn.Module):
for k, op_name in enumerate(self.op_names):
if op_name == 'none': continue
edges.append( (op_name, j, ws[k]) )
# (TODO) xuanyidong:
# Here the selected two edges might come from the same input node.
# And this case could be a problem that two edges will collapse into a single one
# due to our assumption -- at most one edge from an input node during evaluation.
edges = sorted(edges, key=lambda x: -x[-1])
selected_edges = edges[:2]
gene.append( tuple(selected_edges) )