text
stringclasses
10 values
videos
sequencelengths
1
1
__dj__stats__
dict
<__dj__video> A group of children dressed in colorful costumes are participating in an outdoor event where they are decorating pumpkins with bubble wands.
[ "./videos/panda/1AiNG23hyUo_7.mp4" ]
{ "alnum_ratio": 0.8116883117, "char_rep_ratio": 0, "flagged_words_ratio": 0, "lang": "en", "lang_score": 0.9006049037, "num_action": 3, "num_dependency_edges": [ 4, 1, 2, 4, 1, 2, 1, 2 ], "num_token": 32, "num_words": 23, "perplexity": 493.4, "special_char_ratio": 0.1883116883, "stopwords_ratio": 0.4347826087, "text_len": 154, "word_rep_ratio": 0, "video_frames_aesthetics_score": [ 0.4555848837 ], "video_aspect_ratios": [ 1.7777777778 ], "video_duration": [ 2 ], "video_height": [ 720 ], "video_width": [ 1280 ], "video_frames_text_similarity": [ 0.3540362418 ], "video_motion_score": [ 5.4380040169 ], "video_nsfw_score": [ 0.0001698484 ], "video_ocr_area_ratio": [ 0.0103081597 ], "video_watermark_prob": [ 0.5288844109 ] }
<__dj__video> a screenshot of an asian screenshot of a game
[ "./videos/internvid/RVZUInNpCFU_9_1.mp4" ]
{ "alnum_ratio": 0.7457627119, "char_rep_ratio": 0.24, "flagged_words_ratio": 0, "lang": "en", "lang_score": 0.5384221077, "num_action": 0, "num_dependency_edges": [ 2, 4, 2 ], "num_token": 16, "num_words": 10, "perplexity": 1951.5, "special_char_ratio": 0.2542372881, "stopwords_ratio": 0.5, "text_len": 59, "word_rep_ratio": 0, "video_frames_aesthetics_score": [ 0.4426309466 ], "video_aspect_ratios": [ 1.7777777778 ], "video_duration": [ 2 ], "video_height": [ 720 ], "video_width": [ 1280 ], "video_frames_text_similarity": [ 0.3099360466 ], "video_motion_score": [ 3.9707188606 ], "video_nsfw_score": [ 0.0001951887 ], "video_ocr_area_ratio": [ 0.0479736328 ], "video_watermark_prob": [ 0.9463900328 ] }
<__dj__video> the word 'gynynyt' is in the middle of many different fruits and vegetables
[ "./videos/internvid/Ip9TjM5PY9I_1_1.mp4" ]
{ "alnum_ratio": 0.7640449438, "char_rep_ratio": 0, "flagged_words_ratio": 0, "lang": "en", "lang_score": 0.7287845016, "num_action": 0, "num_dependency_edges": [ 2, 1, 3, 5, 1 ], "num_token": 24, "num_words": 14, "perplexity": 813.4, "special_char_ratio": 0.2359550562, "stopwords_ratio": 0.5, "text_len": 89, "word_rep_ratio": 0, "video_frames_aesthetics_score": [ 0.3954536319 ], "video_aspect_ratios": [ 1.7777777778 ], "video_duration": [ 2 ], "video_height": [ 720 ], "video_width": [ 1280 ], "video_frames_text_similarity": [ 0.3075161576 ], "video_motion_score": [ 2.2344789505 ], "video_nsfw_score": [ 0.0001862614 ], "video_ocr_area_ratio": [ 0.045139974 ], "video_watermark_prob": [ 0.5961741209 ] }
<__dj__video> a woman in a yellow shirt is on a tennis court
[ "./videos/internvid/Per2JwtVUwM_3_1.mp4" ]
{ "alnum_ratio": 0.7166666667, "char_rep_ratio": 0, "flagged_words_ratio": 0, "lang": "en", "lang_score": 0.5908017755, "num_action": 0, "num_dependency_edges": [ 3, 3, 1, 3 ], "num_token": 17, "num_words": 12, "perplexity": 1720.9, "special_char_ratio": 0.2833333333, "stopwords_ratio": 0.5, "text_len": 60, "word_rep_ratio": 0, "video_frames_aesthetics_score": [ 0.4742202759 ], "video_aspect_ratios": [ 1.7777777778 ], "video_duration": [ 4 ], "video_height": [ 720 ], "video_width": [ 1280 ], "video_frames_text_similarity": [ 0.319634378 ], "video_motion_score": [ 10.1790513992 ], "video_nsfw_score": [ 0.0001222684 ], "video_ocr_area_ratio": [ 0.0214476612 ], "video_watermark_prob": [ 0.8281876445 ] }
<__dj__video> A man with glasses is sitting in a car and eating food.
[ "./videos/panda/0p8CO6j9U-M_3.mp4" ]
{ "alnum_ratio": 0.7246376812, "char_rep_ratio": 0, "flagged_words_ratio": 0, "lang": "en", "lang_score": 0.923728466, "num_action": 2, "num_dependency_edges": [ 3, 1, 2, 1 ], "num_token": 19, "num_words": 13, "perplexity": 863.1, "special_char_ratio": 0.2753623188, "stopwords_ratio": 0.4615384615, "text_len": 69, "word_rep_ratio": 0, "video_frames_aesthetics_score": [ 0.4376985431 ], "video_aspect_ratios": [ 1.7777777778 ], "video_duration": [ 17 ], "video_height": [ 720 ], "video_width": [ 1280 ], "video_frames_text_similarity": [ 0.3342003822 ], "video_motion_score": [ 3.6470098495 ], "video_nsfw_score": [ 0.000185161 ], "video_ocr_area_ratio": [ 0 ], "video_watermark_prob": [ 0.5817964077 ] }
<__dj__video> an image of someone pouring soil into a wheelbarrow
[ "./videos/internvid/t692w4byVrw_7_2.mp4" ]
{ "alnum_ratio": 0.7692307692, "char_rep_ratio": 0, "flagged_words_ratio": 0, "lang": "en", "lang_score": 0.6552112699, "num_action": 1, "num_dependency_edges": [ 2, 2, 1, 2 ], "num_token": 17, "num_words": 10, "perplexity": 1950, "special_char_ratio": 0.2307692308, "stopwords_ratio": 0.5, "text_len": 65, "word_rep_ratio": 0, "video_frames_aesthetics_score": [ 0.4468871057 ], "video_aspect_ratios": [ 1.7777777778 ], "video_duration": [ 11 ], "video_height": [ 720 ], "video_width": [ 1280 ], "video_frames_text_similarity": [ 0.3253023028 ], "video_motion_score": [ 3.5699613094 ], "video_nsfw_score": [ 0.0001119684 ], "video_ocr_area_ratio": [ 0 ], "video_watermark_prob": [ 0.5961615443 ] }
<__dj__video> a wedding car with a red ribbon and flowers
[ "./videos/internvid/froKkfIPaHA_1_1.mp4" ]
{ "alnum_ratio": 0.7368421053, "char_rep_ratio": 0, "flagged_words_ratio": 0, "lang": "en", "lang_score": 0.6550011635, "num_action": 0, "num_dependency_edges": [ 1, 3, 5, 1 ], "num_token": 15, "num_words": 10, "perplexity": 2739.7, "special_char_ratio": 0.2631578947, "stopwords_ratio": 0.4, "text_len": 57, "word_rep_ratio": 0, "video_frames_aesthetics_score": [ 0.4366855621 ], "video_aspect_ratios": [ 1.7777777778 ], "video_duration": [ 1 ], "video_height": [ 720 ], "video_width": [ 1280 ], "video_frames_text_similarity": [ 0.3092386127 ], "video_motion_score": [ 11.0694389343 ], "video_nsfw_score": [ 0.0001391819 ], "video_ocr_area_ratio": [ 0.0036116536 ], "video_watermark_prob": [ 0.3725105524 ] }
<__dj__video> A man dancing with two young girls in a dance studio.
[ "./videos/panda/-vPbw02IbRc_3.mp4" ]
{ "alnum_ratio": 0.7313432836, "char_rep_ratio": 0, "flagged_words_ratio": 0, "lang": "en", "lang_score": 0.4970883131, "num_action": 1, "num_dependency_edges": [ 2, 3, 1, 3 ], "num_token": 18, "num_words": 12, "perplexity": 967.6, "special_char_ratio": 0.2686567164, "stopwords_ratio": 0.3333333333, "text_len": 67, "word_rep_ratio": 0, "video_frames_aesthetics_score": [ 0.4885383844 ], "video_aspect_ratios": [ 1.7777777778 ], "video_duration": [ 4 ], "video_height": [ 720 ], "video_width": [ 1280 ], "video_frames_text_similarity": [ 0.3328181803 ], "video_motion_score": [ 6.7048120499 ], "video_nsfw_score": [ 0.0001282161 ], "video_ocr_area_ratio": [ 0.0047927517 ], "video_watermark_prob": [ 0.5738118887 ] }
<__dj__video> young people wearing face masks sit in the seats
[ "./videos/internvid/ivU8GO4Zyo0_7_1.mp4" ]
{ "alnum_ratio": 0.7580645161, "char_rep_ratio": 0, "flagged_words_ratio": 0, "lang": "en", "lang_score": 0.7919024229, "num_action": 2, "num_dependency_edges": [ 3, 1, 2, 2 ], "num_token": 15, "num_words": 10, "perplexity": 2879.6, "special_char_ratio": 0.2419354839, "stopwords_ratio": 0.2, "text_len": 62, "word_rep_ratio": 0, "video_frames_aesthetics_score": [ 0.4656786919 ], "video_aspect_ratios": [ 1.7777777778 ], "video_duration": [ 3 ], "video_height": [ 720 ], "video_width": [ 1280 ], "video_frames_text_similarity": [ 0.3210753798 ], "video_motion_score": [ 0.7968443036 ], "video_nsfw_score": [ 0.0001760678 ], "video_ocr_area_ratio": [ 0.0179399957 ], "video_watermark_prob": [ 0.9622330666 ] }
<__dj__video> A man in a blue shirt smiles at the camera while people stand in the background.
[ "./videos/panda/-mdIuelE99E_17.mp4" ]
{ "alnum_ratio": 0.7553191489, "char_rep_ratio": 0, "flagged_words_ratio": 0, "lang": "en", "lang_score": 0.6364951134, "num_action": 2, "num_dependency_edges": [ 3, 3, 2, 1, 2 ], "num_token": 23, "num_words": 17, "perplexity": 622.2, "special_char_ratio": 0.2446808511, "stopwords_ratio": 0.4705882353, "text_len": 94, "word_rep_ratio": 0, "video_frames_aesthetics_score": [ 0.4479264021 ], "video_aspect_ratios": [ 1.7777777778 ], "video_duration": [ 13 ], "video_height": [ 720 ], "video_width": [ 1280 ], "video_frames_text_similarity": [ 0.3299186826 ], "video_motion_score": [ 10.0738592148 ], "video_nsfw_score": [ 0.0001195985 ], "video_ocr_area_ratio": [ 0.0164872685 ], "video_watermark_prob": [ 0.9694447517 ] }

Data-Juicer Sandbox: A Comprehensive Suite for Multimodal Data-Model Co-development

Project description

The emergence of large-scale multi-modal generative models has drastically advanced artificial intelligence, introducing unprecedented levels of performance and functionality. However, optimizing these models remains challenging due to historically isolated paths of model-centric and data-centric developments, leading to suboptimal outcomes and inefficient resource utilization. In response, we present a novel sandbox suite tailored for integrated data-model co-development. This sandbox provides a comprehensive experimental platform, enabling rapid iteration and insight-driven refinement of both data and models. Our proposed "Probe-Analyze-Refine" workflow, validated through applications on T2V-Turbo and achieve a new state-of-the-art on VBench leaderboard with 1.09% improvement from T2V-Turbo. Our experiment code and model are released at Data-Juicer Sandbox.

Dataset Information

  • The whole dataset is available here (About 227.5GB).
  • Number of samples: 147,176 (Include videos and keep ~12.09% from the original dataset)
  • The original dataset totals 1,217k instances from InternVid (606k), Panda-70M (605k), and MSR-VTT (6k).

Refining Recipe

# global parameters
# global parameters
project_name: 'Data-Juicer-recipes-T2V-optimal'
dataset_path: '/path/to/your/dataset'  # path to your dataset directory or file
export_path: '/path/to/your/dataset.jsonl'

np: 4  # number of subprocess to process your dataset

# process schedule
# a list of several process operators with their arguments
process:
  - video_nsfw_filter:
      hf_nsfw_model: Falconsai/nsfw_image_detection
      score_threshold: 0.000195383
      frame_sampling_method: uniform
      frame_num: 3
      reduce_mode: avg
      any_or_all: any
      mem_required: '1GB'
  - video_frames_text_similarity_filter:
      hf_clip: openai/clip-vit-base-patch32
      min_score: 0.306337
      max_score: 1.0
      frame_sampling_method: uniform
      frame_num: 3
      horizontal_flip: false
      vertical_flip: false
      reduce_mode: avg
      any_or_all: any
      mem_required: '10GB'
Downloads last month
2
Edit dataset card