simplecloud/VidChain-exercise
✏️ Data for VidChain Excercise VidChain: Chain-of-Tasks with Metric-based Direct Preference Optimization for Dense Video Captioning Ji Soo Lee*, Jongha Kim*, Jeehye Na, Jinyoung Park, Hyunwoo J. Kim†. AAAI 2025 🎯 Learning Objectives By working through this exercise, you will: Reproduce baseline behavior of a video-language model (VTimeLLM, CVPR 2024 Highlight). Observe the limitations of existing approaches in temporal… See the full description on the dataset page: https://huggingface.co/datasets/simplecloud/VidChain-exercise.
0150
1{2 "python.defaultInterpreterPath": "/home/jisoo/.conda/envs/vtimellm/bin/python",3 "python.terminal.activateEnvironment": true,4 "python.condaPath": "/home/jisoo/.conda/bin/conda",5 "python.envFile": "${workspaceFolder}/.env",6 "python.analysis.extraPaths": [7 "${workspaceFolder}",8 "${workspaceFolder}/vtimellm",9 "${workspaceFolder}/vtimellm/train",10 "${workspaceFolder}/vtimellm/model"11 ],12 "python.analysis.autoImportCompletions": true,13 "python.analysis.typeCheckingMode": "basic",14 "python.analysis.autoSearchPaths": true,15 "python.analysis.diagnosticMode": "workspace",16 "python.analysis.stubPath": "${workspaceFolder}/typings",17 "python.linting.enabled": true,18 "python.linting.pylintEnabled": false,19 "python.linting.flake8Enabled": true,20 "python.linting.mypyEnabled": false,21 "python.formatting.provider": "black",22 "python.sortImports.args": ["--profile", "black"],23 "files.associations": {24 "*.py": "python"25 },26 "terminal.integrated.env.linux": {27 "PYTHONPATH": "${workspaceFolder}:${env:PYTHONPATH}",28 "CUDA_VISIBLE_DEVICES": "1",29 "TRANSFORMERS_OFFLINE": "1",30 "WANDB_PROJECT": "vtimellm"31 },32 "terminal.integrated.defaultProfile.linux": "bash",33 "terminal.integrated.profiles.linux": {34 "bash": {35 "path": "/bin/bash",36 "args": ["-l"]37 }38 }39}40