Merge pull request #2274 from huggingface/bulk_runner_tweaks

Better all res resolution for bulk runner

Author: rwightman

bulk_runner.py

@@ -21,7 +21,7 @@
 from typing import Callable, List, Tuple, Union
 
 
-from timm.models import is_model, list_models, get_pretrained_cfg
+from timm.models import is_model, list_models, get_pretrained_cfg, get_arch_pretrained_cfgs
 
 
 parser = argparse.ArgumentParser(description='Per-model process launcher')
@@ -98,23 +98,44 @@ def _get_model_cfgs(
         num_classes=None,
         expand_train_test=False,
         include_crop=True,
+        expand_arch=False,
 ):
-    model_cfgs = []
-    for n in model_names:
-        pt_cfg = get_pretrained_cfg(n)
-        if num_classes is not None and getattr(pt_cfg, 'num_classes', 0) != num_classes:
-            continue
-        model_cfgs.append((n, pt_cfg.input_size[-1], pt_cfg.crop_pct))
-        if expand_train_test and pt_cfg.test_input_size is not None:
-            if pt_cfg.test_crop_pct is not None:
-                model_cfgs.append((n, pt_cfg.test_input_size[-1], pt_cfg.test_crop_pct))
+    model_cfgs = set()
+
+    for name in model_names:
+        if expand_arch:
+            pt_cfgs = get_arch_pretrained_cfgs(name).values()
+        else:
+            pt_cfg = get_pretrained_cfg(name)
+            pt_cfgs = [pt_cfg] if pt_cfg is not None else []
+
+        for cfg in pt_cfgs:
+            if cfg.input_size is None:
+                continue
+            if num_classes is not None and getattr(cfg, 'num_classes', 0) != num_classes:
+                continue
+
+            # Add main configuration
+            size = cfg.input_size[-1]
+            if include_crop:
+                model_cfgs.add((name, size, cfg.crop_pct))
             else:
-                model_cfgs.append((n, pt_cfg.test_input_size[-1], pt_cfg.crop_pct))
+                model_cfgs.add((name, size))
+
+            # Add test configuration if required
+            if expand_train_test and cfg.test_input_size is not None:
+                test_size = cfg.test_input_size[-1]
+                if include_crop:
+                    test_crop = cfg.test_crop_pct or cfg.crop_pct
+                    model_cfgs.add((name, test_size, test_crop))
+                else:
+                    model_cfgs.add((name, test_size))
+
+    # Format the output
     if include_crop:
-        model_cfgs = [(n, {'img-size': r, 'crop-pct': cp}) for n, r, cp in sorted(model_cfgs)]
+        return [(n, {'img-size': r, 'crop-pct': cp}) for n, r, cp in sorted(model_cfgs)]
     else:
-        model_cfgs = [(n, {'img-size': r}) for n, r, cp in sorted(model_cfgs)]
-    return model_cfgs
+        return [(n, {'img-size': r}) for n, r in sorted(model_cfgs)]
 
 
 def main():
@@ -132,17 +153,16 @@ def main():
         model_cfgs = _get_model_cfgs(model_names, num_classes=1000, expand_train_test=True)
     elif args.model_list == 'all_res':
         model_names = list_models()
-        model_cfgs = _get_model_cfgs(model_names, expand_train_test=True, include_crop=False)
+        model_cfgs = _get_model_cfgs(model_names, expand_train_test=True, include_crop=False, expand_arch=True)
     elif not is_model(args.model_list):
         # model name doesn't exist, try as wildcard filter
         model_names = list_models(args.model_list)
         model_cfgs = [(n, None) for n in model_names]
 
     if not model_cfgs and os.path.exists(args.model_list):
         with open(args.model_list) as f:
-            model_cfgs = []
             model_names = [line.rstrip() for line in f]
-            _get_model_cfgs(
+            model_cfgs = _get_model_cfgs(
                 model_names,
                 #num_classes=1000,
                 expand_train_test=True,

timm/models/__init__.py

@@ -95,4 +95,5 @@
 from ._prune import adapt_model_from_string
 from ._registry import split_model_name_tag, get_arch_name, generate_default_cfgs, register_model, \
     register_model_deprecations, model_entrypoint, list_models, list_pretrained, get_deprecated_models, \
-    is_model, list_modules, is_model_in_modules, is_model_pretrained, get_pretrained_cfg, get_pretrained_cfg_value
+    is_model, list_modules, is_model_in_modules, is_model_pretrained, get_pretrained_cfg, get_pretrained_cfg_value, \
+    get_arch_pretrained_cfgs

timm/models/_registry.py

@@ -16,7 +16,7 @@
 __all__ = [
     'split_model_name_tag', 'get_arch_name', 'register_model', 'generate_default_cfgs',
     'list_models', 'list_pretrained', 'is_model', 'model_entrypoint', 'list_modules', 'is_model_in_modules',
-    'get_pretrained_cfg_value', 'is_model_pretrained'
+    'get_pretrained_cfg_value', 'is_model_pretrained', 'get_pretrained_cfgs_for_arch'
 ]
 
 _module_to_models: Dict[str, Set[str]] = defaultdict(set)  # dict of sets to check membership of model in module
@@ -341,3 +341,12 @@ def get_pretrained_cfg_value(model_name: str, cfg_key: str) -> Optional[Any]:
     """
     cfg = get_pretrained_cfg(model_name, allow_unregistered=False)
     return getattr(cfg, cfg_key, None)
+
+
+def get_arch_pretrained_cfgs(model_name: str) -> Dict[str, PretrainedCfg]:
+    """ Get all pretrained cfgs for a given architecture.
+    """
+    arch_name, _ = split_model_name_tag(model_name)
+    model_names = _model_with_tags[arch_name]
+    cfgs = {m: _model_pretrained_cfgs[m] for m in model_names}
+    return cfgs