Upload 14 files

src/videogen_hub/benchmark/fal_text_guided_t2v.py

@@ -0,0 +1,120 @@
+from typing import Optional
+import os
+from tqdm import tqdm
+import json, requests
+import fal_client
+# import json
+
+def infer_text_guided_vg_bench(
+    model_name,
+    result_folder: str = "results",
+    experiment_name: str = "Exp_Text-Guided_VG",
+    overwrite_model_outputs: bool = False,
+    overwrite_inputs: bool = False,
+    limit_videos_amount: Optional[int] = None,
+):
+    """
+    Performs inference on the VideogenHub dataset using the provided text-guided video generation model.
+
+    Args:
+        model_name: name of the model we want to run inference on
+        result_folder (str, optional): Path to the root directory where the results should be saved.
+               Defaults to 'results'.
+        experiment_name (str, optional): Name of the folder inside 'result_folder' where results
+               for this particular experiment will be stored. Defaults to "Exp_Text-Guided_IG".
+        overwrite_model_outputs (bool, optional): If set to True, will overwrite any pre-existing
+               model outputs. Useful for resuming runs. Defaults to False.
+        overwrite_inputs (bool, optional): If set to True, will overwrite any pre-existing input
+               samples. Typically, should be set to False unless there's a need to update the inputs.
+               Defaults to False.
+        limit_videos_amount (int, optional): Limits the number of videos to be processed. If set to
+               None, all videos in the dataset will be processed.
+
+    Returns:
+        None. Results are saved in the specified directory.
+
+    Notes:
+        The function processes each sample from the dataset, uses the model to infer an video based
+        on text prompts, and then saves the resulting videos in the specified directories.
+    """
+    benchmark_prompt_path = "t2v_vbench_1000.json"
+    prompts = json.load(open(benchmark_prompt_path, "r"))
+    save_path = os.path.join(result_folder, experiment_name, "dataset_lookup.json")
+    if overwrite_inputs or not os.path.exists(save_path):
+        if not os.path.exists(os.path.join(result_folder, experiment_name)):
+            os.makedirs(os.path.join(result_folder, experiment_name))
+        with open(save_path, "w") as f:
+            json.dump(prompts, f, indent=4)
+
+    print(
+        "========> Running Benchmark Dataset:",
+        experiment_name,
+        "| Model:",
+        model_name,
+    )
+    
+    if model_name == 'AnimateDiff':
+        fal_model_name = 'fast-animatediff/text-to-video'
+    elif model_name == 'AnimateDiffTurbo':
+        fal_model_name = 'fast-animatediff/turbo/text-to-video'
+    elif model_name == 'FastSVD':
+        fal_model_name = 'fast-svd/text-to-video'
+    else:
+        raise ValueError("Invalid model_name")
+
+    for file_basename, prompt in tqdm(prompts.items()):
+        idx = int(file_basename.split('_')[0])
+        dest_folder = os.path.join(
+            result_folder, experiment_name, model_name
+        )
+        # file_basename = f"{idx}_{prompt['prompt_en'].replace(' ', '_')}.mp4"
+        if not os.path.exists(dest_folder):
+            os.mkdir(dest_folder)
+        dest_file = os.path.join(dest_folder, file_basename)
+        if overwrite_model_outputs or not os.path.exists(dest_file):
+            print("========> Inferencing", dest_file)
+            
+            handler = fal_client.submit(
+                f"fal-ai/{fal_model_name}",
+                arguments={
+                    "prompt": prompt["prompt_en"]
+                },
+            )
+            
+            # for event in handler.iter_events(with_logs=True):
+            #     if isinstance(event, fal_client.InProgress):
+            #         print('Request in progress')
+            #         print(event.logs)
+
+            result = handler.get()
+            result_url = result['video']['url']
+            download_mp4(result_url, dest_file)
+        else:
+            print("========> Skipping", dest_file, ", it already exists")
+
+        if limit_videos_amount is not None and (idx >= limit_videos_amount):
+            break
+
+def download_mp4(url, filename):
+    try:
+        # Send a GET request to the URL
+        response = requests.get(url, stream=True)
+        response.raise_for_status()  # Check if the request was successful
+
+        # Open a local file with write-binary mode
+        with open(filename, 'wb') as file:
+            # Write the response content to the file in chunks
+            for chunk in response.iter_content(chunk_size=8192):
+                file.write(chunk)
+
+        # print(f"Download complete: {filename}")
+
+    except requests.exceptions.RequestException as e:
+        print(f"Error downloading file: {e}")
+        
+if __name__ == "__main__":
+    pass
+    # infer_text_guided_vg_bench(model_name="AnimateDiff")
+    infer_text_guided_vg_bench(result_folder="/mnt/tjena/maxku/max_projects/VideoGenHub/results", model_name="FastSVD")
+    # infer_text_guided_vg_bench(result_folder="/mnt/tjena/maxku/max_projects/VideoGenHub/results", model_name="AnimateDiff")
+    # infer_text_guided_vg_bench(result_folder="/mnt/tjena/maxku/max_projects/VideoGenHub/results", model_name="AnimateDiffTurbo")

src/videogen_hub/benchmark/merge_prompt.py

@@ -0,0 +1,25 @@
+import json
+
+def main(prmopts_path_1, prompts_path_2):
+    prompts_1 = json.load(open(prmopts_path_1, "r"))
+    prompts_2 = json.load(open(prompts_path_2, "r"))
+    
+    new_prompts = {}
+    new_idx = 0
+    for prompt_key in prompts_1:
+        prompt_key_lst = prompt_key.split("_")
+        prompt_key_lst[0] = str(new_idx)
+        new_prompts['_'.join(prompt_key_lst)] = prompts_1[prompt_key]
+        new_idx += 1
+    
+    for prompt_key in prompts_2:
+        prompt_key_lst = prompt_key.split("_")
+        prompt_key_lst[0] = str(new_idx)
+        new_prompts['_'.join(prompt_key_lst)] = prompts_2[prompt_key]
+        new_idx += 1
+    
+    with open(f"t2v_vbench_1000.json", "w") as f:
+        json.dump(new_prompts, f, indent=4)
+
+if __name__ == "__main__":
+    main("t2v_vbench_200.json", "t2v_vbench_800.json")

src/videogen_hub/benchmark/prompt_generation.py

@@ -0,0 +1,67 @@
+import os
+import json
+import numpy as np
+import random
+
+# Randomly sample a subset of prompts for benchmarking
+def main(prompt_path, overwrite_inputs=False):
+    prompts = json.load(open(prompt_path, "r"))
+
+    # construct dimension_count map
+    dimension_count_map = {}
+    dimension_prompt_idx_map = {}
+    dimensions_count = 0
+    for i in range(len(prompts)):
+        prompt = prompts[i]
+        dimensions = prompt["dimension"]
+        for dimension in dimensions:
+            if dimension not in dimension_prompt_idx_map:
+                dimension_prompt_idx_map[dimension] = []
+            dimension_prompt_idx_map[dimension].append(i)
+
+            if dimension not in dimension_count_map:
+                dimension_count_map[dimension] = 0
+
+            dimension_count_map[dimension] += 1
+
+            dimensions_count += 1
+
+    print(
+        "Dimensions count (each prompt can contribute to more than one dimension count):",
+        dimensions_count,
+    )
+    print(dimension_count_map)
+
+    target_prompts_count = 800
+    # sample prompts based on the distribution of dimensions
+    sampled_prompts = list()
+    remaining_prompts = list()
+    dimension_probs = np.array(list(dimension_count_map.values())) / dimensions_count
+    dimensions = list(dimension_count_map.keys())
+    sample_counts = np.random.multinomial(target_prompts_count, dimension_probs)
+    print(sample_counts)
+    for dimension, count in zip(dimensions, sample_counts):
+
+        sampled_prompts_idx = random.sample(dimension_prompt_idx_map[dimension], count)
+        for idx in range(len(prompts)):
+            if idx in sampled_prompts_idx:
+                sampled_prompts.append(prompts[idx])
+            else:
+                remaining_prompts.append(prompts[idx])
+
+    save_path = "./t2v_vbench_1000.json"
+    remaing_data_save_path = "./t2v_vbench_remain_1000.json"
+    if overwrite_inputs or not os.path.exists(save_path):
+        # if not os.path.exists(os.path.join(result_folder, experiment_name)):
+        #     os.makedirs(os.path.join(result_folder, experiment_name))
+        with open(save_path, "w") as f:
+            json.dump(sampled_prompts, f, indent=4)
+            
+        with open(remaing_data_save_path, "w") as f:
+            json.dump(remaining_prompts, f, indent=4)
+    else:
+        print("Dataset already exists, skipping generation")
+
+if __name__ == "__main__":
+    main(prompt_path="VBench_full_info.json")
+    # main(prompt_path="t2v_vbench_remain_200.json")

src/videogen_hub/benchmark/sample_prompt.py

@@ -0,0 +1,70 @@
+import os
+import json
+import numpy as np
+import random
+
+# Randomly sample a subset of prompts for benchmarking
+def main(prompt_path, overwrite_inputs=False):
+    prompts = json.load(open(prompt_path, "r"))
+
+    # construct dimension_count map
+    dimension_count_map = {}
+    dimension_prompt_idx_map = {}
+    dimensions_count = 0
+    for key, prompt in prompts.items():
+        dimensions = prompt["dimension"]
+        for dimension in dimensions:
+            if dimension not in dimension_prompt_idx_map:
+                dimension_prompt_idx_map[dimension] = []
+            dimension_prompt_idx_map[dimension].append(key)
+
+            if dimension not in dimension_count_map:
+                dimension_count_map[dimension] = 0
+
+            dimension_count_map[dimension] += 1
+
+            dimensions_count += 1
+
+    print(
+        "Dimensions count (each prompt can contribute to more than one dimension count):",
+        dimensions_count,
+    )
+    print(dimension_count_map)
+
+    target_prompts_count = 800
+    # sample prompts based on the distribution of dimensions
+    sampled_prompts = {}
+    remaining_prompts = {}
+    dimension_probs = np.array(list(dimension_count_map.values())) / dimensions_count
+    dimensions = list(dimension_count_map.keys())
+    sample_counts = np.random.multinomial(target_prompts_count, dimension_probs)
+    print(np.sum(sample_counts))
+    print(sample_counts)
+    for dimension, count in zip(dimensions, sample_counts):
+
+        sampled_prompts_keys = random.sample(dimension_prompt_idx_map[dimension], count)
+        for key in prompts.keys():
+            if key in sampled_prompts_keys:
+                while key in sampled_prompts:
+                    key = random.sample(dimension_prompt_idx_map[dimension], 1)[0]
+                sampled_prompts[key] = prompts[key]
+            else:
+                remaining_prompts[key] = prompts[key]
+
+    save_path = "./t2v_vbench_800.json"
+    remaing_data_save_path = "./t2v_vbench_remain_1000.json"
+    print(len(sampled_prompts.keys()))
+    if overwrite_inputs or not os.path.exists(save_path):
+        # if not os.path.exists(os.path.join(result_folder, experiment_name)):
+        #     os.makedirs(os.path.join(result_folder, experiment_name))
+        with open(save_path, "w") as f:
+            json.dump(sampled_prompts, f, indent=4)
+            
+        with open(remaing_data_save_path, "w") as f:
+            json.dump(remaining_prompts, f, indent=4)
+    else:
+        print("Dataset already exists, skipping generation")
+
+if __name__ == "__main__":
+    # main(prompt_path="VBench_full_info.json")
+    main(prompt_path="t2v_vbench_remain_200.json")

src/videogen_hub/benchmark/t2v_vbench_200.json

@@ -0,0 +1,1956 @@
+{
+    "0_a_toilet,_frozen_in_time.mp4": {
+        "prompt_en": "a toilet, frozen in time",
+        "dimension": [
+            "temporal_flickering"
+        ]
+    },
+    "1_a_laptop,_frozen_in_time.mp4": {
+        "prompt_en": "a laptop, frozen in time",
+        "dimension": [
+            "temporal_flickering"
+        ]
+    },
+    "2_In_a_still_frame,_phone_booth.mp4": {
+        "prompt_en": "In a still frame, phone booth",
+        "dimension": [
+            "temporal_flickering"
+        ]
+    },
+    "3_A_tranquil_tableau_of_an_apple.mp4": {
+        "prompt_en": "A tranquil tableau of an apple",
+        "dimension": [
+            "temporal_flickering"
+        ]
+    },
+    "4_A_tranquil_tableau_of_a_bench.mp4": {
+        "prompt_en": "A tranquil tableau of a bench",
+        "dimension": [
+            "temporal_flickering"
+        ]
+    },
+    "5_A_tranquil_tableau_of_an_exquisite_mahogany_dining_table.mp4": {
+        "prompt_en": "A tranquil tableau of an exquisite mahogany dining table",
+        "dimension": [
+            "temporal_flickering"
+        ]
+    },
+    "6_A_tranquil_tableau_of_the_jail_cell_was_small_and_dimly_lit,_with_cold,_steel_bars.mp4": {
+        "prompt_en": "A tranquil tableau of the jail cell was small and dimly lit, with cold, steel bars",
+        "dimension": [
+            "temporal_flickering"
+        ]
+    },
+    "7_In_a_still_frame,_the_Temple_of_Hephaestus,_with_its_timeless_Doric_grace,_stands_stoically_against_the_backdrop_of_a_quiet_Athens.mp4": {
+        "prompt_en": "In a still frame, the Temple of Hephaestus, with its timeless Doric grace, stands stoically against the backdrop of a quiet Athens",
+        "dimension": [
+            "temporal_flickering"
+        ]
+    },
+    "8_In_a_still_frame,_the_ornate_Victorian_streetlamp_stands_solemnly,_adorned_with_intricate_ironwork_and_stained_glass_panels.mp4": {
+        "prompt_en": "In a still frame, the ornate Victorian streetlamp stands solemnly, adorned with intricate ironwork and stained glass panels",
+        "dimension": [
+            "temporal_flickering"
+        ]
+    },
+    "9_A_tranquil_tableau_of_in_the_quaint_village_square,_a_traditional_wrought-iron_streetlamp_featured_delicate_filigree_patterns_and_amber-hued_glass_panels.mp4": {
+        "prompt_en": "A tranquil tableau of in the quaint village square, a traditional wrought-iron streetlamp featured delicate filigree patterns and amber-hued glass panels",
+        "dimension": [
+            "temporal_flickering"
+        ]
+    },
+    "10_In_a_still_frame,_in_the_heart_of_the_old_city,_a_row_of_ornate_lantern-style_streetlamps_bathed_the_narrow_alleyway_in_a_warm,_welcoming_light.mp4": {
+        "prompt_en": "In a still frame, in the heart of the old city, a row of ornate lantern-style streetlamps bathed the narrow alleyway in a warm, welcoming light",
+        "dimension": [
+            "temporal_flickering"
+        ]
+    },
+    "11_In_a_still_frame,_a_tranquil_pond_was_fringed_by_weeping_cherry_trees,_their_blossoms_drifting_lazily_onto_the_glassy_surface.mp4": {
+        "prompt_en": "In a still frame, a tranquil pond was fringed by weeping cherry trees, their blossoms drifting lazily onto the glassy surface",
+        "dimension": [
+            "temporal_flickering"
+        ]
+    },
+    "12_a_bird_and_a_cat.mp4": {
+        "prompt_en": "a bird and a cat",
+        "dimension": [
+            "multiple_objects"
+        ],
+        "auxiliary_info": {
+            "multiple_objects": {
+                "object": "bird and cat"
+            }
+        }
+    },
+    "13_a_cat_and_a_dog.mp4": {
+        "prompt_en": "a cat and a dog",
+        "dimension": [
+            "multiple_objects"
+        ],
+        "auxiliary_info": {
+            "multiple_objects": {
+                "object": "cat and dog"
+            }
+        }
+    },
+    "14_a_sheep_and_a_cow.mp4": {
+        "prompt_en": "a sheep and a cow",
+        "dimension": [
+            "multiple_objects"
+        ],
+        "auxiliary_info": {
+            "multiple_objects": {
+                "object": "sheep and cow"
+            }
+        }
+    },
+    "15_an_elephant_and_a_bear.mp4": {
+        "prompt_en": "an elephant and a bear",
+        "dimension": [
+            "multiple_objects"
+        ],
+        "auxiliary_info": {
+            "multiple_objects": {
+                "object": "elephant and bear"
+            }
+        }
+    },
+    "16_a_giraffe_and_a_bird.mp4": {
+        "prompt_en": "a giraffe and a bird",
+        "dimension": [
+            "multiple_objects"
+        ],
+        "auxiliary_info": {
+            "multiple_objects": {
+                "object": "giraffe and bird"
+            }
+        }
+    },
+    "17_a_couch_and_a_potted_plant.mp4": {
+        "prompt_en": "a couch and a potted plant",
+        "dimension": [
+            "multiple_objects"
+        ],
+        "auxiliary_info": {
+            "multiple_objects": {
+                "object": "couch and potted plant"
+            }
+        }
+    },
+    "18_a_laptop_and_a_remote.mp4": {
+        "prompt_en": "a laptop and a remote",
+        "dimension": [
+            "multiple_objects"
+        ],
+        "auxiliary_info": {
+            "multiple_objects": {
+                "object": "laptop and remote"
+            }
+        }
+    },
+    "19_a_clock_and_a_backpack.mp4": {
+        "prompt_en": "a clock and a backpack",
+        "dimension": [
+            "multiple_objects"
+        ],
+        "auxiliary_info": {
+            "multiple_objects": {
+                "object": "clock and backpack"
+            }
+        }
+    },
+    "20_a_vase_and_scissors.mp4": {
+        "prompt_en": "a vase and scissors",
+        "dimension": [
+            "multiple_objects"
+        ],
+        "auxiliary_info": {
+            "multiple_objects": {
+                "object": "vase and scissors"
+            }
+        }
+    },
+    "21_a_teddy_bear_and_a_frisbee.mp4": {
+        "prompt_en": "a teddy bear and a frisbee",
+        "dimension": [
+            "multiple_objects"
+        ],
+        "auxiliary_info": {
+            "multiple_objects": {
+                "object": "teddy bear and frisbee"
+            }
+        }
+    },
+    "22_skis_and_a_snowboard.mp4": {
+        "prompt_en": "skis and a snowboard",
+        "dimension": [
+            "multiple_objects"
+        ],
+        "auxiliary_info": {
+            "multiple_objects": {
+                "object": "skis and snowboard"
+            }
+        }
+    },
+    "23_a_bottle_and_a_chair.mp4": {
+        "prompt_en": "a bottle and a chair",
+        "dimension": [
+            "multiple_objects"
+        ],
+        "auxiliary_info": {
+            "multiple_objects": {
+                "object": "bottle and chair"
+            }
+        }
+    },
+    "24_a_bicycle_and_a_car.mp4": {
+        "prompt_en": "a bicycle and a car",
+        "dimension": [
+            "multiple_objects"
+        ],
+        "auxiliary_info": {
+            "multiple_objects": {
+                "object": "bicycle and car"
+            }
+        }
+    },
+    "25_a_bowl_and_a_remote.mp4": {
+        "prompt_en": "a bowl and a remote",
+        "dimension": [
+            "multiple_objects"
+        ],
+        "auxiliary_info": {
+            "multiple_objects": {
+                "object": "bowl and remote"
+            }
+        }
+    },
+    "26_broccoli_and_a_backpack.mp4": {
+        "prompt_en": "broccoli and a backpack",
+        "dimension": [
+            "multiple_objects"
+        ],
+        "auxiliary_info": {
+            "multiple_objects": {
+                "object": "broccoli and backpack"
+            }
+        }
+    },
+    "27_a_refrigerator_and_skis.mp4": {
+        "prompt_en": "a refrigerator and skis",
+        "dimension": [
+            "multiple_objects"
+        ],
+        "auxiliary_info": {
+            "multiple_objects": {
+                "object": "refrigerator and skis"
+            }
+        }
+    },
+    "28_A_person_is_riding_a_bike.mp4": {
+        "prompt_en": "A person is riding a bike",
+        "dimension": [
+            "human_action"
+        ]
+    },
+    "29_A_person_is_marching.mp4": {
+        "prompt_en": "A person is marching",
+        "dimension": [
+            "human_action"
+        ]
+    },
+    "30_A_person_is_playing_harp.mp4": {
+        "prompt_en": "A person is playing harp",
+        "dimension": [
+            "human_action"
+        ]
+    },
+    "31_A_person_is_wrestling.mp4": {
+        "prompt_en": "A person is wrestling",
+        "dimension": [
+            "human_action"
+        ]
+    },
+    "32_A_person_is_sweeping_floor.mp4": {
+        "prompt_en": "A person is sweeping floor",
+        "dimension": [
+            "human_action"
+        ]
+    },
+    "33_A_person_is_push_up.mp4": {
+        "prompt_en": "A person is push up",
+        "dimension": [
+            "human_action"
+        ]
+    },
+    "34_A_person_is_filling_eyebrows.mp4": {
+        "prompt_en": "A person is filling eyebrows",
+        "dimension": [
+            "human_action"
+        ]
+    },
+    "35_A_person_is_air_drumming.mp4": {
+        "prompt_en": "A person is air drumming",
+        "dimension": [
+            "human_action"
+        ]
+    },
+    "36_A_person_is_rock_climbing.mp4": {
+        "prompt_en": "A person is rock climbing",
+        "dimension": [
+            "human_action"
+        ]
+    },
+    "37_A_person_is_hula_hooping.mp4": {
+        "prompt_en": "A person is hula hooping",
+        "dimension": [
+            "human_action"
+        ]
+    },
+    "38_A_person_is_crawling_baby.mp4": {
+        "prompt_en": "A person is crawling baby",
+        "dimension": [
+            "human_action"
+        ]
+    },
+    "39_A_person_is_motorcycling.mp4": {
+        "prompt_en": "A person is motorcycling",
+        "dimension": [
+            "human_action"
+        ]
+    },
+    "40_A_person_is_riding_or_walking_with_horse.mp4": {
+        "prompt_en": "A person is riding or walking with horse",
+        "dimension": [
+            "human_action"
+        ]
+    },
+    "41_A_person_is_using_computer.mp4": {
+        "prompt_en": "A person is using computer",
+        "dimension": [
+            "human_action"
+        ]
+    },
+    "42_A_person_is_arranging_flowers.mp4": {
+        "prompt_en": "A person is arranging flowers",
+        "dimension": [
+            "human_action"
+        ]
+    },
+    "43_A_person_is_ice_skating.mp4": {
+        "prompt_en": "A person is ice skating",
+        "dimension": [
+            "human_action"
+        ]
+    },
+    "44_A_person_is_barbequing.mp4": {
+        "prompt_en": "A person is barbequing",
+        "dimension": [
+            "human_action"
+        ]
+    },
+    "45_a_person_swimming_in_ocean.mp4": {
+        "prompt_en": "a person swimming in ocean",
+        "dimension": [
+            "subject_consistency",
+            "dynamic_degree",
+            "motion_smoothness"
+        ]
+    },
+    "46_a_car_turning_a_corner.mp4": {
+        "prompt_en": "a car turning a corner",
+        "dimension": [
+            "subject_consistency",
+            "dynamic_degree",
+            "motion_smoothness"
+        ]
+    },
+    "47_a_dog_enjoying_a_peaceful_walk.mp4": {
+        "prompt_en": "a dog enjoying a peaceful walk",
+        "dimension": [
+            "subject_consistency",
+            "dynamic_degree",
+            "motion_smoothness"
+        ]
+    },
+    "48_a_dog_playing_in_park.mp4": {
+        "prompt_en": "a dog playing in park",
+        "dimension": [
+            "subject_consistency",
+            "dynamic_degree",
+            "motion_smoothness"
+        ]
+    },
+    "49_a_horse_galloping_across_an_open_field.mp4": {
+        "prompt_en": "a horse galloping across an open field",
+        "dimension": [
+            "subject_consistency",
+            "dynamic_degree",
+            "motion_smoothness"
+        ]
+    },
+    "50_a_sheep_bending_down_to_drink_water_from_a_river.mp4": {
+        "prompt_en": "a sheep bending down to drink water from a river",
+        "dimension": [
+            "subject_consistency",
+            "dynamic_degree",
+            "motion_smoothness"
+        ]
+    },
+    "51_a_cow_bending_down_to_drink_water_from_a_river.mp4": {
+        "prompt_en": "a cow bending down to drink water from a river",
+        "dimension": [
+            "subject_consistency",
+            "dynamic_degree",
+            "motion_smoothness"
+        ]
+    },
+    "52_a_cow_running_to_join_a_herd_of_its_kind.mp4": {
+        "prompt_en": "a cow running to join a herd of its kind",
+        "dimension": [
+            "subject_consistency",
+            "dynamic_degree",
+            "motion_smoothness"
+        ]
+    },
+    "53_a_car_slowing_down_to_stop.mp4": {
+        "prompt_en": "a car slowing down to stop",
+        "dimension": [
+            "subject_consistency",
+            "dynamic_degree",
+            "motion_smoothness"
+        ]
+    },
+    "54_a_car_accelerating_to_gain_speed.mp4": {
+        "prompt_en": "a car accelerating to gain speed",
+        "dimension": [
+            "subject_consistency",
+            "dynamic_degree",
+            "motion_smoothness"
+        ]
+    },
+    "55_a_motorcycle_slowing_down_to_stop.mp4": {
+        "prompt_en": "a motorcycle slowing down to stop",
+        "dimension": [
+            "subject_consistency",
+            "dynamic_degree",
+            "motion_smoothness"
+        ]
+    },
+    "56_a_bus_turning_a_corner.mp4": {
+        "prompt_en": "a bus turning a corner",
+        "dimension": [
+            "subject_consistency",
+            "dynamic_degree",
+            "motion_smoothness"
+        ]
+    },
+    "57_a_bird_soaring_gracefully_in_the_sky.mp4": {
+        "prompt_en": "a bird soaring gracefully in the sky",
+        "dimension": [
+            "subject_consistency",
+            "dynamic_degree",
+            "motion_smoothness"
+        ]
+    },
+    "58_a_dog_enjoying_a_peaceful_walk.mp4": {
+        "prompt_en": "a dog enjoying a peaceful walk",
+        "dimension": [
+            "subject_consistency",
+            "dynamic_degree",
+            "motion_smoothness"
+        ]
+    },
+    "59_a_dog_playing_in_park.mp4": {
+        "prompt_en": "a dog playing in park",
+        "dimension": [
+            "subject_consistency",
+            "dynamic_degree",
+            "motion_smoothness"
+        ]
+    },
+    "60_a_dog_drinking_water.mp4": {
+        "prompt_en": "a dog drinking water",
+        "dimension": [
+            "subject_consistency",
+            "dynamic_degree",
+            "motion_smoothness"
+        ]
+    },
+    "61_a_horse_bending_down_to_drink_water_from_a_river.mp4": {
+        "prompt_en": "a horse bending down to drink water from a river",
+        "dimension": [
+            "subject_consistency",
+            "dynamic_degree",
+            "motion_smoothness"
+        ]
+    },
+    "62_a_horse_galloping_across_an_open_field.mp4": {
+        "prompt_en": "a horse galloping across an open field",
+        "dimension": [
+            "subject_consistency",
+            "dynamic_degree",
+            "motion_smoothness"
+        ]
+    },
+    "63_a_bear_sniffing_the_air_for_scents_of_food.mp4": {
+        "prompt_en": "a bear sniffing the air for scents of food",
+        "dimension": [
+            "subject_consistency",
+            "dynamic_degree",
+            "motion_smoothness"
+        ]
+    },
+    "64_a_zebra_taking_a_peaceful_walk.mp4": {
+        "prompt_en": "a zebra taking a peaceful walk",
+        "dimension": [
+            "subject_consistency",
+            "dynamic_degree",
+            "motion_smoothness"
+        ]
+    },
+    "65_a_person_eating_a_burger.mp4": {
+        "prompt_en": "a person eating a burger",
+        "dimension": [
+            "subject_consistency",
+            "dynamic_degree",
+            "motion_smoothness"
+        ]
+    },
+    "66_a_car_stuck_in_traffic_during_rush_hour.mp4": {
+        "prompt_en": "a car stuck in traffic during rush hour",
+        "dimension": [
+            "subject_consistency",
+            "dynamic_degree",
+            "motion_smoothness"
+        ]
+    },
+    "67_a_motorcycle_turning_a_corner.mp4": {
+        "prompt_en": "a motorcycle turning a corner",
+        "dimension": [
+            "subject_consistency",
+            "dynamic_degree",
+            "motion_smoothness"
+        ]
+    },
+    "68_an_airplane_taking_off.mp4": {
+        "prompt_en": "an airplane taking off",
+        "dimension": [
+            "subject_consistency",
+            "dynamic_degree",
+            "motion_smoothness"
+        ]
+    },
+    "69_a_truck_anchored_in_a_tranquil_bay.mp4": {
+        "prompt_en": "a truck anchored in a tranquil bay",
+        "dimension": [
+            "subject_consistency",
+            "dynamic_degree",
+            "motion_smoothness"
+        ]
+    },
+    "70_a_truck_slowing_down_to_stop.mp4": {
+        "prompt_en": "a truck slowing down to stop",
+        "dimension": [
+            "subject_consistency",
+            "dynamic_degree",
+            "motion_smoothness"
+        ]
+    },
+    "71_a_bird_building_a_nest_from_twigs_and_leaves.mp4": {
+        "prompt_en": "a bird building a nest from twigs and leaves",
+        "dimension": [
+            "subject_consistency",
+            "dynamic_degree",
+            "motion_smoothness"
+        ]
+    },
+    "72_a_bird_flying_over_a_snowy_forest.mp4": {
+        "prompt_en": "a bird flying over a snowy forest",
+        "dimension": [
+            "subject_consistency",
+            "dynamic_degree",
+            "motion_smoothness"
+        ]
+    },
+    "73_a_zebra_taking_a_peaceful_walk.mp4": {
+        "prompt_en": "a zebra taking a peaceful walk",
+        "dimension": [
+            "subject_consistency",
+            "dynamic_degree",
+            "motion_smoothness"
+        ]
+    },
+    "74_a_train.mp4": {
+        "prompt_en": "a train",
+        "dimension": [
+            "object_class"
+        ],
+        "auxiliary_info": {
+            "object_class": {
+                "object": "train"
+            }
+        }
+    },
+    "75_a_cat.mp4": {
+        "prompt_en": "a cat",
+        "dimension": [
+            "object_class"
+        ],
+        "auxiliary_info": {
+            "object_class": {
+                "object": "cat"
+            }
+        }
+    },
+    "76_an_elephant.mp4": {
+        "prompt_en": "an elephant",
+        "dimension": [
+            "object_class"
+        ],
+        "auxiliary_info": {
+            "object_class": {
+                "object": "elephant"
+            }
+        }
+    },
+    "77_a_suitcase.mp4": {
+        "prompt_en": "a suitcase",
+        "dimension": [
+            "object_class"
+        ],
+        "auxiliary_info": {
+            "object_class": {
+                "object": "suitcase"
+            }
+        }
+    },
+    "78_an_orange.mp4": {
+        "prompt_en": "an orange",
+        "dimension": [
+            "object_class"
+        ],
+        "auxiliary_info": {
+            "object_class": {
+                "object": "orange"
+            }
+        }
+    },
+    "79_a_hot_dog.mp4": {
+        "prompt_en": "a hot dog",
+        "dimension": [
+            "object_class"
+        ],
+        "auxiliary_info": {
+            "object_class": {
+                "object": "hot dog"
+            }
+        }
+    },
+    "80_a_keyboard.mp4": {
+        "prompt_en": "a keyboard",
+        "dimension": [
+            "object_class"
+        ],
+        "auxiliary_info": {
+            "object_class": {
+                "object": "keyboard"
+            }
+        }
+    },
+    "81_a_sink.mp4": {
+        "prompt_en": "a sink",
+        "dimension": [
+            "object_class"
+        ],
+        "auxiliary_info": {
+            "object_class": {
+                "object": "sink"
+            }
+        }
+    },
+    "82_a_toothbrush.mp4": {
+        "prompt_en": "a toothbrush",
+        "dimension": [
+            "object_class"
+        ],
+        "auxiliary_info": {
+            "object_class": {
+                "object": "toothbrush"
+            }
+        }
+    },
+    "83_a_red_bicycle.mp4": {
+        "prompt_en": "a red bicycle",
+        "dimension": [
+            "color"
+        ],
+        "auxiliary_info": {
+            "color": {
+                "color": "red"
+            }
+        }
+    },
+    "84_a_green_bicycle.mp4": {
+        "prompt_en": "a green bicycle",
+        "dimension": [
+            "color"
+        ],
+        "auxiliary_info": {
+            "color": {
+                "color": "green"
+            }
+        }
+    },
+    "85_a_yellow_bicycle.mp4": {
+        "prompt_en": "a yellow bicycle",
+        "dimension": [
+            "color"
+        ],
+        "auxiliary_info": {
+            "color": {
+                "color": "yellow"
+            }
+        }
+    },
+    "86_a_black_bicycle.mp4": {
+        "prompt_en": "a black bicycle",
+        "dimension": [
+            "color"
+        ],
+        "auxiliary_info": {
+            "color": {
+                "color": "black"
+            }
+        }
+    },
+    "87_a_purple_bird.mp4": {
+        "prompt_en": "a purple bird",
+        "dimension": [
+            "color"
+        ],
+        "auxiliary_info": {
+            "color": {
+                "color": "purple"
+            }
+        }
+    },
+    "88_a_yellow_cat.mp4": {
+        "prompt_en": "a yellow cat",
+        "dimension": [
+            "color"
+        ],
+        "auxiliary_info": {
+            "color": {
+                "color": "yellow"
+            }
+        }
+    },
+    "89_a_pink_suitcase.mp4": {
+        "prompt_en": "a pink suitcase",
+        "dimension": [
+            "color"
+        ],
+        "auxiliary_info": {
+            "color": {
+                "color": "pink"
+            }
+        }
+    },
+    "90_a_purple_chair.mp4": {
+        "prompt_en": "a purple chair",
+        "dimension": [
+            "color"
+        ],
+        "auxiliary_info": {
+            "color": {
+                "color": "purple"
+            }
+        }
+    },
+    "91_a_green_clock.mp4": {
+        "prompt_en": "a green clock",
+        "dimension": [
+            "color"
+        ],
+        "auxiliary_info": {
+            "color": {
+                "color": "green"
+            }
+        }
+    },
+    "92_a_yellow_clock.mp4": {
+        "prompt_en": "a yellow clock",
+        "dimension": [
+            "color"
+        ],
+        "auxiliary_info": {
+            "color": {
+                "color": "yellow"
+            }
+        }
+    },
+    "93_a_purple_clock.mp4": {
+        "prompt_en": "a purple clock",
+        "dimension": [
+            "color"
+        ],
+        "auxiliary_info": {
+            "color": {
+                "color": "purple"
+            }
+        }
+    },
+    "94_a_green_vase.mp4": {
+        "prompt_en": "a green vase",
+        "dimension": [
+            "color"
+        ],
+        "auxiliary_info": {
+            "color": {
+                "color": "green"
+            }
+        }
+    },
+    "95_The_bund_Shanghai,_watercolor_painting.mp4": {
+        "prompt_en": "The bund Shanghai, watercolor painting",
+        "dimension": [
+            "appearance_style"
+        ],
+        "auxiliary_info": {
+            "appearance_style": {
+                "appearance_style": "watercolor painting"
+            }
+        }
+    },
+    "96_a_shark_is_swimming_in_the_ocean_by_Hokusai,_in_the_style_of_Ukiyo.mp4": {
+        "prompt_en": "a shark is swimming in the ocean by Hokusai, in the style of Ukiyo",
+        "dimension": [
+            "appearance_style"
+        ],
+        "auxiliary_info": {
+            "appearance_style": {
+                "appearance_style": "by Hokusai, in the style of Ukiyo"
+            }
+        }
+    },
+    "97_a_shark_is_swimming_in_the_ocean,_pixel_art.mp4": {
+        "prompt_en": "a shark is swimming in the ocean, pixel art",
+        "dimension": [
+            "appearance_style"
+        ],
+        "auxiliary_info": {
+            "appearance_style": {
+                "appearance_style": "pixel art"
+            }
+        }
+    },
+    "98_Gwen_Stacy_reading_a_book,_pixel_art.mp4": {
+        "prompt_en": "Gwen Stacy reading a book, pixel art",
+        "dimension": [
+            "appearance_style"
+        ],
+        "auxiliary_info": {
+            "appearance_style": {
+                "appearance_style": "pixel art"
+            }
+        }
+    },
+    "99_A_boat_sailing_leisurely_along_the_Seine_River_with_the_Eiffel_Tower_in_background,_pixel_art.mp4": {
+        "prompt_en": "A boat sailing leisurely along the Seine River with the Eiffel Tower in background, pixel art",
+        "dimension": [
+            "appearance_style"
+        ],
+        "auxiliary_info": {
+            "appearance_style": {
+                "appearance_style": "pixel art"
+            }
+        }
+    },
+    "100_A_couple_in_formal_evening_wear_going_home_get_caught_in_a_heavy_downpour_with_umbrellas_by_Hokusai,_in_the_style_of_Ukiyo.mp4": {
+        "prompt_en": "A couple in formal evening wear going home get caught in a heavy downpour with umbrellas by Hokusai, in the style of Ukiyo",
+        "dimension": [
+            "appearance_style"
+        ],
+        "auxiliary_info": {
+            "appearance_style": {
+                "appearance_style": "by Hokusai, in the style of Ukiyo"
+            }
+        }
+    },
+    "101_An_astronaut_flying_in_space,_Van_Gogh_style.mp4": {
+        "prompt_en": "An astronaut flying in space, Van Gogh style",
+        "dimension": [
+            "appearance_style"
+        ],
+        "auxiliary_info": {
+            "appearance_style": {
+                "appearance_style": "Van Gogh style"
+            }
+        }
+    },
+    "102_An_astronaut_flying_in_space,_oil_painting.mp4": {
+        "prompt_en": "An astronaut flying in space, oil painting",
+        "dimension": [
+            "appearance_style"
+        ],
+        "auxiliary_info": {
+            "appearance_style": {
+                "appearance_style": "oil painting"
+            }
+        }
+    },
+    "103_An_astronaut_flying_in_space_by_Hokusai,_in_the_style_of_Ukiyo.mp4": {
+        "prompt_en": "An astronaut flying in space by Hokusai, in the style of Ukiyo",
+        "dimension": [
+            "appearance_style"
+        ],
+        "auxiliary_info": {
+            "appearance_style": {
+                "appearance_style": "by Hokusai, in the style of Ukiyo"
+            }
+        }
+    },
+    "104_An_astronaut_flying_in_space,_in_cyberpunk_style.mp4": {
+        "prompt_en": "An astronaut flying in space, in cyberpunk style",
+        "dimension": [
+            "appearance_style"
+        ],
+        "auxiliary_info": {
+            "appearance_style": {
+                "appearance_style": "in cyberpunk style"
+            }
+        }
+    },
+    "105_Snow_rocky_mountains_peaks_canyon._snow_blanketed_rocky_mountains_surround_and_shadow_deep_canyons._the_canyons_twist_and_bend_through_the_high_elevated_mountain_peaks,_watercolor_painting.mp4": {
+        "prompt_en": "Snow rocky mountains peaks canyon. snow blanketed rocky mountains surround and shadow deep canyons. the canyons twist and bend through the high elevated mountain peaks, watercolor painting",
+        "dimension": [
+            "appearance_style"
+        ],
+        "auxiliary_info": {
+            "appearance_style": {
+                "appearance_style": "watercolor painting"
+            }
+        }
+    },
+    "106_A_beautiful_coastal_beach_in_spring,_waves_lapping_on_sand,_zoom_out.mp4": {
+        "prompt_en": "A beautiful coastal beach in spring, waves lapping on sand, zoom out",
+        "dimension": [
+            "temporal_style"
+        ]
+    },
+    "107_A_beautiful_coastal_beach_in_spring,_waves_lapping_on_sand,_pan_right.mp4": {
+        "prompt_en": "A beautiful coastal beach in spring, waves lapping on sand, pan right",
+        "dimension": [
+            "temporal_style"
+        ]
+    },
+    "108_A_beautiful_coastal_beach_in_spring,_waves_lapping_on_sand,_tilt_up.mp4": {
+        "prompt_en": "A beautiful coastal beach in spring, waves lapping on sand, tilt up",
+        "dimension": [
+            "temporal_style"
+        ]
+    },
+    "109_The_bund_Shanghai,_zoom_out.mp4": {
+        "prompt_en": "The bund Shanghai, zoom out",
+        "dimension": [
+            "temporal_style"
+        ]
+    },
+    "110_a_shark_is_swimming_in_the_ocean,_featuring_a_steady_and_smooth_perspective.mp4": {
+        "prompt_en": "a shark is swimming in the ocean, featuring a steady and smooth perspective",
+        "dimension": [
+            "temporal_style"
+        ]
+    },
+    "111_A_panda_drinking_coffee_in_a_cafe_in_Paris,_in_super_slow_motion.mp4": {
+        "prompt_en": "A panda drinking coffee in a cafe in Paris, in super slow motion",
+        "dimension": [
+            "temporal_style"
+        ]
+    },
+    "112_Gwen_Stacy_reading_a_book,_zoom_out.mp4": {
+        "prompt_en": "Gwen Stacy reading a book, zoom out",
+        "dimension": [
+            "temporal_style"
+        ]
+    },
+    "113_Gwen_Stacy_reading_a_book,_featuring_a_steady_and_smooth_perspective.mp4": {
+        "prompt_en": "Gwen Stacy reading a book, featuring a steady and smooth perspective",
+        "dimension": [
+            "temporal_style"
+        ]
+    },
+    "114_A_couple_in_formal_evening_wear_going_home_get_caught_in_a_heavy_downpour_with_umbrellas,_tilt_down.mp4": {
+        "prompt_en": "A couple in formal evening wear going home get caught in a heavy downpour with umbrellas, tilt down",
+        "dimension": [
+            "temporal_style"
+        ]
+    },
+    "115_Snow_rocky_mountains_peaks_canyon._snow_blanketed_rocky_mountains_surround_and_shadow_deep_canyons._the_canyons_twist_and_bend_through_the_high_elevated_mountain_peaks,_tilt_up.mp4": {
+        "prompt_en": "Snow rocky mountains peaks canyon. snow blanketed rocky mountains surround and shadow deep canyons. the canyons twist and bend through the high elevated mountain peaks, tilt up",
+        "dimension": [
+            "temporal_style"
+        ]
+    },
+    "116_Snow_rocky_mountains_peaks_canyon._snow_blanketed_rocky_mountains_surround_and_shadow_deep_canyons._the_canyons_twist_and_bend_through_the_high_elevated_mountain_peaks,_racking_focus.mp4": {
+        "prompt_en": "Snow rocky mountains peaks canyon. snow blanketed rocky mountains surround and shadow deep canyons. the canyons twist and bend through the high elevated mountain peaks, racking focus",
+        "dimension": [
+            "temporal_style"
+        ]
+    },
+    "117_Flying_through_fantasy_landscapes..mp4": {
+        "prompt_en": "Flying through fantasy landscapes.",
+        "dimension": [
+            "overall_consistency",
+            "aesthetic_quality",
+            "imaging_quality"
+        ]
+    },
+    "118_Aerial_panoramic_video_from_a_drone_of_a_fantasy_land..mp4": {
+        "prompt_en": "Aerial panoramic video from a drone of a fantasy land.",
+        "dimension": [
+            "overall_consistency",
+            "aesthetic_quality",
+            "imaging_quality"
+        ]
+    },
+    "119_Balloon_full_of_water_exploding_in_extreme_slow_motion..mp4": {
+        "prompt_en": "Balloon full of water exploding in extreme slow motion.",
+        "dimension": [
+            "overall_consistency",
+            "aesthetic_quality",
+            "imaging_quality"
+        ]
+    },
+    "120_Few_big_purple_plums_rotating_on_the_turntable._water_drops_appear_on_the_skin_during_rotation._isolated_on_the_white_background._close-up._macro..mp4": {
+        "prompt_en": "Few big purple plums rotating on the turntable. water drops appear on the skin during rotation. isolated on the white background. close-up. macro.",
+        "dimension": [
+            "overall_consistency",
+            "aesthetic_quality",
+            "imaging_quality"
+        ]
+    },
+    "121_A_fantasy_landscape.mp4": {
+        "prompt_en": "A fantasy landscape",
+        "dimension": [
+            "overall_consistency",
+            "aesthetic_quality",
+            "imaging_quality"
+        ]
+    },
+    "122_A_steam_train_moving_on_a_mountainside.mp4": {
+        "prompt_en": "A steam train moving on a mountainside",
+        "dimension": [
+            "overall_consistency",
+            "aesthetic_quality",
+            "imaging_quality"
+        ]
+    },
+    "123_A_beautiful_coastal_beach_in_spring,_waves_lapping_on_sand_by_Hokusai,_in_the_style_of_Ukiyo.mp4": {
+        "prompt_en": "A beautiful coastal beach in spring, waves lapping on sand by Hokusai, in the style of Ukiyo",
+        "dimension": [
+            "overall_consistency",
+            "aesthetic_quality",
+            "imaging_quality"
+        ]
+    },
+    "124_A_polar_bear_is_playing_guitar.mp4": {
+        "prompt_en": "A polar bear is playing guitar",
+        "dimension": [
+            "overall_consistency",
+            "aesthetic_quality",
+            "imaging_quality"
+        ]
+    },
+    "125_An_astronaut_feeding_ducks_on_a_sunny_afternoon,_reflection_from_the_water..mp4": {
+        "prompt_en": "An astronaut feeding ducks on a sunny afternoon, reflection from the water.",
+        "dimension": [
+            "overall_consistency",
+            "aesthetic_quality",
+            "imaging_quality"
+        ]
+    },
+    "126_Sunset_time_lapse_at_the_beach_with_moving_clouds_and_colors_in_the_sky..mp4": {
+        "prompt_en": "Sunset time lapse at the beach with moving clouds and colors in the sky.",
+        "dimension": [
+            "overall_consistency",
+            "aesthetic_quality",
+            "imaging_quality"
+        ]
+    },
+    "127_Flying_through_fantasy_landscapes..mp4": {
+        "prompt_en": "Flying through fantasy landscapes.",
+        "dimension": [
+            "overall_consistency",
+            "aesthetic_quality",
+            "imaging_quality"
+        ]
+    },
+    "128_A_squirrel_eating_a_burger..mp4": {
+        "prompt_en": "A squirrel eating a burger.",
+        "dimension": [
+            "overall_consistency",
+            "aesthetic_quality",
+            "imaging_quality"
+        ]
+    },
+    "129_A_drone_view_of_celebration_with_Christmas_tree_and_fireworks,_starry_sky_-_background..mp4": {
+        "prompt_en": "A drone view of celebration with Christmas tree and fireworks, starry sky - background.",
+        "dimension": [
+            "overall_consistency",
+            "aesthetic_quality",
+            "imaging_quality"
+        ]
+    },
+    "130_Robot_dancing_in_Times_Square..mp4": {
+        "prompt_en": "Robot dancing in Times Square.",
+        "dimension": [
+            "overall_consistency",
+            "aesthetic_quality",
+            "imaging_quality"
+        ]
+    },
+    "131_Few_big_purple_plums_rotating_on_the_turntable._water_drops_appear_on_the_skin_during_rotation._isolated_on_the_white_background._close-up._macro..mp4": {
+        "prompt_en": "Few big purple plums rotating on the turntable. water drops appear on the skin during rotation. isolated on the white background. close-up. macro.",
+        "dimension": [
+            "overall_consistency",
+            "aesthetic_quality",
+            "imaging_quality"
+        ]
+    },
+    "132_Ashtray_full_of_butts_on_table,_smoke_flowing_on_black_background,_close-up.mp4": {
+        "prompt_en": "Ashtray full of butts on table, smoke flowing on black background, close-up",
+        "dimension": [
+            "overall_consistency",
+            "aesthetic_quality",
+            "imaging_quality"
+        ]
+    },
+    "133_A_future_where_humans_have_achieved_teleportation_technology.mp4": {
+        "prompt_en": "A future where humans have achieved teleportation technology",
+        "dimension": [
+            "overall_consistency",
+            "aesthetic_quality",
+            "imaging_quality"
+        ]
+    },
+    "134_Gwen_Stacy_reading_a_book.mp4": {
+        "prompt_en": "Gwen Stacy reading a book",
+        "dimension": [
+            "overall_consistency",
+            "aesthetic_quality",
+            "imaging_quality"
+        ]
+    },
+    "135_Yoda_playing_guitar_on_the_stage.mp4": {
+        "prompt_en": "Yoda playing guitar on the stage",
+        "dimension": [
+            "overall_consistency",
+            "aesthetic_quality",
+            "imaging_quality"
+        ]
+    },
+    "136_A_cat_eating_food_out_of_a_bowl.mp4": {
+        "prompt_en": "A cat eating food out of a bowl",
+        "dimension": [
+            "overall_consistency",
+            "aesthetic_quality",
+            "imaging_quality"
+        ]
+    },
+    "137_A_cute_raccoon_playing_guitar_in_a_boat_on_the_ocean.mp4": {
+        "prompt_en": "A cute raccoon playing guitar in a boat on the ocean",
+        "dimension": [
+            "overall_consistency",
+            "aesthetic_quality",
+            "imaging_quality"
+        ]
+    },
+    "138_A_happy_fuzzy_panda_playing_guitar_nearby_a_campfire,_snow_mountain_in_the_background.mp4": {
+        "prompt_en": "A happy fuzzy panda playing guitar nearby a campfire, snow mountain in the background",
+        "dimension": [
+            "overall_consistency",
+            "aesthetic_quality",
+            "imaging_quality"
+        ]
+    },
+    "139_A_polar_bear_is_playing_guitar.mp4": {
+        "prompt_en": "A polar bear is playing guitar",
+        "dimension": [
+            "overall_consistency",
+            "aesthetic_quality",
+            "imaging_quality"
+        ]
+    },
+    "140_A_bigfoot_walking_in_the_snowstorm..mp4": {
+        "prompt_en": "A bigfoot walking in the snowstorm.",
+        "dimension": [
+            "overall_consistency",
+            "aesthetic_quality",
+            "imaging_quality"
+        ]
+    },
+    "141_A_drone_view_of_celebration_with_Christmas_tree_and_fireworks,_starry_sky_-_background..mp4": {
+        "prompt_en": "A drone view of celebration with Christmas tree and fireworks, starry sky - background.",
+        "dimension": [
+            "overall_consistency",
+            "aesthetic_quality",
+            "imaging_quality"
+        ]
+    },
+    "142_An_astronaut_is_riding_a_horse_in_the_space_in_a_photorealistic_style..mp4": {
+        "prompt_en": "An astronaut is riding a horse in the space in a photorealistic style.",
+        "dimension": [
+            "overall_consistency",
+            "aesthetic_quality",
+            "imaging_quality"
+        ]
+    },
+    "143_Sewing_machine,_old_sewing_machine_working..mp4": {
+        "prompt_en": "Sewing machine, old sewing machine working.",
+        "dimension": [
+            "overall_consistency",
+            "aesthetic_quality",
+            "imaging_quality"
+        ]
+    },
+    "144_A_boat_sailing_leisurely_along_the_Seine_River_with_the_Eiffel_Tower_in_background_by_Vincent_van_Gogh.mp4": {
+        "prompt_en": "A boat sailing leisurely along the Seine River with the Eiffel Tower in background by Vincent van Gogh",
+        "dimension": [
+            "overall_consistency",
+            "aesthetic_quality",
+            "imaging_quality"
+        ]
+    },
+    "145_A_Mars_rover_moving_on_Mars.mp4": {
+        "prompt_en": "A Mars rover moving on Mars",
+        "dimension": [
+            "overall_consistency",
+            "aesthetic_quality",
+            "imaging_quality"
+        ]
+    },
+    "146_A_super_cool_giant_robot_in_Cyberpunk_Beijing.mp4": {
+        "prompt_en": "A super cool giant robot in Cyberpunk Beijing",
+        "dimension": [
+            "overall_consistency",
+            "aesthetic_quality",
+            "imaging_quality"
+        ]
+    },
+    "147_Iron_Man_flying_in_the_sky.mp4": {
+        "prompt_en": "Iron Man flying in the sky",
+        "dimension": [
+            "overall_consistency",
+            "aesthetic_quality",
+            "imaging_quality"
+        ]
+    },
+    "148_A_beautiful_coastal_beach_in_spring,_waves_lapping_on_sand_by_Hokusai,_in_the_style_of_Ukiyo.mp4": {
+        "prompt_en": "A beautiful coastal beach in spring, waves lapping on sand by Hokusai, in the style of Ukiyo",
+        "dimension": [
+            "overall_consistency",
+            "aesthetic_quality",
+            "imaging_quality"
+        ]
+    },
+    "149_A_cat_eating_food_out_of_a_bowl.mp4": {
+        "prompt_en": "A cat eating food out of a bowl",
+        "dimension": [
+            "overall_consistency",
+            "aesthetic_quality",
+            "imaging_quality"
+        ]
+    },
+    "150_A_cute_fluffy_panda_eating_Chinese_food_in_a_restaurant.mp4": {
+        "prompt_en": "A cute fluffy panda eating Chinese food in a restaurant",
+        "dimension": [
+            "overall_consistency",
+            "aesthetic_quality",
+            "imaging_quality"
+        ]
+    },
+    "151_A_cute_raccoon_playing_guitar_in_a_boat_on_the_ocean.mp4": {
+        "prompt_en": "A cute raccoon playing guitar in a boat on the ocean",
+        "dimension": [
+            "overall_consistency",
+            "aesthetic_quality",
+            "imaging_quality"
+        ]
+    },
+    "152_Clown_fish_swimming_through_the_coral_reef.mp4": {
+        "prompt_en": "Clown fish swimming through the coral reef",
+        "dimension": [
+            "overall_consistency",
+            "aesthetic_quality",
+            "imaging_quality"
+        ]
+    },
+    "153_The_bund_Shanghai,_vibrant_color.mp4": {
+        "prompt_en": "The bund Shanghai, vibrant color",
+        "dimension": [
+            "overall_consistency",
+            "aesthetic_quality",
+            "imaging_quality"
+        ]
+    },
+    "154_alley.mp4": {
+        "prompt_en": "alley",
+        "dimension": [
+            "scene",
+            "background_consistency"
+        ],
+        "auxiliary_info": {
+            "scene": {
+                "scene": {
+                    "scene": "alley"
+                }
+            }
+        }
+    },
+    "155_bridge.mp4": {
+        "prompt_en": "bridge",
+        "dimension": [
+            "scene",
+            "background_consistency"
+        ],
+        "auxiliary_info": {
+            "scene": {
+                "scene": {
+                    "scene": "bridge"
+                }
+            }
+        }
+    },
+    "156_botanical_garden.mp4": {
+        "prompt_en": "botanical garden",
+        "dimension": [
+            "scene",
+            "background_consistency"
+        ],
+        "auxiliary_info": {
+            "scene": {
+                "scene": {
+                    "scene": "botanical garden"
+                }
+            }
+        }
+    },
+    "157_campsite.mp4": {
+        "prompt_en": "campsite",
+        "dimension": [
+            "scene",
+            "background_consistency"
+        ],
+        "auxiliary_info": {
+            "scene": {
+                "scene": {
+                    "scene": "campsite"
+                }
+            }
+        }
+    },
+    "158_castle.mp4": {
+        "prompt_en": "castle",
+        "dimension": [
+            "scene",
+            "background_consistency"
+        ],
+        "auxiliary_info": {
+            "scene": {
+                "scene": {
+                    "scene": "castle"
+                }
+            }
+        }
+    },
+    "159_construction_site.mp4": {
+        "prompt_en": "construction site",
+        "dimension": [
+            "scene",
+            "background_consistency"
+        ],
+        "auxiliary_info": {
+            "scene": {
+                "scene": {
+                    "scene": "construction site"
+                }
+            }
+        }
+    },
+    "160_food_court.mp4": {
+        "prompt_en": "food court",
+        "dimension": [
+            "scene",
+            "background_consistency"
+        ],
+        "auxiliary_info": {
+            "scene": {
+                "scene": {
+                    "scene": "food court"
+                }
+            }
+        }
+    },
+    "161_hospital.mp4": {
+        "prompt_en": "hospital",
+        "dimension": [
+            "scene",
+            "background_consistency"
+        ],
+        "auxiliary_info": {
+            "scene": {
+                "scene": {
+                    "scene": "hospital"
+                }
+            }
+        }
+    },
+    "162_industrial_area.mp4": {
+        "prompt_en": "industrial area",
+        "dimension": [
+            "scene",
+            "background_consistency"
+        ],
+        "auxiliary_info": {
+            "scene": {
+                "scene": {
+                    "scene": "industrial area"
+                }
+            }
+        }
+    },
+    "163_junkyard.mp4": {
+        "prompt_en": "junkyard",
+        "dimension": [
+            "scene",
+            "background_consistency"
+        ],
+        "auxiliary_info": {
+            "scene": {
+                "scene": {
+                    "scene": "junkyard"
+                }
+            }
+        }
+    },
+    "164_lighthouse.mp4": {
+        "prompt_en": "lighthouse",
+        "dimension": [
+            "scene",
+            "background_consistency"
+        ],
+        "auxiliary_info": {
+            "scene": {
+                "scene": {
+                    "scene": "lighthouse"
+                }
+            }
+        }
+    },
+    "165_indoor_movie_theater.mp4": {
+        "prompt_en": "indoor movie theater",
+        "dimension": [
+            "scene",
+            "background_consistency"
+        ],
+        "auxiliary_info": {
+            "scene": {
+                "scene": {
+                    "scene": "indoor movie theater"
+                }
+            }
+        }
+    },
+    "166_nursery.mp4": {
+        "prompt_en": "nursery",
+        "dimension": [
+            "scene",
+            "background_consistency"
+        ],
+        "auxiliary_info": {
+            "scene": {
+                "scene": {
+                    "scene": "nursery"
+                }
+            }
+        }
+    },
+    "167_ocean.mp4": {
+        "prompt_en": "ocean",
+        "dimension": [
+            "scene",
+            "background_consistency"
+        ],
+        "auxiliary_info": {
+            "scene": {
+                "scene": {
+                    "scene": "ocean"
+                }
+            }
+        }
+    },
+    "168_office.mp4": {
+        "prompt_en": "office",
+        "dimension": [
+            "scene",
+            "background_consistency"
+        ],
+        "auxiliary_info": {
+            "scene": {
+                "scene": {
+                    "scene": "office"
+                }
+            }
+        }
+    },
+    "169_river.mp4": {
+        "prompt_en": "river",
+        "dimension": [
+            "scene",
+            "background_consistency"
+        ],
+        "auxiliary_info": {
+            "scene": {
+                "scene": {
+                    "scene": "river"
+                }
+            }
+        }
+    },
+    "170_shower.mp4": {
+        "prompt_en": "shower",
+        "dimension": [
+            "scene",
+            "background_consistency"
+        ],
+        "auxiliary_info": {
+            "scene": {
+                "scene": {
+                    "scene": "shower"
+                }
+            }
+        }
+    },
+    "171_supermarket.mp4": {
+        "prompt_en": "supermarket",
+        "dimension": [
+            "scene",
+            "background_consistency"
+        ],
+        "auxiliary_info": {
+            "scene": {
+                "scene": {
+                    "scene": "supermarket"
+                }
+            }
+        }
+    },
+    "172_tower.mp4": {
+        "prompt_en": "tower",
+        "dimension": [
+            "scene",
+            "background_consistency"
+        ],
+        "auxiliary_info": {
+            "scene": {
+                "scene": {
+                    "scene": "tower"
+                }
+            }
+        }
+    },
+    "173_bakery_shop.mp4": {
+        "prompt_en": "bakery shop",
+        "dimension": [
+            "scene",
+            "background_consistency"
+        ],
+        "auxiliary_info": {
+            "scene": {
+                "scene": {
+                    "scene": "bakery shop"
+                }
+            }
+        }
+    },
+    "174_ballroom.mp4": {
+        "prompt_en": "ballroom",
+        "dimension": [
+            "scene",
+            "background_consistency"
+        ],
+        "auxiliary_info": {
+            "scene": {
+                "scene": {
+                    "scene": "ballroom"
+                }
+            }
+        }
+    },
+    "175_botanical_garden.mp4": {
+        "prompt_en": "botanical garden",
+        "dimension": [
+            "scene",
+            "background_consistency"
+        ],
+        "auxiliary_info": {
+            "scene": {
+                "scene": {
+                    "scene": "botanical garden"
+                }
+            }
+        }
+    },
+    "176_cafeteria.mp4": {
+        "prompt_en": "cafeteria",
+        "dimension": [
+            "scene",
+            "background_consistency"
+        ],
+        "auxiliary_info": {
+            "scene": {
+                "scene": {
+                    "scene": "cafeteria"
+                }
+            }
+        }
+    },
+    "177_crosswalk.mp4": {
+        "prompt_en": "crosswalk",
+        "dimension": [
+            "scene",
+            "background_consistency"
+        ],
+        "auxiliary_info": {
+            "scene": {
+                "scene": {
+                    "scene": "crosswalk"
+                }
+            }
+        }
+    },
+    "178_construction_site.mp4": {
+        "prompt_en": "construction site",
+        "dimension": [
+            "scene",
+            "background_consistency"
+        ],
+        "auxiliary_info": {
+            "scene": {
+                "scene": {
+                    "scene": "construction site"
+                }
+            }
+        }
+    },
+    "179_courtyard.mp4": {
+        "prompt_en": "courtyard",
+        "dimension": [
+            "scene",
+            "background_consistency"
+        ],
+        "auxiliary_info": {
+            "scene": {
+                "scene": {
+                    "scene": "courtyard"
+                }
+            }
+        }
+    },
+    "180_food_court.mp4": {
+        "prompt_en": "food court",
+        "dimension": [
+            "scene",
+            "background_consistency"
+        ],
+        "auxiliary_info": {
+            "scene": {
+                "scene": {
+                    "scene": "food court"
+                }
+            }
+        }
+    },
+    "181_indoor_gymnasium.mp4": {
+        "prompt_en": "indoor gymnasium",
+        "dimension": [
+            "scene",
+            "background_consistency"
+        ],
+        "auxiliary_info": {
+            "scene": {
+                "scene": {
+                    "scene": "indoor gymnasium"
+                }
+            }
+        }
+    },
+    "182_indoor_library.mp4": {
+        "prompt_en": "indoor library",
+        "dimension": [
+            "scene",
+            "background_consistency"
+        ],
+        "auxiliary_info": {
+            "scene": {
+                "scene": {
+                    "scene": "indoor library"
+                }
+            }
+        }
+    },
+    "183_marsh.mp4": {
+        "prompt_en": "marsh",
+        "dimension": [
+            "scene",
+            "background_consistency"
+        ],
+        "auxiliary_info": {
+            "scene": {
+                "scene": {
+                    "scene": "marsh"
+                }
+            }
+        }
+    },
+    "184_mountain.mp4": {
+        "prompt_en": "mountain",
+        "dimension": [
+            "scene",
+            "background_consistency"
+        ],
+        "auxiliary_info": {
+            "scene": {
+                "scene": {
+                    "scene": "mountain"
+                }
+            }
+        }
+    },
+    "185_science_museum.mp4": {
+        "prompt_en": "science museum",
+        "dimension": [
+            "scene",
+            "background_consistency"
+        ],
+        "auxiliary_info": {
+            "scene": {
+                "scene": {
+                    "scene": "science museum"
+                }
+            }
+        }
+    },
+    "186_a_parking_meter_on_the_right_of_a_bench,_front_view.mp4": {
+        "prompt_en": "a parking meter on the right of a bench, front view",
+        "dimension": [
+            "spatial_relationship"
+        ],
+        "auxiliary_info": {
+            "spatial_relationship": {
+                "spatial_relationship": {
+                    "object_a": "parking meter",
+                    "object_b": "bench",
+                    "relationship": "on the right of"
+                }
+            }
+        }
+    },
+    "187_a_cow_on_the_right_of_an_elephant,_front_view.mp4": {
+        "prompt_en": "a cow on the right of an elephant, front view",
+        "dimension": [
+            "spatial_relationship"
+        ],
+        "auxiliary_info": {
+            "spatial_relationship": {
+                "spatial_relationship": {
+                    "object_a": "cow",
+                    "object_b": "elephant",
+                    "relationship": "on the right of"
+                }
+            }
+        }
+    },
+    "188_a_sports_ball_on_the_right_of_a_baseball_bat,_front_view.mp4": {
+        "prompt_en": "a sports ball on the right of a baseball bat, front view",
+        "dimension": [
+            "spatial_relationship"
+        ],
+        "auxiliary_info": {
+            "spatial_relationship": {
+                "spatial_relationship": {
+                    "object_a": "sports ball",
+                    "object_b": "baseball bat",
+                    "relationship": "on the right of"
+                }
+            }
+        }
+    },
+    "189_a_baseball_bat_on_the_left_of_a_baseball_glove,_front_view.mp4": {
+        "prompt_en": "a baseball bat on the left of a baseball glove, front view",
+        "dimension": [
+            "spatial_relationship"
+        ],
+        "auxiliary_info": {
+            "spatial_relationship": {
+                "spatial_relationship": {
+                    "object_a": "baseball bat",
+                    "object_b": "baseball glove",
+                    "relationship": "on the left of"
+                }
+            }
+        }
+    },
+    "190_an_oven_on_the_bottom_of_a_toaster,_front_view.mp4": {
+        "prompt_en": "an oven on the bottom of a toaster, front view",
+        "dimension": [
+            "spatial_relationship"
+        ],
+        "auxiliary_info": {
+            "spatial_relationship": {
+                "spatial_relationship": {
+                    "object_a": "oven",
+                    "object_b": "toaster",
+                    "relationship": "on the bottom of"
+                }
+            }
+        }
+    },
+    "191_a_hot_dog_on_the_top_of_a_pizza,_front_view.mp4": {
+        "prompt_en": "a hot dog on the top of a pizza, front view",
+        "dimension": [
+            "spatial_relationship"
+        ],
+        "auxiliary_info": {
+            "spatial_relationship": {
+                "spatial_relationship": {
+                    "object_a": "hot dog",
+                    "object_b": "pizza",
+                    "relationship": "on the top of"
+                }
+            }
+        }
+    },
+    "192_a_donut_on_the_top_of_broccoli,_front_view.mp4": {
+        "prompt_en": "a donut on the top of broccoli, front view",
+        "dimension": [
+            "spatial_relationship"
+        ],
+        "auxiliary_info": {
+            "spatial_relationship": {
+                "spatial_relationship": {
+                    "object_a": "donut",
+                    "object_b": "broccoli",
+                    "relationship": "on the top of"
+                }
+            }
+        }
+    },
+    "193_a_donut_on_the_bottom_of_broccoli,_front_view.mp4": {
+        "prompt_en": "a donut on the bottom of broccoli, front view",
+        "dimension": [
+            "spatial_relationship"
+        ],
+        "auxiliary_info": {
+            "spatial_relationship": {
+                "spatial_relationship": {
+                    "object_a": "donut",
+                    "object_b": "broccoli",
+                    "relationship": "on the bottom of"
+                }
+            }
+        }
+    },
+    "194_broccoli_on_the_bottom_of_a_banana,_front_view.mp4": {
+        "prompt_en": "broccoli on the bottom of a banana, front view",
+        "dimension": [
+            "spatial_relationship"
+        ],
+        "auxiliary_info": {
+            "spatial_relationship": {
+                "spatial_relationship": {
+                    "object_a": "broccoli",
+                    "object_b": "banana",
+                    "relationship": "on the bottom of"
+                }
+            }
+        }
+    },
+    "195_skis_on_the_top_of_a_snowboard,_front_view.mp4": {
+        "prompt_en": "skis on the top of a snowboard, front view",
+        "dimension": [
+            "spatial_relationship"
+        ],
+        "auxiliary_info": {
+            "spatial_relationship": {
+                "spatial_relationship": {
+                    "object_a": "skis",
+                    "object_b": "snowboard",
+                    "relationship": "on the top of"
+                }
+            }
+        }
+    },
+    "196_a_snowboard_on_the_bottom_of_a_kite,_front_view.mp4": {
+        "prompt_en": "a snowboard on the bottom of a kite, front view",
+        "dimension": [
+            "spatial_relationship"
+        ],
+        "auxiliary_info": {
+            "spatial_relationship": {
+                "spatial_relationship": {
+                    "object_a": "snowboard",
+                    "object_b": "kite",
+                    "relationship": "on the bottom of"
+                }
+            }
+        }
+    },
+    "197_a_kite_on_the_bottom_of_a_skateboard,_front_view.mp4": {
+        "prompt_en": "a kite on the bottom of a skateboard, front view",
+        "dimension": [
+            "spatial_relationship"
+        ],
+        "auxiliary_info": {
+            "spatial_relationship": {
+                "spatial_relationship": {
+                    "object_a": "kite",
+                    "object_b": "skateboard",
+                    "relationship": "on the bottom of"
+                }
+            }
+        }
+    },
+    "198_a_skateboard_on_the_top_of_a_surfboard,_front_view.mp4": {
+        "prompt_en": "a skateboard on the top of a surfboard, front view",
+        "dimension": [
+            "spatial_relationship"
+        ],
+        "auxiliary_info": {
+            "spatial_relationship": {
+                "spatial_relationship": {
+                    "object_a": "skateboard",
+                    "object_b": "surfboard",
+                    "relationship": "on the top of"
+                }
+            }
+        }
+    },
+    "199_a_surfboard_on_the_top_of_skis,_front_view.mp4": {
+        "prompt_en": "a surfboard on the top of skis, front view",
+        "dimension": [
+            "spatial_relationship"
+        ],
+        "auxiliary_info": {
+            "spatial_relationship": {
+                "spatial_relationship": {
+                    "object_a": "surfboard",
+                    "object_b": "skis",
+                    "relationship": "on the top of"
+                }
+            }
+        }
+    }
+}

src/videogen_hub/benchmark/text_guided_t2v.py

@@ -0,0 +1,137 @@
+from typing import Optional
+import os
+from tqdm import tqdm
+from videogen_hub.infermodels import load_model
+import cv2, json 
+import numpy as np
+import argparse
+from videogen_hub.utils.file_helper import get_file_path
+from moviepy.editor import ImageSequenceClip
+
+
+def infer_text_guided_vg_bench(
+    model,
+    result_folder: str = "results",
+    experiment_name: str = "Exp_Text-Guided_VG",
+    overwrite_model_outputs: bool = False,
+    overwrite_inputs: bool = False,
+    limit_videos_amount: Optional[int] = None,
+):
+    """
+    Performs inference on the VideogenHub dataset using the provided text-guided video generation model.
+
+    Args:
+        model: Instance of a model that supports text-guided video generation. Expected to have
+               a method 'infer_one_video' for inferencing.
+        result_folder (str, optional): Path to the root directory where the results should be saved.
+               Defaults to 'results'.
+        experiment_name (str, optional): Name of the folder inside 'result_folder' where results
+               for this particular experiment will be stored. Defaults to "Exp_Text-Guided_IG".
+        overwrite_model_outputs (bool, optional): If set to True, will overwrite any pre-existing
+               model outputs. Useful for resuming runs. Defaults to False.
+        overwrite_inputs (bool, optional): If set to True, will overwrite any pre-existing input
+               samples. Typically, should be set to False unless there's a need to update the inputs.
+               Defaults to False.
+        limit_videos_amount (int, optional): Limits the number of videos to be processed. If set to
+               None, all videos in the dataset will be processed.
+
+    Returns:
+        None. Results are saved in the specified directory.
+
+    Notes:
+        The function processes each sample from the dataset, uses the model to infer an video based
+        on text prompts, and then saves the resulting videos in the specified directories.
+    """
+    benchmark_prompt_path = "t2v_vbench_1000.json"
+    prompts = json.load(open(get_file_path(benchmark_prompt_path), "r"))
+    save_path = os.path.join(result_folder, experiment_name, "dataset_lookup.json")
+    if overwrite_inputs or not os.path.exists(save_path):
+        if not os.path.exists(os.path.join(result_folder, experiment_name)):
+            os.makedirs(os.path.join(result_folder, experiment_name))
+        with open(save_path, "w") as f:
+            json.dump(prompts, f, indent=4)
+
+    print(
+        "========> Running Benchmark Dataset:",
+        experiment_name,
+        "| Model:",
+        model.__class__.__name__,
+    )
+
+    for file_basename, prompt in tqdm(prompts.items()):
+        idx = int(file_basename.split("_")[0])
+        dest_folder = os.path.join(
+            result_folder, experiment_name, model.__class__.__name__
+        )
+        # file_basename = f"{idx}_{prompt['prompt_en'].replace(' ', '_')}.mp4"
+        if not os.path.exists(dest_folder):
+            os.mkdir(dest_folder)
+        dest_file = os.path.join(dest_folder, file_basename)
+        if overwrite_model_outputs or not os.path.exists(dest_file):
+            print("========> Inferencing", dest_file)
+            frames = model.infer_one_video(prompt=prompt["prompt_en"])
+
+            #special_treated_list = ["LaVie", "ModelScope", "T2VTurbo"]
+            special_treated_list = []
+            if model.__class__.__name__ in special_treated_list:
+                print("======> Saved through cv2.VideoWriter_fourcc")
+                # save the video
+                fps = 8
+                fourcc = cv2.VideoWriter_fourcc(*"mp4v")  # Codec
+                out = cv2.VideoWriter(
+                    dest_file, fourcc, fps, (frames.shape[2], frames.shape[1])
+                )
+
+                # Convert each tensor frame to numpy and write it to the video
+                for i in range(frames.shape[0]):
+                    frame = frames[i].numpy().astype(np.uint8)
+                    out.write(frame)
+
+                out.release()
+            else:
+                def tensor_to_video(tensor, output_path, fps=8):
+                    """
+                    Converts a PyTorch tensor to a video file.
+                    
+                    Args:
+                        tensor (torch.Tensor): The input tensor of shape (T, C, H, W).
+                        output_path (str): The path to save the output video.
+                        fps (int): Frames per second for the output video.
+                    """
+                    # Ensure the tensor is on the CPU and convert to NumPy array
+                    tensor = tensor.cpu().numpy()
+                    
+                    # Normalize the tensor values to [0, 1]
+                    tensor_min = tensor.min()
+                    tensor_max = tensor.max()
+                    tensor = (tensor - tensor_min) / (tensor_max - tensor_min)
+                    
+                    # Permute dimensions from (T, C, H, W) to (T, H, W, C) and scale to [0, 255]
+                    video_frames = (tensor.transpose(0, 2, 3, 1) * 255).astype(np.uint8)
+                    
+                    # Create a video clip from the frames
+                    clip = ImageSequenceClip(list(video_frames), fps=fps)
+                    
+                    # Write the video file
+                    clip.write_videofile(output_path, codec='libx264')
+
+                if frames.shape[-1] == 3:
+                    frames = frames.permute(0, 3, 1, 2)
+                    print("======> corrected frames.shape", frames.shape)
+
+                tensor_to_video(frames, dest_file)
+        else:
+            print("========> Skipping", dest_file, ", it already exists")
+
+        if limit_videos_amount is not None and (idx >= limit_videos_amount):
+            break
+
+
+# for testing
+if __name__ == "__main__":
+    parser = argparse.ArgumentParser(description="Load a model by name")
+    parser.add_argument("--model_name", type=str, required=True, help="Name of the model to load")
+    args = parser.parse_args()
+    
+    model = load_model(args.model_name)
+    infer_text_guided_vg_bench(model)

src/videogen_hub/benchmark/transform.py

@@ -0,0 +1,16 @@
+import json
+
+def main(prompt_path):
+    new_prompt = {}
+    prompts = json.load(open(prompt_path, "r"))
+    
+    for idx, prompt in enumerate(prompts):
+        new_prompt[f"{idx}_{prompt['prompt_en'].replace(' ', '_')}.mp4"] = prompt
+    
+    with open(f"new_{prompt_path}", "w") as f:
+        json.dump(new_prompt, f, indent=4)
+    
+    
+if __name__ == "__main__":
+    # main("t2v_vbench_200.json")
+    main("t2v_vbench_remain.json")