CoRL
Collection
CoRL • 3 items • Updated • 1
images listlengths 2 2 | data_source large_stringclasses 4
values | prompt listlengths 2 2 | ability large_stringclasses 4
values | reward_model dict | metadata dict |
|---|---|---|---|---|---|
[
{
"bytes": [
255,
216,
255,
224,
0,
16,
74,
70,
73,
70,
0,
1,
1,
0,
0,
1,
0,
1,
0,
0,
255,
219,
0,
67,
0,
10,
7,
7,
8,
7,
6,
... | corl_l1_sim_val | [
{
"content": "You are analyzing two images from different robot perspectives viewing the same environment. These images represent two distinct robot viewpoints that may show overlapping or complementary parts of the scene. IMPORTANT FORMAT REQUIREMENTS:\n1. ALL reasoning must be enclosed within <think> </think>... | spatial_reasoning_object_counting | {
"ground_truth": "<think>\\nI need to analyze images from two robot perspectives to answer the question about 'How many fork and kettle are there in this space?'.\\n\\nLet me first examine the main perspective (Image A):\\n- I can see a fork located at [192, 66, 224, 70]\\n\\nNow examining the auxiliary perspective (Image B):\\n- I can see a fork located at [110, 143, 142, 146]\\n- There's a kettle positioned at coordinates [120, 120, 138, 142]\\n\\nAnalyzing spatial correspondence between the two perspectives:\\n- The fork_0 at [110, 143, 142, 146] in auxiliary perspective is not visible in main perspective\\n- The fork_1 at [192, 66, 224, 70] in main perspective is not visible in auxiliary perspective\\n- The kettle_0 at [120, 120, 138, 142] in auxiliary perspective is not visible in main perspective\\n\\nCombining observations from both perspectives, the total number of objects you asked me to count is 3.\\n</think>\\n\\n\\\\boxed{3}",
"style": "rule"
} | {
"category": null,
"dataset_name": "ObjectCount_4",
"grasping_instruction": null,
"image_name": null,
"init_pos": null,
"layout_id": 1,
"noise_objects": [
"cardboardbox_noise0",
"cup_noise1"
],
"objects": [
"fork_0",
"fork_1",
"kettle_0",
"cardboardbox_noise0",
"cup_noise1"
],
"original_answer": null,
"original_question": null,
"question_id": null,
"robots": [
"anymal_c",
"unitree_h1"
],
"source_dataset": null,
"target_center": null,
"target_coordinates": null,
"target_image": null,
"target_object": null,
"task_id": "7817_2-1",
"task_name": "count_objects_TWO_TYPES",
"task_type": null
} |
[
{
"bytes": [
255,
216,
255,
224,
0,
16,
74,
70,
73,
70,
0,
1,
1,
0,
0,
1,
0,
1,
0,
0,
255,
219,
0,
67,
0,
10,
7,
7,
8,
7,
6,
... | corl_l1_sim_val | [
{
"content": "You are analyzing two images from different robot perspectives viewing the same environment. These images represent two distinct robot viewpoints that may show overlapping or complementary parts of the scene. IMPORTANT FORMAT REQUIREMENTS:\n1. ALL reasoning must be enclosed within <think> </think>... | spatial_reasoning_object_counting | {
"ground_truth": "<think>\\nI need to analyze images from two robot perspectives to answer the question about 'Can you assess how many apple are present?'.\\n\\nLet me first examine the main perspective (Image A):\\n- I can see a apple located at [39, 14, 41, 17]\\n\\nNow examining the auxiliary perspective (Image B):\\n- I can see a apple located at [22, 25, 24, 27]\\n\\nAnalyzing spatial correspondence between the two perspectives:\\n- The apple_0 at [22, 25, 24, 27] in auxiliary perspective is not visible in main perspective\\n- The apple_1 at [39, 14, 41, 17] in main perspective is not visible in auxiliary perspective\\n\\nCombining observations from both perspectives, the total number of objects you asked me to count is 2.\\n</think>\\n\\n\\\\boxed{2}",
"style": "rule"
} | {
"category": null,
"dataset_name": "ObjectCount_1",
"grasping_instruction": null,
"image_name": null,
"init_pos": null,
"layout_id": 1,
"noise_objects": [
"green_peas_noise0",
"cardboardbox_noise1"
],
"objects": [
"apple_0",
"apple_1",
"green_peas_noise0",
"cardboardbox_noise1"
],
"original_answer": null,
"original_question": null,
"question_id": null,
"robots": [
"anymal_c",
"unitree_h1"
],
"source_dataset": null,
"target_center": null,
"target_coordinates": null,
"target_image": null,
"target_object": null,
"task_id": "9381_1-1",
"task_name": "count_objects_EASY",
"task_type": null
} |
[
{
"bytes": [
255,
216,
255,
224,
0,
16,
74,
70,
73,
70,
0,
1,
1,
0,
0,
1,
0,
1,
0,
0,
255,
219,
0,
67,
0,
10,
7,
7,
8,
7,
6,
... | corl_l1_sim_val | [
{
"content": "You are analyzing two images from different robot perspectives viewing the same environment. These images represent two distinct robot viewpoints that may show overlapping or complementary parts of the scene. IMPORTANT FORMAT REQUIREMENTS:\n1. ALL reasoning must be enclosed within <think> </think>... | spatial_reasoning_object_counting | {
"ground_truth": "<think>\\nI need to analyze images from two robot perspectives to answer the question about 'Could you tell me how many banana and hammer exist?'.\\n\\nLet me first examine the main perspective (Image A):\\n- I can see the first banana at [65, 84, 73, 90]\\n- I also see a second banana at [144, 174, 173, 215]\\n- I notice a hammer at [198, 67, 210, 75]\\n\\nNow examining the auxiliary perspective (Image B):\\n- In this view, I can see the first banana at [247, 76, 251, 82]\\n- I also notice a second banana located at [225, 173, 240, 211]\\n\\nAnalyzing spatial correspondence between the two perspectives:\\n- The banana_0 visible in main perspective at [65, 84, 73, 90] corresponds to the same object visible in auxiliary perspective at [247, 76, 251, 82]\\n- The banana_1 visible in main perspective at [144, 174, 173, 215] corresponds to the same object visible in auxiliary perspective at [225, 173, 240, 211]\\n- The hammer_0 at [198, 67, 210, 75] in main perspective is not visible in auxiliary perspective\\n\\nCombining observations from both perspectives, the total number of objects you asked me to count is 3.\\n</think>\\n\\n\\\\boxed{3}",
"style": "rule"
} | {
"category": null,
"dataset_name": "ObjectCount_5",
"grasping_instruction": null,
"image_name": null,
"init_pos": null,
"layout_id": 6,
"noise_objects": [
"shoe_noise0",
"kettle_noise1"
],
"objects": [
"banana_0",
"banana_1",
"hammer_0",
"shoe_noise0",
"kettle_noise1"
],
"original_answer": null,
"original_question": null,
"question_id": null,
"robots": [
"panda",
"fetch"
],
"source_dataset": null,
"target_center": null,
"target_coordinates": null,
"target_image": null,
"target_object": null,
"task_id": "6805_2-1",
"task_name": "count_objects_TWO_TYPES",
"task_type": null
} |
[
{
"bytes": [
255,
216,
255,
224,
0,
16,
74,
70,
73,
70,
0,
1,
1,
0,
0,
1,
0,
1,
0,
0,
255,
219,
0,
67,
0,
10,
7,
7,
8,
7,
6,
... | corl_l1_sim_val | [
{
"content": "You are analyzing two images from different robot perspectives viewing the same environment. These images represent two distinct robot viewpoints that may show overlapping or complementary parts of the scene. IMPORTANT FORMAT REQUIREMENTS:\n1. ALL reasoning must be enclosed within <think> </think>... | spatial_reasoning_object_counting | {
"ground_truth": "<think>\\nI need to analyze images from two robot perspectives to answer the question about 'How many pizza do you need to move?'.\\n\\nLet me first examine the main perspective (Image A):\\n- I can see the first pizza at [0, 27, 5, 30]\\n- I also see a second pizza at [24, 28, 30, 33]\\n\\nNow examining the auxiliary perspective (Image B):\\n- I can see a pizza located at [49, 27, 55, 30]\\n\\nAnalyzing spatial correspondence between the two perspectives:\\n- The pizza_0 visible in main perspective at [0, 27, 5, 30] corresponds to the same object visible in auxiliary perspective at [49, 27, 55, 30]\\n- The pizza_1 at [24, 28, 30, 33] in main perspective is not visible in auxiliary perspective\\n\\nCombining observations from both perspectives, the total number of objects you asked me to count is 2.\\n</think>\\n\\n\\\\boxed{2}",
"style": "rule"
} | {
"category": null,
"dataset_name": "ObjectCount_1",
"grasping_instruction": null,
"image_name": null,
"init_pos": null,
"layout_id": 0,
"noise_objects": [
"peach_noise0",
"spoon_noise1"
],
"objects": [
"pizza_0",
"pizza_1",
"peach_noise0",
"spoon_noise1"
],
"original_answer": null,
"original_question": null,
"question_id": null,
"robots": [
"unitree_h1"
],
"source_dataset": null,
"target_center": null,
"target_coordinates": null,
"target_image": null,
"target_object": null,
"task_id": "8319_1-1",
"task_name": "count_objects_EASY",
"task_type": null
} |
[
{
"bytes": [
255,
216,
255,
224,
0,
16,
74,
70,
73,
70,
0,
1,
1,
0,
0,
1,
0,
1,
0,
0,
255,
219,
0,
67,
0,
10,
7,
7,
8,
7,
6,
... | corl_l1_sim_val | [
{
"content": "You are analyzing two images from different robot perspectives viewing the same environment. These images represent two distinct robot viewpoints that may show overlapping or complementary parts of the scene. IMPORTANT FORMAT REQUIREMENTS:\n1. ALL reasoning must be enclosed within <think> </think>... | spatial_reasoning_object_counting | {
"ground_truth": "<think>\\nI need to analyze images from two robot perspectives to answer the question about 'How many pizza and hammer are located here?'.\\n\\nLet me first examine the main perspective (Image A):\\n- I can see a pizza located at [36, 22, 39, 25]\\n- There's a hammer positioned at coordinates [45, 20, 48, 22]\\n\\nNow examining the auxiliary perspective (Image B):\\n- I can see a hammer located at [10, 25, 11, 26]\\n\\nAnalyzing spatial correspondence between the two perspectives:\\n- The pizza_0 at [36, 22, 39, 25] in main perspective is not visible in auxiliary perspective\\n- The hammer_0 visible in main perspective at [45, 20, 48, 22] corresponds to the same object visible in auxiliary perspective at [10, 25, 11, 26]\\n\\nCombining observations from both perspectives, the total number of objects you asked me to count is 2.\\n</think>\\n\\n\\\\boxed{2}",
"style": "rule"
} | {
"category": null,
"dataset_name": "ObjectCount_1",
"grasping_instruction": null,
"image_name": null,
"init_pos": null,
"layout_id": 9,
"noise_objects": [
"wine_noise0",
"green_peas_noise1"
],
"objects": [
"pizza_0",
"hammer_0",
"wine_noise0",
"green_peas_noise1"
],
"original_answer": null,
"original_question": null,
"question_id": null,
"robots": [
"fetch"
],
"source_dataset": null,
"target_center": null,
"target_coordinates": null,
"target_image": null,
"target_object": null,
"task_id": "9761_2-1",
"task_name": "count_objects_TWO_TYPES",
"task_type": null
} |
[{"bytes":"/9j/4AAQSkZJRgABAQAAAQABAAD/2wBDAAoHBwgHBgoICAgLCgoLDhgQDg0NDh0VFhEYIx8lJCIfIiEmKzcvJik0K(...TRUNCATED) | corl_l1_sim_val | [{"content":"You are analyzing two images from different robot perspectives viewing the same environ(...TRUNCATED) | spatial_reasoning_object_counting | {"ground_truth":"<think>\\nI need to analyze images from two robot perspectives to answer the questi(...TRUNCATED) | {"category":null,"dataset_name":"ObjectCount_1","grasping_instruction":null,"image_name":null,"init_(...TRUNCATED) |
[{"bytes":"/9j/4AAQSkZJRgABAQAAAQABAAD/2wBDAAoHBwgHBgoICAgLCgoLDhgQDg0NDh0VFhEYIx8lJCIfIiEmKzcvJik0K(...TRUNCATED) | corl_l1_sim_val | [{"content":"You are analyzing two images from different robot perspectives viewing the same environ(...TRUNCATED) | spatial_reasoning_object_counting | {"ground_truth":"<think>\\nI need to analyze images from two robot perspectives to answer the questi(...TRUNCATED) | {"category":null,"dataset_name":"ObjectCount_4","grasping_instruction":null,"image_name":null,"init_(...TRUNCATED) |
[{"bytes":"/9j/4AAQSkZJRgABAQAAAQABAAD/2wBDAAoHBwgHBgoICAgLCgoLDhgQDg0NDh0VFhEYIx8lJCIfIiEmKzcvJik0K(...TRUNCATED) | corl_l1_sim_val | [{"content":"You are analyzing two images from different robot perspectives viewing the same environ(...TRUNCATED) | spatial_reasoning_object_counting | {"ground_truth":"<think>\\nI need to analyze images from two robot perspectives to answer the questi(...TRUNCATED) | {"category":null,"dataset_name":"ObjectCount_3","grasping_instruction":null,"image_name":null,"init_(...TRUNCATED) |
[{"bytes":"/9j/4AAQSkZJRgABAQAAAQABAAD/2wBDAAoHBwgHBgoICAgLCgoLDhgQDg0NDh0VFhEYIx8lJCIfIiEmKzcvJik0K(...TRUNCATED) | corl_l1_sim_val | [{"content":"You are analyzing two images from different robot perspectives viewing the same environ(...TRUNCATED) | spatial_reasoning_object_counting | {"ground_truth":"<think>\\nI need to analyze images from two robot perspectives to answer the questi(...TRUNCATED) | {"category":null,"dataset_name":"ObjectCount_4","grasping_instruction":null,"image_name":null,"init_(...TRUNCATED) |
[{"bytes":"/9j/4AAQSkZJRgABAQAAAQABAAD/2wBDAAoHBwgHBgoICAgLCgoLDhgQDg0NDh0VFhEYIx8lJCIfIiEmKzcvJik0K(...TRUNCATED) | corl_l1_sim_val | [{"content":"You are analyzing two images from different robot perspectives viewing the same environ(...TRUNCATED) | spatial_reasoning_object_counting | {"ground_truth":"<think>\\nI need to analyze images from two robot perspectives to answer the questi(...TRUNCATED) | {"category":null,"dataset_name":"ObjectCount_5","grasping_instruction":null,"image_name":null,"init_(...TRUNCATED) |
End of preview. Expand in Data Studio
No dataset card yet
- Downloads last month
- 12
Size of downloaded dataset files:
48.5 MB
Size of the auto-converted Parquet files:
48.5 MB
Number of rows:
2,000