Add files using upload-large-folder tool

.gitattributes

@@ -1,6 +1,5 @@
 *.7z filter=lfs diff=lfs merge=lfs -text
 *.arrow filter=lfs diff=lfs merge=lfs -text
-*.avro filter=lfs diff=lfs merge=lfs -text
 *.bin filter=lfs diff=lfs merge=lfs -text
 *.bz2 filter=lfs diff=lfs merge=lfs -text
 *.ckpt filter=lfs diff=lfs merge=lfs -text

README.md

@@ -1,3 +1,202 @@
----
-license: apache-2.0
----
+---
+license: mit
+task_categories:
+- robotics
+- reinforcement-learning
+tags:
+- metaworld
+- robotics
+- manipulation
+- multi-task
+- r3m
+- vision-language
+- imitation
+size_categories:
+- 1K<n<10K
+language:
+- en
+pretty_name: Short-MetaWorld Dataset
+dataset_info:
+  features:
+  - name: image
+    dtype: image
+  - name: state
+    dtype: 
+      sequence: float32
+  - name: action
+    dtype:
+      sequence: float32
+  - name: prompt
+    dtype: string
+  - name: task_name
+    dtype: string
+  splits:
+  - name: train
+    num_bytes: 1900000000
+    num_examples: 40000
+  download_size: 1900000000
+  dataset_size: 1900000000
+---
+
+# Short-MetaWorld Dataset
+
+## Overview
+
+Short-MetaWorld is a curated dataset from Meta-World containing **Multi-Task 10 (MT10)** and **Meta-Learning 10 (ML10)** tasks with **100 successful trajectories per task** and **20 steps per trajectory**. This dataset is specifically designed for multi-task robot learning, imitation learning, and vision-language robotics research.
+
+## 🚀 Quick Start
+
+```python
+from short_metaworld_loader import load_short_metaworld
+from torch.utils.data import DataLoader
+
+# Load the dataset
+dataset = load_short_metaworld("./", image_size=224)
+
+# Create a DataLoader
+dataloader = DataLoader(dataset, batch_size=32, shuffle=True)
+
+# Get a sample
+sample = dataset[0]
+print(f"Image shape: {sample['image'].shape}")
+print(f"State: {sample['state']}")
+print(f"Action: {sample['action']}")
+print(f"Task: {sample['task_name']}")
+print(f"Prompt: {sample['prompt']}")
+```
+
+## 📁 Dataset Structure
+
+```
+short-MetaWorld/
+├── README.txt                     # Original dataset documentation
+├── short-MetaWorld/
+│   ├── img_only/                    # 224x224 RGB images
+│   │   ├── button-press-topdown-v2/
+│   │   │   ├── 0/                   # Trajectory 0
+│   │   │   │   ├── 0.jpg           # Step 0
+│   │   │   │   ├── 1.jpg           # Step 1
+│   │   │   │   └── ...
+│   │   │   ├── 1/                   # Trajectory 1
+│   │   │   └── ...
+│   │   ├── door-open-v2/
+│   │   └── ...
+│   └── r3m-processed/              # R3M processed features
+│       └── r3m_MT10_20/
+│           ├── button-press-topdown-v2.pkl
+│           ├── door-open-v2.pkl
+│           └── ...
+└── r3m-processed/                  # Additional R3M data
+    └── r3m_MT10_20/
+├── mt50_task_prompts.json          # Task descriptions & prompts
+├── short_metaworld_loader.py       # Dataset loader
+└── requirements.txt
+```
+
+## 🎯 Tasks Included
+
+### Multi-Task 10 (MT10)
+- `button-press-topdown-v2` - Press button from above
+- `door-open-v2` - Open door by pulling handle  
+- `drawer-close-v2` - Close drawer
+- `drawer-open-v2` - Open drawer
+- `peg-insert-side-v2` - Insert peg into hole
+- `pick-place-v2` - Pick up object and place on target
+
+### Meta-Learning 10 (ML10)
+Additional tasks for meta-learning evaluation.
+
+## 📊 Data Format
+
+- **Images**: 224×224 RGB images in JPEG format
+- **States**: 7-dimensional robot state vectors (joint positions)
+- **Actions**: 4-dimensional continuous control actions
+- **Prompts**: Natural language task descriptions in 3 styles:
+  - `simple`: Brief task description
+  - `detailed`: Comprehensive task explanation  
+  - `task_specific`: Context-specific variations
+- **R3M Features**: Pre-processed visual representations using R3M model
+
+## 💾 Loading the Dataset
+
+The dataset comes with a comprehensive loader (`short_metaworld_loader.py`):
+
+```python
+# Load specific tasks
+mt10_tasks = [
+    "reach-v2", "push-v2", "pick-place-v2", "door-open-v2", 
+    "drawer-open-v2", "drawer-close-v2", "button-press-topdown-v2",
+    "button-press-v2", "button-press-wall-v2", "button-press-topdown-wall-v2"
+]
+dataset = load_short_metaworld("./", tasks=mt10_tasks)
+
+# Load all available tasks
+dataset = load_short_metaworld("./")
+
+# Get dataset statistics
+stats = dataset.get_dataset_stats()
+print(f"Total steps: {stats['total_steps']}")
+print(f"Tasks: {stats['tasks']}")
+
+# Get task-specific prompts
+task_info = dataset.get_task_info("pick-place-v2")
+print(task_info['detailed'])  # Detailed task description
+```
+
+## 🔬 Research Applications
+
+This dataset is designed for:
+
+- **Multi-task Reinforcement Learning**: Train policies across multiple manipulation tasks
+- **Imitation Learning**: Learn from demonstration trajectories
+- **Vision-Language Robotics**: Connect visual observations with natural language instructions
+- **Meta-Learning**: Adapt quickly to new manipulation tasks
+- **Robot Policy Training**: End-to-end visuomotor control
+
+## 📈 Dataset Statistics
+
+- **Total trajectories**: 2,000 (100 per task × 20 tasks)
+- **Total steps**: ~40,000 (20 steps per trajectory)
+- **Image resolution**: 224×224 RGB
+- **State dimension**: 7 (robot joint positions)
+- **Action dimension**: 4 (continuous control)
+- **Dataset size**: ~1.9GB
+
+## 🛠️ Installation
+
+```bash
+pip install torch torchvision Pillow numpy
+```
+
+## 📖 Citation
+
+If you use this dataset, please cite:
+
+```bibtex
+@inproceedings{yu2020meta,
+  title={Meta-world: A benchmark and evaluation for multi-task and meta reinforcement learning},
+  author={Yu, Tianhe and Quillen, Deirdre and He, Zhanpeng and Julian, Ryan and Hausman, Karol and Finn, Chelsea and Levine, Sergey},
+  booktitle={Conference on robot learning},
+  pages={1094--1100},
+  year={2020},
+  organization={PMLR}
+}
+
+@inproceedings{nair2022r3m,
+  title={R3M: A Universal Visual Representation for Robot Manipulation},
+  author={Nair, Suraj and Rajeswaran, Aravind and Kumar, Vikash and Finn, Chelsea and Gupta, Abhinav},
+  booktitle={Conference on Robot Learning},
+  pages={892--902},
+  year={2023},
+  organization={PMLR}
+}
+```
+
+## 📧 Contact
+
+- Original dataset: liangzx@connect.hku.hk
+- Questions about this upload: Open an issue in the dataset repository
+
+## ⚖️ License
+
+MIT License - See LICENSE file for details.

README.txt

@@ -0,0 +1,33 @@
+1. Overview
+Short-MetaWorld is a dataset rendered from modified environment of Meta-World [1], which contains Multi-Task10(MT10) and Meta-Learning10(ML10) in total 20 tasks with 100 successful trajectories for each task. Each trajectory is padded to 20 steps.
+
+2. File Structure
+This directory contains 3 sub-directories.
+
+└── short-MetaWorld
+    ├── task_description.py           # language instructions for each task
+    ├── img_only                      # rendered visual inputs for all tasks
+    │   ├── button-press-topdown-v2   # task name
+    │   │     ├── 0                   # trajectory id     
+    │   │     │   ├── 0.jpg           # step 0 observation (224*224)
+    │   │     │	  ├── 1.jpg           # step 1 observation
+    │   │     │   └── ...
+    │   │     ├── 1
+    │   │     └── ...
+    │   ├── door-open-v2
+    │   └── ...
+    ├── unprocessed     	   # trajectory actions with visual observations (256*256)
+    │   ├── unprocessed_MT10_20    # task name
+    │   │   ├── data.pkl           # a file contains all 10 tasks
+    │   │   ├── door-open-v2.pkl   # door-open task file
+    │   │   └── ...
+    │   └── unprocessed_ML10_20
+    └── r3m-processed              # trajectory actions with visual obs processed by R3M [2]
+
+3. Contact
+If you have any questions, please contact liangzx@connect.hku.hk
+
+[1] Yu, Tianhe, et al. "Meta-world: A benchmark and evaluation for multi-task and meta reinforcement learning." Conference on robot learning. PMLR, 2020.
+[2] Nair, Suraj, et al. "R3M: A Universal Visual Representation for Robot Manipulation." Conference on Robot Learning. PMLR, 2023.
+
+

mt50_task_prompts.json

@@ -0,0 +1,313 @@
+{
+  "reach-v2": {
+    "simple": "Reach the target location with the end-effector.",
+    "detailed": "Move the robot's hand to touch the highlighted target position. Ensure the end-effector is precisely aligned with the target for optimal task completion. This task requires careful control of the robot's arm to avoid overshooting the target.",
+    "task_specific": "The target is located at the far end of the workspace, requiring full extension of the arm."
+  },
+  "push-v2": {
+    "simple": "Push the object to the goal position.",
+    "detailed": "Use the gripper to push the small cylinder until it reaches the marked target location. Maintain a steady force to ensure the object moves smoothly without toppling. This task emphasizes the importance of controlled linear motion.",
+    "task_specific": "The target location is slightly elevated, requiring an upward push."
+  },
+  "pick-place-v2": {
+    "simple": "Pick up the object and place it on the target.",
+    "detailed": "Grasp the cube, lift it and accurately place it onto the designated goal site. Ensure the cube is stable and centered on the target to prevent it from falling. This task tests the precision of the pick-and-place operation.",
+    "task_specific": "The goal site is surrounded by obstacles, requiring careful navigation."
+  },
+  "door-open-v2": {
+    "simple": "Open the door by pulling the handle.",
+    "detailed": "Grip the door handle, pull it towards you to fully open the door. Apply a consistent force to overcome the door's resistance. This task requires understanding the mechanics of door operation.",
+    "task_specific": "The door is heavier than usual, requiring additional force."
+  },
+  "drawer-open-v2": {
+    "simple": "Open the drawer.",
+    "detailed": "Grab the drawer handle and pull it outwards until the drawer is fully open. Ensure the drawer is pulled straight to avoid jamming. This task focuses on the smoothness of the pulling action.",
+    "task_specific": "The drawer contains fragile items, necessitating a gentle pull."
+  },
+  "drawer-close-v2": {
+    "simple": "Close the drawer.",
+    "detailed": "Push the drawer until it is completely closed. Apply a steady force to ensure the drawer closes evenly. This task highlights the importance of even force distribution.",
+    "task_specific": "The drawer has a soft-close mechanism, requiring a gentle push at the end."
+  },
+  "button-press-topdown-v2": {
+    "simple": "Press the button from above.",
+    "detailed": "Move the gripper top-down to press the red button until it clicks. Ensure the button is pressed with sufficient force to activate it. This task tests the accuracy of vertical motion.",
+    "task_specific": "The button is located on a sloped surface, requiring angle adjustment."
+  },
+  "button-press-v2": {
+    "simple": "Press the button.",
+    "detailed": "Approach and press the blue button on the front panel. Apply enough pressure to ensure the button is fully depressed. This task requires precision in targeting the button.",
+    "task_specific": "The button is smaller than usual, requiring precise alignment."
+  },
+  "button-press-wall-v2": {
+    "simple": "Press the wall-mounted button.",
+    "detailed": "Move forward to press the button located on the vertical wall. Ensure the button is pressed with a direct force to avoid slipping. This task emphasizes the importance of approach angle.",
+    "task_specific": "The button is recessed, requiring a deeper press."
+  },
+  "button-press-topdown-wall-v2": {
+    "simple": "Press the wall button from above.",
+    "detailed": "Descend onto the wall-mounted button and press it. Apply a controlled force to ensure the button is activated without damaging the wall. This task tests the precision of top-down approaches.",
+    "task_specific": "The button is surrounded by a protective casing, requiring careful navigation."
+  },
+  "door-close-v2": {
+    "simple": "Close the door.",
+    "detailed": "Push the door handle away to fully close the door. Apply a consistent force to ensure the door latches securely. This task requires understanding the mechanics of door closure.",
+    "task_specific": "The door has a spring mechanism, requiring a firm push."
+  },
+  "hammer-v2": {
+    "simple": "Hammer the nail.",
+    "detailed": "Pick up the hammer and hit the nail until it is flush. Ensure each strike is accurate to avoid bending the nail. This task tests the precision and force of hammering actions.",
+    "task_specific": "The nail is positioned at an angle, requiring careful alignment."
+  },
+  "handle-press-side-v2": {
+    "simple": "Press the handle sideways.",
+    "detailed": "Push the handle laterally to activate it. Ensure the handle is fully depressed to engage the mechanism. This task emphasizes the importance of lateral force application.",
+    "task_specific": "The handle is stiff, requiring additional force."
+  },
+  "handle-press-v2": {
+    "simple": "Press the handle down.",
+    "detailed": "Press the top of the handle downward until the mechanism clicks. Apply a steady force to ensure the handle is fully engaged. This task tests the precision of downward force application.",
+    "task_specific": "The handle is located at an awkward angle, requiring wrist adjustment."
+  },
+  "handle-pull-v2": {
+    "simple": "Pull the handle.",
+    "detailed": "Grip and pull the handle towards you. Apply a consistent force to ensure the handle is fully extended. This task requires understanding the mechanics of handle operation.",
+    "task_specific": "The handle is slippery, requiring a firm grip."
+  },
+  "handle-pull-side-v2": {
+    "simple": "Pull the handle sideways.",
+    "detailed": "Pull the handle laterally along its slot. Ensure the handle is fully extended to engage the mechanism. This task emphasizes the importance of lateral motion.",
+    "task_specific": "The handle is partially obstructed, requiring careful navigation."
+  },
+  "lever-pull-v2": {
+    "simple": "Pull the lever.",
+    "detailed": "Grab the lever and pull it down to the end. Apply a steady force to ensure the lever is fully engaged. This task tests the precision of downward force application.",
+    "task_specific": "The lever is rusted, requiring additional force."
+  },
+  "peg-insert-side-v2": {
+    "simple": "Insert the peg into the hole.",
+    "detailed": "Pick the peg and slide it into the side hole until fully seated. Ensure the peg is aligned with the hole to avoid jamming. This task tests the precision of insertion actions.",
+    "task_specific": "The hole is slightly smaller, requiring precise alignment."
+  },
+  "pick-place-wall-v2": {
+    "simple": "Pick object and place it on the wall shelf.",
+    "detailed": "Lift the cube then place it onto the shelf attached to the wall. Ensure the cube is stable and centered on the shelf to prevent it from falling. This task tests the precision of pick-and-place operations.",
+    "task_specific": "The shelf is narrow, requiring careful placement."
+  },
+  "push-wall-v2": {
+    "simple": "Push the object along the wall to the target.",
+    "detailed": "From the side, push the cylinder along the wall until it reaches the goal mark. Maintain a steady force to ensure the object moves smoothly without toppling. This task emphasizes the importance of controlled linear motion.",
+    "task_specific": "The wall is uneven, requiring angle adjustment."
+  },
+  "reach-wall-v2": {
+    "simple": "Reach to the wall target.",
+    "detailed": "Extend the gripper to touch the target point on the wall. Ensure the end-effector is precisely aligned with the target for optimal task completion. This task requires careful control of the robot's arm to avoid overshooting the target.",
+    "task_specific": "The target is located at a high point, requiring full extension of the arm."
+  },
+  "shelf-place-v2": {
+    "simple": "Place the object on the shelf.",
+    "detailed": "Pick up the cube and place it carefully onto the shelf surface. Ensure the cube is stable and centered on the shelf to prevent it from falling. This task tests the precision of pick-and-place operations.",
+    "task_specific": "The shelf is cluttered, requiring careful navigation."
+  },
+  "sweep-into-v2": {
+    "simple": "Sweep the object into the bowl.",
+    "detailed": "Use the tool to sweep the ball into the container. Ensure the ball is fully inside the container to complete the task. This task tests the precision of sweeping actions.",
+    "task_specific": "The container is small, requiring precise aim."
+  },
+  "sweep-v2": {
+    "simple": "Sweep the object off the table.",
+    "detailed": "Sweep the debris across the tabletop and drop it over the edge. Ensure all debris is cleared from the table to complete the task. This task tests the precision of sweeping actions.",
+    "task_specific": "The debris is scattered, requiring multiple sweeps."
+  },
+  "window-open-v2": {
+    "simple": "Open the window.",
+    "detailed": "Slide the window latch and pull to open it. Apply a consistent force to ensure the window opens smoothly. This task requires understanding the mechanics of window operation.",
+    "task_specific": "The window is stuck, requiring additional force."
+  },
+  "window-close-v2": {
+    "simple": "Close the window.",
+    "detailed": "Push the window until it is fully closed and latched. Apply a steady force to ensure the window closes evenly. This task highlights the importance of even force distribution.",
+    "task_specific": "The window has a soft-close mechanism, requiring a gentle push at the end."
+  },
+  "coffee-button-v2": {
+    "simple": "Press the coffee machine button.",
+    "detailed": "Press the brew button on the coffee machine panel. Ensure the button is pressed with sufficient force to activate it. This task tests the accuracy of button pressing actions.",
+    "task_specific": "The button is smaller than usual, requiring precise alignment."
+  },
+  "coffee-pull-v2": {
+    "simple": "Pull the coffee lever.",
+    "detailed": "Pull down the lever to start the coffee extraction. Apply a steady force to ensure the lever is fully engaged. This task tests the precision of downward force application.",
+    "task_specific": "The lever is stiff, requiring additional force."
+  },
+  "coffee-push-v2": {
+    "simple": "Push the coffee lever back up.",
+    "detailed": "Push the lever upward to reset the coffee machine. Apply a steady force to ensure the lever is fully engaged. This task tests the precision of upward force application.",
+    "task_specific": "The lever is located at an awkward angle, requiring wrist adjustment."
+  },
+  "dial-turn-v2": {
+    "simple": "Turn the dial to the target angle.",
+    "detailed": "Grip the dial knob and rotate it until it points at the target value. Ensure the dial is turned smoothly to avoid overshooting the target. This task tests the precision of rotational actions.",
+    "task_specific": "The dial is stiff, requiring additional force."
+  },
+  "disassemble-v2": {
+    "simple": "Disassemble the object by pulling it apart.",
+    "detailed": "Pull the two pieces apart to disassemble the structure. Apply a steady force to ensure the pieces separate cleanly. This task tests the precision of disassembly actions.",
+    "task_specific": "The pieces are tightly fitted, requiring additional force."
+  },
+  "door-lock-v2": {
+    "simple": "Lock the door.",
+    "detailed": "Rotate the lock mechanism to secure the door. Ensure the lock is fully engaged to prevent the door from opening. This task tests the precision of rotational actions.",
+    "task_specific": "The lock is stiff, requiring additional force."
+  },
+  "hand-insert-v2": {
+    "simple": "Insert the peg with the gripper hand.",
+    "detailed": "Align and insert the gripper into the slot until fully seated. Ensure the gripper is aligned with the slot to avoid jamming. This task tests the precision of insertion actions.",
+    "task_specific": "The slot is narrow, requiring precise alignment."
+  },
+  "handover-v2": {
+    "simple": "Hand over the object.",
+    "detailed": "Pick up the object and move it to the hand-over position. Ensure the object is stable during the transfer to prevent dropping. This task tests the precision of handover actions.",
+    "task_specific": "The hand-over position is elevated, requiring careful lifting."
+  },
+  "lamp-turn-on-v2": {
+    "simple": "Turn on the lamp.",
+    "detailed": "Flip the lamp switch upwards to turn it on. Ensure the switch is fully engaged to activate the lamp. This task tests the precision of switch flipping actions.",
+    "task_specific": "The switch is located at an awkward angle, requiring wrist adjustment."
+  },
+  "pour-v2": {
+    "simple": "Pour the liquid into the cup.",
+    "detailed": "Lift the pitcher and pour its contents into the mug until full. Ensure the liquid is poured smoothly to avoid spillage. This task tests the precision of pouring actions.",
+    "task_specific": "The mug is small, requiring precise aim."
+  },
+  "soccer-v2": {
+    "simple": "Kick the ball into the goal.",
+    "detailed": "Push the soccer ball forward until it crosses the goal line. Ensure the ball is kicked with sufficient force to reach the goal. This task tests the precision of kicking actions.",
+    "task_specific": "The goal is narrow, requiring precise aim."
+  },
+  "stick-push-v2": {
+    "simple": "Push the stick to the target.",
+    "detailed": "Use the stick to push the object toward the goal location. Ensure the stick is applied with a steady force to avoid overshooting the target. This task tests the precision of pushing actions.",
+    "task_specific": "The target is elevated, requiring an upward push."
+  },
+  "stick-pull-v2": {
+    "simple": "Pull the stick object towards you.",
+    "detailed": "Hook the stick and pull it to the designated area. Ensure the stick is applied with a steady force to avoid overshooting the target. This task tests the precision of pulling actions.",
+    "task_specific": "The target is recessed, requiring a deeper pull."
+  },
+  "switch-v2": {
+    "simple": "Flip the switch.",
+    "detailed": "Toggle the electrical switch from off to on. Ensure the switch is fully engaged to activate the circuit. This task tests the precision of switch toggling actions.",
+    "task_specific": "The switch is located at an awkward angle, requiring wrist adjustment."
+  },
+  "basketball-v2": {
+    "simple": "Shoot the basketball into the hoop.",
+    "detailed": "Lift and throw the ball so it passes through the basket. Ensure the ball is thrown with sufficient force to reach the basket. This task tests the precision of throwing actions.",
+    "task_specific": "The basket is elevated, requiring a high throw."
+  },
+  "faucet-open-v2": {
+    "simple": "Open the faucet.",
+    "detailed": "Rotate the faucet handle to start water flow. Ensure the handle is turned smoothly to avoid overshooting the target. This task tests the precision of rotational actions.",
+    "task_specific": "The handle is stiff, requiring additional force."
+  },
+  "faucet-close-v2": {
+    "simple": "Close the faucet.",
+    "detailed": "Turn the faucet handle to stop the water. Ensure the handle is turned smoothly to avoid overshooting the target. This task tests the precision of rotational actions.",
+    "task_specific": "The handle is stiff, requiring additional force."
+  },
+  "ladle-pick-v2": {
+    "simple": "Pick up the ladle.",
+    "detailed": "Grab the ladle handle and lift it from the pot. Ensure the ladle is lifted smoothly to avoid spillage. This task tests the precision of lifting actions.",
+    "task_specific": "The ladle is heavy, requiring additional force."
+  },
+  "safepick-v2": {
+    "simple": "Pick up the object from the safe.",
+    "detailed": "Open the safe and retrieve the object inside. Ensure the object is handled carefully to avoid damage. This task tests the precision of retrieval actions.",
+    "task_specific": "The object is fragile, requiring gentle handling."
+  },
+  "safepush-v2": {
+    "simple": "Push the safe door closed.",
+    "detailed": "Push on the safe door until it clicks shut. Ensure the door is pushed with sufficient force to engage the lock. This task tests the precision of pushing actions.",
+    "task_specific": "The door is heavy, requiring additional force."
+  },
+  "scale-v2": {
+    "simple": "Place the object on the scale.",
+    "detailed": "Pick up the cube and put it onto the scale platform. Ensure the cube is placed gently to avoid tipping the scale. This task tests the precision of placement actions.",
+    "task_specific": "The scale is sensitive, requiring gentle placement."
+  },
+  "spoon-pick-v2": {
+    "simple": "Pick up the spoon.",
+    "detailed": "Grip the spoon handle and lift it. Ensure the spoon is lifted smoothly to avoid spillage. This task tests the precision of lifting actions.",
+    "task_specific": "The spoon is slippery, requiring a firm grip."
+  },
+  "tray-pick-v2": {
+    "simple": "Pick up the tray.",
+    "detailed": "Slide under the tray and lift it from the table. Ensure the tray is lifted smoothly to avoid spillage. This task tests the precision of lifting actions.",
+    "task_specific": "The tray is heavy, requiring additional force."
+  },
+  "utensil-pick-v2": {
+    "simple": "Pick up the utensil.",
+    "detailed": "Grab the utensil handle and lift it off the table. Ensure the utensil is lifted smoothly to avoid spillage. This task tests the precision of lifting actions.",
+    "task_specific": "The utensil is slippery, requiring a firm grip."
+  },
+  "basketball-v3": {
+    "simple": "Shoot the basketball into the hoop.",
+    "detailed": "Lift and throw the ball so it passes through the basket. Ensure the ball is thrown with sufficient force to reach the basket. This task tests the precision of throwing actions.",
+    "task_specific": "The basket is elevated, requiring a high throw."
+  },
+  "button-press-topdown-v3": {
+    "simple": "Press the button from above.",
+    "detailed": "Move the gripper top-down to press the red button until it clicks. Ensure the button is pressed with sufficient force to activate it. This task tests the accuracy of vertical motion.",
+    "task_specific": "The button is located on a sloped surface, requiring angle adjustment."
+  },
+  "door-open-v3": {
+    "simple": "Open the door by pulling the handle.",
+    "detailed": "Grip the door handle, pull it towards you to fully open the door. Apply a consistent force to overcome the door's resistance. This task requires understanding the mechanics of door operation.",
+    "task_specific": "The door is heavier than usual, requiring additional force."
+  },
+  "drawer-close-v3": {
+    "simple": "Close the drawer.",
+    "detailed": "Push the drawer until it is completely closed. Apply a steady force to ensure the drawer closes evenly. This task highlights the importance of even force distribution.",
+    "task_specific": "The drawer has a soft-close mechanism, requiring a gentle push at the end."
+  },
+  "peg-insert-side-v3": {
+    "simple": "Insert the peg into the hole.",
+    "detailed": "Pick the peg and slide it into the side hole until fully seated. Ensure the peg is aligned with the hole to avoid jamming. This task tests the precision of insertion actions.",
+    "task_specific": "The hole is slightly smaller, requiring precise alignment."
+  },
+  "pick-place-v3": {
+    "simple": "Pick up the object and place it on the target.",
+    "detailed": "Grasp the cube, lift it and accurately place it onto the designated goal site. Ensure the cube is stable and centered on the target to prevent it from falling. This task tests the precision of the pick-and-place operation.",
+    "task_specific": "The goal site is surrounded by obstacles, requiring careful navigation."
+  },
+  "push-v3": {
+    "simple": "Push the object to the goal position.",
+    "detailed": "Use the gripper to push the small cylinder until it reaches the marked target location. Maintain a steady force to ensure the object moves smoothly without toppling. This task emphasizes the importance of controlled linear motion.",
+    "task_specific": "The target location is slightly elevated, requiring an upward push."
+  },
+  "reach-v3": {
+    "simple": "Reach the target location with the end-effector.",
+    "detailed": "Move the robot's hand to touch the highlighted target position. Ensure the end-effector is precisely aligned with the target for optimal task completion. This task requires careful control of the robot's arm to avoid overshooting the target.",
+    "task_specific": "The target is located at the far end of the workspace, requiring full extension of the arm."
+  },
+  "sweep-v3": {
+    "simple": "Sweep the object off the table.",
+    "detailed": "Sweep the debris across the tabletop and drop it over the edge. Ensure all debris is cleared from the table to complete the task. This task tests the precision of sweeping actions.",
+    "task_specific": "The debris is scattered, requiring multiple sweeps."
+  },
+  "window-open-v3": {
+    "simple": "Open the window.",
+    "detailed": "Slide the window latch and pull to open it. Apply a consistent force to ensure the window opens smoothly. This task requires understanding the mechanics of window operation.",
+    "task_specific": "The window is stuck, requiring additional force."
+  },
+  "stick-pull-v3": {
+    "simple": "Pull the stick object towards you.",
+    "detailed": "Hook the stick and pull it to the designated area. Ensure the stick is applied with a steady force to avoid overshooting the target. This task tests the precision of pulling actions.",
+    "task_specific": "The target is recessed, requiring a deeper pull."
+  },
+  "handle-pull-v3": {
+    "simple": "Pull the handle.",
+    "detailed": "Grip and pull the handle towards you. Apply a consistent force to ensure the handle is fully extended. This task requires understanding the mechanics of handle operation.",
+    "task_specific": "The handle is slippery, requiring a firm grip."
+  }
+}
+

requirements.txt

@@ -0,0 +1,5 @@
+# Short-MetaWorld Dataset Requirements
+torch>=1.9.0
+torchvision>=0.10.0
+Pillow>=8.0.0
+numpy>=1.20.0

short_metaworld_loader.py

@@ -0,0 +1,241 @@
+#!/usr/bin/env python3
+"""
+Short-MetaWorld Dataset Loader
+Loads the dataset with proper structure preservation and task prompts.
+"""
+
+import os
+import pickle
+import json
+import glob
+from pathlib import Path
+from PIL import Image
+import torch
+import torchvision.transforms as T
+from torch.utils.data import Dataset
+
+class ShortMetaWorldDataset(Dataset):
+    """
+    Short-MetaWorld dataset loader that preserves the original structure.
+    
+    Args:
+        data_root (str): Path to the dataset root directory
+        task_list (list): List of tasks to load (default: all available tasks)
+        image_size (int): Target image size for transforms (default: 224)
+        transform (callable): Optional custom transform
+        load_prompts (bool): Whether to load task prompts (default: True)
+    """
+    
+    def __init__(self, data_root, task_list=None, image_size=224, transform=None, load_prompts=True):
+        self.data_root = Path(data_root)
+        self.image_size = image_size
+        
+        # Load task prompts
+        self.prompts = {}
+        if load_prompts:
+            prompt_file = self.data_root / "mt50_task_prompts.json"
+            if prompt_file.exists():
+                with open(prompt_file, 'r') as f:
+                    self.prompts = json.load(f)
+                print(f"📖 Loaded {len(self.prompts)} task prompts")
+            else:
+                print("⚠️ Task prompts not found, using fallback prompts")
+        
+        # Set up paths
+        self.img_root = self.data_root / "short-MetaWorld" / "short-MetaWorld" / "img_only"
+        self.data_pkl_root = self.data_root / "short-MetaWorld" / "r3m-processed" / "r3m_MT10_20"
+        
+        # Discover available tasks
+        available_tasks = []
+        if self.data_pkl_root.exists():
+            for pkl_file in self.data_pkl_root.glob("*.pkl"):
+                task_name = pkl_file.stem
+                if (self.img_root / task_name).exists():
+                    available_tasks.append(task_name)
+        
+        # Filter tasks if task_list provided
+        if task_list is not None:
+            self.tasks = [task for task in task_list if task in available_tasks]
+        else:
+            self.tasks = available_tasks
+        
+        print(f"📊 Loading {len(self.tasks)} tasks: {self.tasks}")
+        
+        # Default transform
+        if transform is None:
+            self.transform = T.Compose([
+                T.Resize((image_size, image_size)),
+                T.ToTensor(),
+            ])
+        else:
+            self.transform = transform
+        
+        # Load all trajectories
+        self.trajectories = self._load_trajectories()
+        print(f"✅ Loaded {len(self.trajectories)} trajectory steps")
+    
+    def _load_trajectories(self):
+        """Load all trajectory data"""
+        all_trajectories = []
+        
+        for task in self.tasks:
+            # Load pickle data
+            pkl_path = self.data_pkl_root / f"{task}.pkl"
+            if not pkl_path.exists():
+                print(f"⚠️ Pickle file not found: {pkl_path}")
+                continue
+                
+            with open(pkl_path, "rb") as f:
+                data = pickle.load(f)
+            
+            # Get image directories
+            task_img_dir = self.img_root / task
+            if not task_img_dir.exists():
+                print(f"⚠️ Image directory not found: {task_img_dir}")
+                continue
+            
+            # Process each trajectory
+            traj_dirs = sorted(task_img_dir.glob("*"), key=lambda x: int(x.name))
+            
+            for traj_idx, traj_dir in enumerate(traj_dirs):
+                if traj_idx >= len(data['actions']):
+                    continue
+                
+                # Get image paths
+                img_paths = sorted(traj_dir.glob("*.jpg"), 
+                                 key=lambda x: int(x.stem))
+                
+                num_steps = len(data['actions'][traj_idx])
+                num_images = len(img_paths)
+                
+                # Use minimum length to handle mismatched data
+                min_steps = min(num_images, num_steps)
+                if min_steps < 1:
+                    continue
+                
+                # Create trajectory entries
+                for step_idx in range(min_steps):
+                    trajectory_entry = {
+                        'task_name': task,
+                        'trajectory_id': traj_idx,
+                        'step_id': step_idx,
+                        'image_path': str(img_paths[step_idx]),
+                        'state': data['state'][traj_idx][step_idx][:7],  # First 7 dims
+                        'action': data['actions'][traj_idx][step_idx],
+                        'prompt': self._get_prompt(task)
+                    }
+                    all_trajectories.append(trajectory_entry)
+        
+        return all_trajectories
+    
+    def _get_prompt(self, task_name):
+        """Get prompt for a task"""
+        if task_name in self.prompts:
+            # Use simple prompt by default
+            return self.prompts[task_name].get('simple', f"Perform the task: {task_name.replace('-', ' ')}")
+        else:
+            return f"Perform the task: {task_name.replace('-', ' ')}"
+    
+    def get_task_info(self, task_name):
+        """Get comprehensive task information"""
+        if task_name in self.prompts:
+            return self.prompts[task_name]
+        return {"simple": f"Perform the task: {task_name.replace('-', ' ')}"}
+    
+    def get_available_tasks(self):
+        """Get list of available tasks"""
+        return self.tasks.copy()
+    
+    def get_dataset_stats(self):
+        """Get dataset statistics"""
+        task_counts = {}
+        for traj in self.trajectories:
+            task = traj['task_name']
+            task_counts[task] = task_counts.get(task, 0) + 1
+        
+        return {
+            'total_steps': len(self.trajectories),
+            'num_tasks': len(self.tasks),
+            'task_step_counts': task_counts,
+            'tasks': self.tasks
+        }
+    
+    def __len__(self):
+        return len(self.trajectories)
+    
+    def __getitem__(self, idx):
+        """Get a single trajectory step"""
+        traj = self.trajectories[idx]
+        
+        # Load and transform image
+        image = Image.open(traj['image_path']).convert("RGB")
+        image_tensor = self.transform(image)
+        
+        # Convert to tensors
+        state = torch.tensor(traj['state'], dtype=torch.float32)
+        action = torch.tensor(traj['action'], dtype=torch.float32)
+        
+        return {
+            'image': image_tensor,
+            'state': state,
+            'action': action,
+            'prompt': traj['prompt'],
+            'task_name': traj['task_name'],
+            'trajectory_id': traj['trajectory_id'],
+            'step_id': traj['step_id']
+        }
+
+# Example usage functions
+def load_short_metaworld(data_root, tasks=None, image_size=224):
+    """
+    Convenience function to load the dataset.
+    
+    Args:
+        data_root (str): Path to dataset root
+        tasks (list): List of tasks to load (None for all)
+        image_size (int): Image size for transforms
+    
+    Returns:
+        ShortMetaWorldDataset: The loaded dataset
+    """
+    return ShortMetaWorldDataset(
+        data_root=data_root,
+        task_list=tasks,
+        image_size=image_size
+    )
+
+def get_mt10_tasks():
+    """Get the MT10 task list"""
+    return [
+        "reach-v2", "push-v2", "pick-place-v2", "door-open-v2", "drawer-open-v2",
+        "drawer-close-v2", "button-press-topdown-v2", "button-press-v2", 
+        "button-press-wall-v2", "button-press-topdown-wall-v2"
+    ]
+
+def demo_usage():
+    """Demonstrate how to use the dataset"""
+    print("📖 Short-MetaWorld Dataset Usage Example")
+    print("=" * 50)
+    
+    # Load dataset
+    dataset = load_short_metaworld("./", tasks=get_mt10_tasks())
+    
+    # Print stats
+    stats = dataset.get_dataset_stats()
+    print(f"📊 Dataset Statistics:")
+    print(f"   Total steps: {stats['total_steps']}")
+    print(f"   Number of tasks: {stats['num_tasks']}")
+    print(f"   Available tasks: {stats['tasks']}")
+    
+    # Get a sample
+    if len(dataset) > 0:
+        sample = dataset[0]
+        print(f"\n📋 Sample data:")
+        print(f"   Image shape: {sample['image'].shape}")
+        print(f"   State shape: {sample['state'].shape}")
+        print(f"   Action shape: {sample['action'].shape}")
+        print(f"   Task: {sample['task_name']}")
+        print(f"   Prompt: {sample['prompt']}")
+
+if __name__ == "__main__":
+    demo_usage()