Add list_benchmarks to compare models easily

requirements.txt

@@ -1,3 +1,4 @@
 torch>=1.4.0
 torchvision>=0.5.0
 pyyaml
+pandas

timm/models/registry.py

@@ -7,6 +7,7 @@
 import fnmatch
 from collections import defaultdict
 from copy import deepcopy
+import pandas as pd
 
 __all__ = ['list_models', 'is_model', 'model_entrypoint', 'list_modules', 'is_model_in_modules',
  'is_pretrained_cfg_key', 'has_pretrained_cfg_key', 'get_pretrained_cfg_value', 'is_model_pretrained']
@@ -66,8 +67,8 @@ def list_models(filter='', module='', pretrained=False, exclude_filters='', name
  name_matches_cfg (bool) - Include only models w/ model_name matching default_cfg name (excludes some aliases)
 
  Example:
- model_list('gluon_resnet*') -- returns all models starting with 'gluon_resnet'
- model_list('*resnext*, 'resnet') -- returns all models with 'resnext' in 'resnet' module
+ list_models('gluon_resnet*') -- returns all models starting with 'gluon_resnet'
+ list_models('*resnext*, 'resnet') -- returns all models with 'resnext' in 'resnet' module
  """
  if module:
  all_models = list(_module_to_models[module])
@@ -96,6 +97,25 @@ def list_models(filter='', module='', pretrained=False, exclude_filters='', name
  return list(sorted(models, key=_natural_key))
 
 
+def list_benchmarks(filter='', module='', pretrained=False, exclude_filters='', name_matches_cfg=False):
+ """ Return list of available benchmarks on imagenet, sorted alphabetically
+
+ Args:
+ filter (str) - Wildcard filter string that works with fnmatch
+ module (str) - Limit model selection to a specific sub-module (ie 'gen_efficientnet')
+ pretrained (bool) - Include only models with pretrained weights if True
+ exclude_filters (str or list[str]) - Wildcard filters to exclude models after including them with filter
+ name_matches_cfg (bool) - Include only models w/ model_name matching default_cfg name (excludes some aliases)
+
+ Example:
+ list_benchmarks('gluon_resnet*') -- returns a pandas dataframe with all the benchmarks starting with 'gluon_resnet'
+ """
+ models = list_models(filter=filter, module=module, pretrained=pretrained, exclude_filters=exclude_filters, name_matches_cfg=name_matches_cfg)
+ df = pd.read_csv("https://raw.githubusercontent.com/rwightman/pytorch-image-models/master/results/results-imagenet.csv") 
+
+ return df[df["model"].isin(models)]
+
+
 def is_model(model_name):
  """ Check if a model name exists
  """
@@ -156,4 +176,4 @@ def get_pretrained_cfg_value(model_name, cfg_key):
  """
  if model_name in _model_pretrained_cfgs:
  return _model_pretrained_cfgs[model_name].get(cfg_key, None)
- return None
+ return None