fixed test multimodal search on cpu and gpu machines

misinformation/test/test_multimodal_search.py

@@ -1,4 +1,5 @@
 import pytest
+import math
 from PIL import Image
 import numpy
 from torch import device, cuda, no_grad
@@ -17,6 +18,8 @@ TEST_IMAGE_9 = "./test/data/IMG_3755.jpg"
 TEST_IMAGE_10 = "./test/data/IMG_3756.jpg"
 TEST_IMAGE_11 = "./test/data/IMG_3757.jpg"
 TEST_IMAGE_12 = "./test/data/pic1.png"
+related_error = 1e-3
+gpu_is_not_available = not cuda.is_available()
 
 
 def test_read_img():
@@ -51,8 +54,7 @@ def test_load_feature_extractor_model_blip2():
             extracted_feature_text = model.extract_features(
                 {"image": "", "text_input": processed_text}, mode="text"
             )
-
-    assert processed_pic[0, 0, 0, 25:35].tolist() == [
+    check_list_processed_pic = [
         -1.0039474964141846,
         -1.0039474964141846,
         -0.8433647751808167,
@@ -64,13 +66,18 @@ def test_load_feature_extractor_model_blip2():
         -0.7119789123535156,
         -1.0623412132263184,
     ]
+    for i, num in zip(range(10), processed_pic[0, 0, 0, 25:35].tolist()):
+        assert (
+            math.isclose(num, check_list_processed_pic[i], rel_tol=related_error)
+            is True
+        )
 
     assert (
         processed_text
         == "the bird sat on a tree located at the intersection of 23rd and 43rd streets"
     )
 
-    assert extracted_feature_img["image_embeds_proj"][0, 0, 10:20].tolist() == [
+    check_list_extracted_feature_img = [
         0.04566730558872223,
         -0.042554520070552826,
         -0.06970272958278656,
@@ -82,8 +89,17 @@ def test_load_feature_extractor_model_blip2():
         0.12923966348171234,
         0.006452132016420364,
     ]
+    for i, num in zip(
+        range(10), extracted_feature_img["image_embeds_proj"][0, 0, 10:20].tolist()
+    ):
+        assert (
+            math.isclose(
+                num, check_list_extracted_feature_img[i], rel_tol=related_error
+            )
+            is True
+        )
 
-    assert extracted_feature_text["text_embeds_proj"][0, 0, 10:20].tolist() == [
+    check_list_extracted_feature_text = [
         -0.1384519338607788,
         -0.008663734421133995,
         0.006240826100111008,
@@ -95,14 +111,57 @@ def test_load_feature_extractor_model_blip2():
         -0.05823372304439545,
         0.036941494792699814,
     ]
+    for i, num in zip(
+        range(10), extracted_feature_text["text_embeds_proj"][0, 0, 10:20].tolist()
+    ):
+        assert (
+            math.isclose(
+                num, check_list_extracted_feature_text[i], rel_tol=related_error
+            )
+            is True
+        )
+
+    image_paths = [TEST_IMAGE_2, TEST_IMAGE_3]
+    raw_images, images_tensors = ms.MultimodalSearch.read_and_process_images(
+        my_dict, image_paths, vis_processor
+    )
+
+    check_list_images_tensors = [
+        -1.0039474964141846,
+        -1.0039474964141846,
+        -0.8433647751808167,
+        -0.6097899675369263,
+        -0.5951915383338928,
+        -0.6243883967399597,
+        -0.6827820539474487,
+        -0.6097899675369263,
+        -0.7119789123535156,
+        -1.0623412132263184,
+    ]
+    for i, num in zip(range(10), images_tensors[0, 0, 0, 0, 25:35].tolist()):
+        assert (
+            math.isclose(num, check_list_images_tensors[i], rel_tol=related_error)
+            is True
+        )
 
     del model, vis_processor, txt_processor
     cuda.empty_cache()
 
 
-def test_load_feature_extractor_model_blip():
+@pytest.mark.parametrize(
+    ("multimodal_device"),
+    [
+        device("cpu"),
+        pytest.param(
+            device("cuda"),
+            marks=pytest.mark.skipif(
+                gpu_is_not_available, reason="gpu_is_not_availible"
+            ),
+        ),
+    ],
+)
+def test_load_feature_extractor_model_blip(multimodal_device):
     my_dict = {}
-    multimodal_device = device("cuda" if cuda.is_available() else "cpu")
     (
         model,
         vis_processor,
@@ -125,7 +184,7 @@ def test_load_feature_extractor_model_blip():
             {"image": "", "text_input": processed_text}, mode="text"
         )
 
-    assert processed_pic[0, 0, 0, 25:35].tolist() == [
+    check_list_processed_pic = [
         -1.0039474964141846,
         -1.0039474964141846,
         -0.8433647751808167,
@@ -137,13 +196,18 @@ def test_load_feature_extractor_model_blip():
         -0.7119789123535156,
         -1.0623412132263184,
     ]
+    for i, num in zip(range(10), processed_pic[0, 0, 0, 25:35].tolist()):
+        assert (
+            math.isclose(num, check_list_processed_pic[i], rel_tol=related_error)
+            is True
+        )
 
     assert (
         processed_text
         == "the bird sat on a tree located at the intersection of 23rd and 43rd streets"
     )
 
-    assert extracted_feature_img["image_embeds_proj"][0, 0, 10:20].tolist() == [
+    check_list_extracted_feature_img = [
         -0.02480311505496502,
         0.05037587881088257,
         0.039517853409051895,
@@ -155,8 +219,17 @@ def test_load_feature_extractor_model_blip():
         -0.07324369996786118,
         0.06994668394327164,
     ]
+    for i, num in zip(
+        range(10), extracted_feature_img["image_embeds_proj"][0, 0, 10:20].tolist()
+    ):
+        assert (
+            math.isclose(
+                num, check_list_extracted_feature_img[i], rel_tol=related_error
+            )
+            is True
+        )
 
-    assert extracted_feature_text["text_embeds_proj"][0, 0, 10:20].tolist() == [
+    check_list_extracted_feature_text = [
         0.0118643119931221,
         -0.01291718054562807,
         -0.0009687161073088646,
@@ -168,14 +241,34 @@ def test_load_feature_extractor_model_blip():
         0.0062415082938969135,
         0.0034833091776818037,
     ]
+    for i, num in zip(
+        range(10), extracted_feature_text["text_embeds_proj"][0, 0, 10:20].tolist()
+    ):
+        assert (
+            math.isclose(
+                num, check_list_extracted_feature_text[i], rel_tol=related_error
+            )
+            is True
+        )
 
     del model, vis_processor, txt_processor
     cuda.empty_cache()
 
 
-def test_load_feature_extractor_model_albef():
+@pytest.mark.parametrize(
+    ("multimodal_device"),
+    [
+        device("cpu"),
+        pytest.param(
+            device("cuda"),
+            marks=pytest.mark.skipif(
+                gpu_is_not_available, reason="gpu_is_not_availible"
+            ),
+        ),
+    ],
+)
+def test_load_feature_extractor_model_albef(multimodal_device):
     my_dict = {}
-    multimodal_device = device("cuda" if cuda.is_available() else "cpu")
     (
         model,
         vis_processor,
@@ -198,7 +291,7 @@ def test_load_feature_extractor_model_albef():
             {"image": "", "text_input": processed_text}, mode="text"
         )
 
-    assert processed_pic[0, 0, 0, 25:35].tolist() == [
+    check_list_processed_pic = [
         -1.0039474964141846,
         -1.0039474964141846,
         -0.8433647751808167,
@@ -210,13 +303,18 @@ def test_load_feature_extractor_model_albef():
         -0.7119789123535156,
         -1.0623412132263184,
     ]
+    for i, num in zip(range(10), processed_pic[0, 0, 0, 25:35].tolist()):
+        assert (
+            math.isclose(num, check_list_processed_pic[i], rel_tol=related_error)
+            is True
+        )
 
     assert (
         processed_text
         == "the bird sat on a tree located at the intersection of 23rd and 43rd streets"
     )
 
-    assert extracted_feature_img["image_embeds_proj"][0, 0, 10:20].tolist() == [
+    check_list_extracted_feature_img = [
         0.08971136063337326,
         -0.10915573686361313,
         -0.020636577159166336,
@@ -228,8 +326,17 @@ def test_load_feature_extractor_model_albef():
         -0.03284582123160362,
         -0.1037328764796257,
     ]
+    for i, num in zip(
+        range(10), extracted_feature_img["image_embeds_proj"][0, 0, 10:20].tolist()
+    ):
+        assert (
+            math.isclose(
+                num, check_list_extracted_feature_img[i], rel_tol=related_error
+            )
+            is True
+        )
 
-    assert extracted_feature_text["text_embeds_proj"][0, 0, 10:20].tolist() == [
+    check_list_extracted_feature_text = [
         -0.06229640915989876,
         0.11278597265481949,
         0.06628583371639252,
@@ -241,14 +348,34 @@ def test_load_feature_extractor_model_albef():
         0.050752390176057816,
         -0.0421440489590168,
     ]
+    for i, num in zip(
+        range(10), extracted_feature_text["text_embeds_proj"][0, 0, 10:20].tolist()
+    ):
+        assert (
+            math.isclose(
+                num, check_list_extracted_feature_text[i], rel_tol=related_error
+            )
+            is True
+        )
 
     del model, vis_processor, txt_processor
     cuda.empty_cache()
 
 
-def test_load_feature_extractor_model_clip_base():
+@pytest.mark.parametrize(
+    ("multimodal_device"),
+    [
+        device("cpu"),
+        pytest.param(
+            device("cuda"),
+            marks=pytest.mark.skipif(
+                gpu_is_not_available, reason="gpu_is_not_availible"
+            ),
+        ),
+    ],
+)
+def test_load_feature_extractor_model_clip_base(multimodal_device):
     my_dict = {}
-    multimodal_device = device("cuda" if cuda.is_available() else "cpu")
     (
         model,
         vis_processor,
@@ -267,7 +394,7 @@ def test_load_feature_extractor_model_clip_base():
         extracted_feature_img = model.extract_features({"image": processed_pic})
         extracted_feature_text = model.extract_features({"text_input": processed_text})
 
-    assert processed_pic[0, 0, 0, 25:35].tolist() == [
+    check_list_processed_pic = [
         -0.7995694875717163,
         -0.7849710583686829,
         -0.7849710583686829,
@@ -279,13 +406,18 @@ def test_load_feature_extractor_model_clip_base():
         -0.7703726291656494,
         -0.7703726291656494,
     ]
+    for i, num in zip(range(10), processed_pic[0, 0, 0, 25:35].tolist()):
+        assert (
+            math.isclose(num, check_list_processed_pic[i], rel_tol=related_error)
+            is True
+        )
 
     assert (
         processed_text
         == "The bird sat on a tree located at the intersection of 23rd and 43rd streets."
     )
 
-    assert extracted_feature_img[0, 10:20].tolist() == [
+    check_list_extracted_feature_img = [
         0.15101124346256256,
         -0.03759124130010605,
         -0.40093156695365906,
@@ -297,8 +429,15 @@ def test_load_feature_extractor_model_clip_base():
         -0.030034437775611877,
         0.19082790613174438,
     ]
+    for i, num in zip(range(10), extracted_feature_img[0, 10:20].tolist()):
+        assert (
+            math.isclose(
+                num, check_list_extracted_feature_img[i], rel_tol=related_error
+            )
+            is True
+        )
 
-    assert extracted_feature_text[0, 10:20].tolist() == [
+    check_list_extracted_feature_text = [
         0.15391531586647034,
         0.3078577518463135,
         0.21737979352474213,
@@ -310,14 +449,32 @@ def test_load_feature_extractor_model_clip_base():
         -0.1858849972486496,
         0.20347601175308228,
     ]
+    for i, num in zip(range(10), extracted_feature_text[0, 10:20].tolist()):
+        assert (
+            math.isclose(
+                num, check_list_extracted_feature_text[i], rel_tol=related_error
+            )
+            is True
+        )
 
     del model, vis_processor, txt_processor
     cuda.empty_cache()
 
 
-def test_load_feature_extractor_model_clip_vitl14():
+@pytest.mark.parametrize(
+    ("multimodal_device"),
+    [
+        device("cpu"),
+        pytest.param(
+            device("cuda"),
+            marks=pytest.mark.skipif(
+                gpu_is_not_available, reason="gpu_is_not_availible"
+            ),
+        ),
+    ],
+)
+def test_load_feature_extractor_model_clip_vitl14(multimodal_device):
     my_dict = {}
-    multimodal_device = device("cuda" if cuda.is_available() else "cpu")
     (
         model,
         vis_processor,
@@ -336,7 +493,7 @@ def test_load_feature_extractor_model_clip_vitl14():
         extracted_feature_img = model.extract_features({"image": processed_pic})
         extracted_feature_text = model.extract_features({"text_input": processed_text})
 
-    assert processed_pic[0, 0, 0, 25:35].tolist() == [
+    check_list_processed_pic = [
         -0.7995694875717163,
         -0.7849710583686829,
         -0.7849710583686829,
@@ -348,13 +505,18 @@ def test_load_feature_extractor_model_clip_vitl14():
         -0.7703726291656494,
         -0.7703726291656494,
     ]
+    for i, num in zip(range(10), processed_pic[0, 0, 0, 25:35].tolist()):
+        assert (
+            math.isclose(num, check_list_processed_pic[i], rel_tol=related_error)
+            is True
+        )
 
     assert (
         processed_text
         == "The bird sat on a tree located at the intersection of 23rd and 43rd streets."
     )
 
-    assert extracted_feature_img[0, 10:20].tolist() == [
+    check_list_extracted_feature_img = [
         -0.3911527395248413,
         -0.35456305742263794,
         0.5724918842315674,
@@ -366,8 +528,15 @@ def test_load_feature_extractor_model_clip_vitl14():
         0.19491413235664368,
         0.01419895887374878,
     ]
+    for i, num in zip(range(10), extracted_feature_img[0, 10:20].tolist()):
+        assert (
+            math.isclose(
+                num, check_list_extracted_feature_img[i], rel_tol=related_error
+            )
+            is True
+        )
 
-    assert extracted_feature_text[0, 10:20].tolist() == [
+    check_list_extracted_feature_text = [
         -0.07539052516222,
         0.0939129889011383,
         -0.2643853425979614,
@@ -379,14 +548,32 @@ def test_load_feature_extractor_model_clip_vitl14():
         0.18563221395015717,
         -0.3183072805404663,
     ]
+    for i, num in zip(range(10), extracted_feature_text[0, 10:20].tolist()):
+        assert (
+            math.isclose(
+                num, check_list_extracted_feature_text[i], rel_tol=related_error
+            )
+            is True
+        )
 
     del model, vis_processor, txt_processor
     cuda.empty_cache()
 
 
-def test_load_feature_extractor_model_clip_vitl14_336():
+@pytest.mark.parametrize(
+    ("multimodal_device"),
+    [
+        device("cpu"),
+        pytest.param(
+            device("cuda"),
+            marks=pytest.mark.skipif(
+                gpu_is_not_available, reason="gpu_is_not_availible"
+            ),
+        ),
+    ],
+)
+def test_load_feature_extractor_model_clip_vitl14_336(multimodal_device):
     my_dict = {}
-    multimodal_device = device("cuda" if cuda.is_available() else "cpu")
     (
         model,
         vis_processor,
@@ -405,7 +592,7 @@ def test_load_feature_extractor_model_clip_vitl14_336():
         extracted_feature_img = model.extract_features({"image": processed_pic})
         extracted_feature_text = model.extract_features({"text_input": processed_text})
 
-    assert processed_pic[0, 0, 0, 25:35].tolist() == [
+    check_list_processed_pic = [
         -0.7995694875717163,
         -0.7849710583686829,
         -0.7849710583686829,
@@ -417,13 +604,18 @@ def test_load_feature_extractor_model_clip_vitl14_336():
         -1.149931788444519,
         -1.0039474964141846,
     ]
+    for i, num in zip(range(10), processed_pic[0, 0, 0, 25:35].tolist()):
+        assert (
+            math.isclose(num, check_list_processed_pic[i], rel_tol=related_error)
+            is True
+        )
 
     assert (
         processed_text
         == "The bird sat on a tree located at the intersection of 23rd and 43rd streets."
     )
 
-    assert extracted_feature_img[0, 10:20].tolist() == [
+    check_list_extracted_feature_img = [
         -0.15060146152973175,
         -0.1998099535703659,
         0.5503129363059998,
@@ -435,8 +627,15 @@ def test_load_feature_extractor_model_clip_vitl14_336():
         0.02220013737678528,
         0.01086437702178955,
     ]
+    for i, num in zip(range(10), extracted_feature_img[0, 10:20].tolist()):
+        assert (
+            math.isclose(
+                num, check_list_extracted_feature_img[i], rel_tol=related_error
+            )
+            is True
+        )
 
-    assert extracted_feature_text[0, 10:20].tolist() == [
+    check_list_extracted_feature_text = [
         -0.1172553077340126,
         0.07105237245559692,
         -0.283934086561203,
@@ -448,6 +647,13 @@ def test_load_feature_extractor_model_clip_vitl14_336():
         0.22669515013694763,
         -0.32044747471809387,
     ]
+    for i, num in zip(range(10), extracted_feature_text[0, 10:20].tolist()):
+        assert (
+            math.isclose(
+                num, check_list_extracted_feature_text[i], rel_tol=related_error
+            )
+            is True
+        )
 
     del model, vis_processor, txt_processor
     cuda.empty_cache()