feat: 🧑‍💻 add **rough** score range for each metric

pyiqa/api_helpers.py

@@ -1,6 +1,7 @@
 import fnmatch
 import re
 from pyiqa.default_model_configs import DEFAULT_CONFIGS
+from pyiqa.dataset_info import DATASET_INFO
 
 from pyiqa.utils import get_root_logger
 from pyiqa.models.inference_model import InferenceModel
@@ -49,3 +50,8 @@ def list_models(metric_mode=None, filter='', exclude_filters=''):
             if len(exclude_models):
                 models = set(models).difference(exclude_models)
     return list(sorted(models, key=_natural_key))
+
+
+def get_dataset_info(dataset_name):
+    assert dataset_name in DATASET_INFO.keys(), f'Dataset {dataset_name} not implemented yet.'
+    return DATASET_INFO[dataset_name]

pyiqa/archs/ahiq_arch.py

@@ -198,7 +198,7 @@ def __init__(
             )
         elif pretrained:
             weight_path = load_file_from_url(default_model_urls["pipal"])
-            checkpoint = torch.load(weight_path)
+            checkpoint = torch.load(weight_path, map_location='cpu', weights_only=False)
             self.regressor.load_state_dict(checkpoint["regressor_model_state_dict"])
             self.deform_net.load_state_dict(checkpoint["deform_net_model_state_dict"])
 

pyiqa/archs/arch_util.py

@@ -163,7 +163,7 @@ def load_pretrained_network(net, model_path, strict=True, weight_keys=None):
     if model_path.startswith("https://") or model_path.startswith("http://"):
         model_path = load_file_from_url(model_path)
     print(f"Loading pretrained model {net.__class__.__name__} from {model_path}")
-    state_dict = torch.load(model_path, map_location=torch.device("cpu"))
+    state_dict = torch.load(model_path, map_location=torch.device("cpu"), weights_only=False)
     if weight_keys is not None:
         state_dict = state_dict[weight_keys]
     state_dict = clean_state_dict(state_dict)

pyiqa/archs/brisque_arch.py

@@ -61,7 +61,7 @@ def brisque(x: torch.Tensor,
     scaled_features = scale_features(features)
 
     if pretrained_model_path:
-        sv_coef, sv = torch.load(pretrained_model_path)
+        sv_coef, sv = torch.load(pretrained_model_path, weights_only=False)
         sv_coef = sv_coef.to(x)
         sv = sv.to(x)
 

pyiqa/archs/clipiqa_arch.py

@@ -138,7 +138,7 @@ def __init__(self,
 
         if pretrained and 'clipiqa+' in model_type:
             if model_type == 'clipiqa+' and backbone == 'RN50':
-                self.prompt_learner.ctx.data = torch.load(load_file_from_url(default_model_urls['clipiqa+']))
+                self.prompt_learner.ctx.data = torch.load(load_file_from_url(default_model_urls['clipiqa+']), weights_only=False)
             elif model_type in default_model_urls.keys():
                 load_pretrained_network(self, default_model_urls[model_type], True, 'params')
             else:

pyiqa/archs/liqe_arch.py

@@ -75,7 +75,7 @@ def __init__(self,
             text_feat_cache_path = os.path.expanduser("~/.cache/pyiqa/liqe_text_feat.pt")
 
         if os.path.exists(text_feat_cache_path):
-            self.text_features = torch.load(text_feat_cache_path, map_location='cpu')
+            self.text_features = torch.load(text_feat_cache_path, map_location='cpu', weights_only=False)
         else:
             print(f'Generating text features for LIQE model, will be cached at {text_feat_cache_path}.')
             if self.mtl:

pyiqa/dataset_info.py

@@ -0,0 +1,39 @@
+
+DATASET_INFO = {
+    "live": {
+        "score_range": (1, 100),
+        "mos_type": "dmos"
+    },
+    "csiq": {
+        "score_range": (0, 1),
+        "mos_type": "dmos"
+    },
+    "tid": {
+        "score_range": (0, 9),
+        "mos_type": "mos"
+    },
+    "kadid": {
+        "score_range": (1, 5),
+        "mos_type": "mos"
+    },
+    "koniq": {
+        "score_range": (1, 100),
+        "mos_type": "mos"
+    },
+    "clive": {
+        "score_range": (1, 100),
+        "mos_type": "mos"
+    },
+    "flive": {
+        "score_range": (1, 100),
+        "mos_type": "mos"
+    },
+    "spaq": {
+        "score_range": (1, 100),
+        "mos_type": "mos"
+    },
+    "ava": {
+        "score_range": (1, 10),
+        "mos_type": "mos"
+    },
+}