Seq Of Colors

Language Instruction

Observe which colored cubes appear during the cue, wait, then touch all of them in any order and press the center button.

Task Description

Sequential cue, unordered execution: colors are shown sequentially, but the remembered targets can be touched in any order.

Source

• Module: mikasa_robo_suite.vla.memory_envs.seq_of_colors_vla

• Source file: mikasa_robo_suite/vla/memory_envs/seq_of_colors_vla.py

Difficulty and Parameters

More colors and Long variants increase presentation-time and retention burden.

Variants

Env ID	Horizon	Data Source
SeqOfColors3-Long-VLA-v0	800	MP
SeqOfColors3-VLA-v0	400	MP
SeqOfColors5-Long-VLA-v0	1000	MP
SeqOfColors5-VLA-v0	400	MP
SeqOfColors7-Long-VLA-v0	1200	MP
SeqOfColors7-VLA-v0	400	MP

Run Example

import gymnasium as gym
import torch

import mikasa_robo_suite.vla.memory_envs  # registers VLA env IDs
from mikasa_robo_suite.vla.utils.apply_wrappers import apply_mikasa_vla_wrappers

env = gym.make(
    "SeqOfColors3-VLA-v0",
    num_envs=1,
    obs_mode="rgb",
    control_mode="pd_ee_delta_pose",
    render_mode="all",
)
env = apply_mikasa_vla_wrappers(env)
obs, info = env.reset(seed=42)
for _ in range(env.max_episode_steps):
    action = env.action_space.sample()
    if not torch.is_tensor(action):
        action = torch.as_tensor(action, device=env.unwrapped.device)
    obs, reward, terminated, truncated, info = env.step(action)
env.close()

Dataset Collection

Motion-planning MIKASA-Robo-90 variants use planner plus replay collection:

uv run python mikasa_robo_suite/vla/dataset_collectors/get_mikasa_robo_datasets_motion_planning.py \
  --env-id SeqOfColors3-VLA-v0 \
  --path-to-save-data data_mikasa_robo \
  --num-train-data 250 \
  --max-attempts 5000 \
  --seed 0

Render Videos

Generate a fresh MP4/GIF render with:

uv run python utils/prepare_benchmark_demo_videos.py \
  --tasks SeqOfColors3-VLA-v0 \
  --output-dir videos/benchmark_demos \
  --max-attempts-per-task 8 \
  --overwrite

Generated media stays under videos/ and should be published deliberately rather than committed by default.