Upload folder using huggingface_hub

.gitattributes

@@ -92,3 +92,7 @@ LLaVA-Next-3D/results/scanrefer/prompt-llavanext-qwen-3d-uniform-16gpu-uniform.j
 LLaVA-Next-3D/results/scanrefer/prompt-llavanext-qwen-3d-uniform-16gpu-double-vg-wo3Dpe-uniform.jsonl filter=lfs diff=lfs merge=lfs -text
 LLaVA-Next-3D/data/scannet filter=lfs diff=lfs merge=lfs -text
 LLaVA-Next-3D/results/scanrefer filter=lfs diff=lfs merge=lfs -text
+LLaVA-Next-3D/results/3dvg_bench_ori/prompt-llavanext-qwen-uniform-16bs-scanrefer-wosam-uniform_absolute_distance.jsonl filter=lfs diff=lfs merge=lfs -text
+LLaVA-Next-3D/results/3dvg_bench_ori/prompt-llavanext-qwen-uniform-16bs-scanrefer-wosam-uniform_relative_distance.jsonl filter=lfs diff=lfs merge=lfs -text
+LLaVA-Next-3D/results/3dvg_bench_ori/prompt-llavanext-qwen-uniform-try5-uniform_absolute_distance.jsonl filter=lfs diff=lfs merge=lfs -text
+LLaVA-Next-3D/results/3dvg_bench_ori/prompt-llavanext-qwen-uniform-try5-uniform_relative_distance.jsonl filter=lfs diff=lfs merge=lfs -text

LLaVA-Next-3D/results/3dvg_bench_ori/prompt-llavanext-qwen-uniform-16bs-scanrefer-wosam-uniform_absolute_distance.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:351d4aa692a5781c3ef0eb58c1246217c0dbc40fe0b0527ec2bdd09df706ba48
+size 11073996

LLaVA-Next-3D/results/3dvg_bench_ori/prompt-llavanext-qwen-uniform-16bs-scanrefer-wosam-uniform_relative_distance.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8e6f961542e680acca26616b18b2e9b5b1c8e1b72c7f6537fce5346b83f4d99a
+size 10976432

LLaVA-Next-3D/results/3dvg_bench_ori/prompt-llavanext-qwen-uniform-try5-uniform_absolute_distance.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5987f58d5fe21d47392ad6d47d31ae7c830e2104ebcb4a26be5cdb58fa079b27
+size 14403501

LLaVA-Next-3D/results/3dvg_bench_ori/prompt-llavanext-qwen-uniform-try5-uniform_relative_distance.jsonl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dcc06658892e58573449265638e251e8fe2185fad043f006eb1388f7221ad2e1
+size 14202811

LLaVA-Next-3D/results/3dvg_bench_ori/prompt-llavanext-qwen-uniform-try5-uniform_small_object_sample.jsonl

@@ -0,0 +1,15 @@
+{"dataset": "SCANNET", "sample_id": 4, "scan_id": "scannet/scene0599_02", "pred_sentence": "3: [0.72, 0.12, 1.0, 0.34], Frame-4: [0.51, 0.2, 0.78, 0.43], Frame-5: [0.18, 0.35, 0.46, 0.68], Frame-12: [0.41, 0.01, 0.57, 0.05], Frame-13: [0.08, 0.41, 0.3, 0.54], Frame-21: [0.48, 0.26, 0.67, 0.43], Frame-22: [0.14, 0.3, 0.33, 0.68], Frame-23: [0.16, 0.0, 0.3, 0.3], Frame-30: [0.61, 0.04, 0.8, 0.2].", "prompt": "The video captures 3D spatial information of a scene. Please focus on the spatial relationships in the video and answer the following questions.\n<image> A wooden chair with a brown seat.", "pred_response": {"data/scannet/posed_images/scene0599_02/00150.jpg": [0.72, 0.12, 1.0, 0.34], "data/scannet/posed_images/scene0599_02/00230.jpg": [0.51, 0.2, 0.78, 0.43], "data/scannet/posed_images/scene0599_02/00310.jpg": [0.18, 0.35, 0.46, 0.68], "data/scannet/posed_images/scene0599_02/00870.jpg": [0.41, 0.01, 0.57, 0.05], "data/scannet/posed_images/scene0599_02/00950.jpg": [0.08, 0.41, 0.3, 0.54], "data/scannet/posed_images/scene0599_02/01590.jpg": [0.48, 0.26, 0.67, 0.43], "data/scannet/posed_images/scene0599_02/01670.jpg": [0.14, 0.3, 0.33, 0.68], "data/scannet/posed_images/scene0599_02/01750.jpg": [0.16, 0.0, 0.3, 0.3], "data/scannet/posed_images/scene0599_02/02310.jpg": [0.61, 0.04, 0.8, 0.2]}, "gt_response": [0.16027873754501343, -0.19867190718650818, 0.41337287425994873, 0.515144944190979, 0.5371779203414917, 0.8692851066589355], "model_id": "llavanext-qwen-uniform-try5", "question_type": "small_object"}
+{"dataset": "SCANNET", "sample_id": 7, "scan_id": "scannet/scene0599_02", "pred_sentence": "7: [0.46, 0.45, 0.51, 0.49], Frame-15: [0.32, 0.48, 0.37, 0.52].", "prompt": "The video captures 3D spatial information of a scene. Please focus on the spatial relationships in the video and answer the following questions.\n<image> A cell phone is sitting on a wooden table.", "pred_response": {"data/scannet/posed_images/scene0599_02/00470.jpg": [0.46, 0.45, 0.51, 0.49], "data/scannet/posed_images/scene0599_02/01110.jpg": [0.32, 0.48, 0.37, 0.52]}, "gt_response": [0.626155436038971, 0.16719213873147964, 0.41738826874643564, 0.6412688493728638, 0.6449674814939499, 0.8774370979517698], "model_id": "llavanext-qwen-uniform-try5", "question_type": "small_object"}
+{"dataset": "SCANNET", "sample_id": 1, "scan_id": "scannet/scene0599_02", "pred_sentence": "5: [0.8, 0.43, 0.85, 0.47], Frame-13: [0.9, 0.6, 0.97, 0.65], Frame-22: [0.72, 0.38, 0.78, 0.43].", "prompt": "The video captures 3D spatial information of a scene. Please focus on the spatial relationships in the video and answer the following questions.\n<image> A screw with a spiral pattern.", "pred_response": {"data/scannet/posed_images/scene0599_02/00310.jpg": [0.8, 0.43, 0.85, 0.47], "data/scannet/posed_images/scene0599_02/00950.jpg": [0.9, 0.6, 0.97, 0.65], "data/scannet/posed_images/scene0599_02/01670.jpg": [0.72, 0.38, 0.78, 0.43]}, "gt_response": [-1.1879916787147522, 0.020348578691482544, 0.9686046838760376, 0.10984361171722412, 0.135649174451828, 0.7286810874938965], "model_id": "llavanext-qwen-uniform-try5", "question_type": "small_object"}
+{"dataset": "SCANNET", "sample_id": 0, "scan_id": "scannet/scene0599_02", "pred_sentence": "12: [0.79, 0.01, 1.0, 0.38], Frame-14: [0.23, 0.09, 0.33, 0.38], Frame-22: [0.51, 0.28, 0.6, 0.38], Frame-23: [0.55, 0.01, 0.64, 0.06].", "prompt": "The video captures 3D spatial information of a scene. Please focus on the spatial relationships in the video and answer the following questions.\n<image> The back of a person's head.", "pred_response": {"data/scannet/posed_images/scene0599_02/00870.jpg": [0.79, 0.01, 1.0, 0.38], "data/scannet/posed_images/scene0599_02/01030.jpg": [0.23, 0.09, 0.33, 0.38], "data/scannet/posed_images/scene0599_02/01670.jpg": [0.51, 0.28, 0.6, 0.38], "data/scannet/posed_images/scene0599_02/01750.jpg": [0.55, 0.01, 0.64, 0.06]}, "gt_response": [0.3022874593734741, 0.45377445220947266, 0.046521518379449844, 3.178182601928711, 4.496844291687012, 0.19208692759275436], "model_id": "llavanext-qwen-uniform-try5", "question_type": "small_object"}
+{"dataset": "SCANNET", "sample_id": 2, "scan_id": "scannet/scene0599_02", "pred_sentence": "4: [0.78, 0.17, 0.85, 0.22], Frame-5: [0.39, 0.31, 0.45, 0.36], Frame-13: [0.45, 0.41, 0.5, 0.44], Frame-21: [0.73, 0.25, 0.78, 0.28], Frame-22: [0.34, 0.29, 0.39, 0.33].", "prompt": "The video captures 3D spatial information of a scene. Please focus on the spatial relationships in the video and answer the following questions.\n<image> A knife with a black handle.", "pred_response": {"data/scannet/posed_images/scene0599_02/00230.jpg": [0.78, 0.17, 0.85, 0.22], "data/scannet/posed_images/scene0599_02/00310.jpg": [0.39, 0.31, 0.45, 0.36], "data/scannet/posed_images/scene0599_02/00950.jpg": [0.45, 0.41, 0.5, 0.44], "data/scannet/posed_images/scene0599_02/01590.jpg": [0.73, 0.25, 0.78, 0.28], "data/scannet/posed_images/scene0599_02/01670.jpg": [0.34, 0.29, 0.39, 0.33]}, "gt_response": [-0.8266880214214325, -1.6382505297660828, 1.5492892861366272, 0.34376198053359985, 0.23970186710357666, 0.3640240430831909], "model_id": "llavanext-qwen-uniform-try5", "question_type": "small_object"}
+{"dataset": "SCANNET", "sample_id": 6, "scan_id": "scannet/scene0599_02", "pred_sentence": "1: [0.02, 0.0, 0.14, 0.69], Frame-8: [0.01, 0.0, 0.66, 0.35], Frame-9: [0.02, 0.0, 0.13, 0.15], Frame-17: [0.01, 0.0, 0.51, 0.26], Frame-18: [0.03, 0.0, 0.12, 0.43], Frame-31: [0.01, 0.0, 0.1, 0.76], Frame-32: [0.02, 0.0, 0.24, 0.5].", "prompt": "The video captures 3D spatial information of a scene. Please focus on the spatial relationships in the video and answer the following questions.\n<image> A white wall with a light shining on it.", "pred_response": {"data/scannet/posed_images/scene0599_02/00000.jpg": [0.02, 0.0, 0.14, 0.69], "data/scannet/posed_images/scene0599_02/00550.jpg": [0.01, 0.0, 0.66, 0.35], "data/scannet/posed_images/scene0599_02/00630.jpg": [0.02, 0.0, 0.13, 0.15], "data/scannet/posed_images/scene0599_02/01270.jpg": [0.01, 0.0, 0.51, 0.26], "data/scannet/posed_images/scene0599_02/01350.jpg": [0.03, 0.0, 0.12, 0.43], "data/scannet/posed_images/scene0599_02/02390.jpg": [0.01, 0.0, 0.1, 0.76], "data/scannet/posed_images/scene0599_02/02470.jpg": [0.02, 0.0, 0.24, 0.5]}, "gt_response": [-1.2765843272209167, 1.1532146707177162, 1.4085116386413574, 0.2697528600692749, 1.8991275280714035, 1.6238937377929688], "model_id": "llavanext-qwen-uniform-try5", "question_type": "small_object"}
+{"dataset": "SCANNET", "sample_id": 12, "scan_id": "scannet/scene0599_02", "pred_sentence": "6: [0.65, 0.0, 1.0, 0.36], Frame-7: [0.01, 0.0, 0.33, 0.34], Frame-14: [0.66, 0.0, 1.0, 0.29], Frame-15: [0.01, 0.0, 0.63, 0.23], Frame-22: [0.92, 0.0, 1.0, 0.38], Frame-24: [0.54, 0.0, 1.0, 0.13], Frame-25: [0.01, 0.0, 0.41, 0.27].", "prompt": "The video captures 3D spatial information of a scene. Please focus on the spatial relationships in the video and answer the following questions.\n<image> The sky is black.", "pred_response": {"data/scannet/posed_images/scene0599_02/00390.jpg": [0.65, 0.0, 1.0, 0.36], "data/scannet/posed_images/scene0599_02/00470.jpg": [0.01, 0.0, 0.33, 0.34], "data/scannet/posed_images/scene0599_02/01030.jpg": [0.66, 0.0, 1.0, 0.29], "data/scannet/posed_images/scene0599_02/01110.jpg": [0.01, 0.0, 0.63, 0.23], "data/scannet/posed_images/scene0599_02/01670.jpg": [0.92, 0.0, 1.0, 0.38], "data/scannet/posed_images/scene0599_02/01830.jpg": [0.54, 0.0, 1.0, 0.13], "data/scannet/posed_images/scene0599_02/01910.jpg": [0.01, 0.0, 0.41, 0.27]}, "gt_response": [1.8700751662254333, 0.3069557547569275, 0.956647599581629, 0.20162761211395264, 4.113177418708801, 1.9032436115667224], "model_id": "llavanext-qwen-uniform-try5", "question_type": "small_object"}
+{"dataset": "SCANNET", "sample_id": 3, "scan_id": "scannet/scene0599_02", "pred_sentence": "3: [0.72, 0.12, 1.0, 0.34], Frame-4: [0.51, 0.2, 0.78, 0.43], Frame-5: [0.18, 0.35, 0.46, 0.68], Frame-12: [0.41, 0.01, 0.57, 0.05], Frame-13: [0.08, 0.41, 0.3, 0.54], Frame-21: [0.48, 0.26, 0.67, 0.43], Frame-22: [0.14, 0.3, 0.33, 0.68], Frame-23: [0.16, 0.0, 0.3, 0.3], Frame-30: [0.61, 0.04, 0.8, 0.2].", "prompt": "The video captures 3D spatial information of a scene. Please focus on the spatial relationships in the video and answer the following questions.\n<image> A wooden chair with a brown seat.", "pred_response": {"data/scannet/posed_images/scene0599_02/00150.jpg": [0.72, 0.12, 1.0, 0.34], "data/scannet/posed_images/scene0599_02/00230.jpg": [0.51, 0.2, 0.78, 0.43], "data/scannet/posed_images/scene0599_02/00310.jpg": [0.18, 0.35, 0.46, 0.68], "data/scannet/posed_images/scene0599_02/00870.jpg": [0.41, 0.01, 0.57, 0.05], "data/scannet/posed_images/scene0599_02/00950.jpg": [0.08, 0.41, 0.3, 0.54], "data/scannet/posed_images/scene0599_02/01590.jpg": [0.48, 0.26, 0.67, 0.43], "data/scannet/posed_images/scene0599_02/01670.jpg": [0.14, 0.3, 0.33, 0.68], "data/scannet/posed_images/scene0599_02/01750.jpg": [0.16, 0.0, 0.3, 0.3], "data/scannet/posed_images/scene0599_02/02310.jpg": [0.61, 0.04, 0.8, 0.2]}, "gt_response": [0.16926710307598114, 1.2876077890396118, 0.38904501125216484, 0.9938067495822906, 3.3490450382232666, 0.821777917444706], "model_id": "llavanext-qwen-uniform-try5", "question_type": "small_object"}
+{"dataset": "SCANNET", "sample_id": 8, "scan_id": "scannet/scene0599_02", "pred_sentence": "12: [0.78, 0.0, 0.99, 0.36], Frame-15: [0.98, 0.36, 1.0, 0.48], Frame-24: [0.79, 0.31, 1.0, 0.77], Frame-25: [0.35, 0.46, 0.62, 0.83].", "prompt": "The video captures 3D spatial information of a scene. Please focus on the spatial relationships in the video and answer the following questions.\n<image> The arm of a person.", "pred_response": {"data/scannet/posed_images/scene0599_02/00870.jpg": [0.78, 0.0, 0.99, 0.36], "data/scannet/posed_images/scene0599_02/01110.jpg": [0.98, 0.36, 1.0, 0.48], "data/scannet/posed_images/scene0599_02/01830.jpg": [0.79, 0.31, 1.0, 0.77], "data/scannet/posed_images/scene0599_02/01910.jpg": [0.35, 0.46, 0.62, 0.83]}, "gt_response": [-0.20232124626636505, 0.7036197036504745, 0.4204786643385887, 0.4789751470088959, 0.4709928333759308, 0.8961470872163773], "model_id": "llavanext-qwen-uniform-try5", "question_type": "small_object"}
+{"dataset": "SCANNET", "sample_id": 9, "scan_id": "scannet/scene0599_02", "pred_sentence": "3: [0.82, 0.16, 1.0, 0.32], Frame-4: [0.54, 0.21, 0.74, 0.43], Frame-5: [0.19, 0.36, 0.39, 0.61], Frame-12: [0.41, 0.01, 0.61, 0.04], Frame-13: [0.13, 0.43, 0.32, 0.54], Frame-21: [0.51, 0.28, 0.67, 0.42], Frame-22: [0.16, 0.32, 0.33, 0.6], Frame-23: [0.2, 0.0, 0.34, 0.22], Frame-30: [0.61, 0.04, 0.75, 0.16].", "prompt": "The video captures 3D spatial information of a scene. Please focus on the spatial relationships in the video and answer the following questions.\n<image> A wooden bench.", "pred_response": {"data/scannet/posed_images/scene0599_02/00150.jpg": [0.82, 0.16, 1.0, 0.32], "data/scannet/posed_images/scene0599_02/00230.jpg": [0.54, 0.21, 0.74, 0.43], "data/scannet/posed_images/scene0599_02/00310.jpg": [0.19, 0.36, 0.39, 0.61], "data/scannet/posed_images/scene0599_02/00870.jpg": [0.41, 0.01, 0.61, 0.04], "data/scannet/posed_images/scene0599_02/00950.jpg": [0.13, 0.43, 0.32, 0.54], "data/scannet/posed_images/scene0599_02/01590.jpg": [0.51, 0.28, 0.67, 0.42], "data/scannet/posed_images/scene0599_02/01670.jpg": [0.16, 0.32, 0.33, 0.6], "data/scannet/posed_images/scene0599_02/01750.jpg": [0.2, 0.0, 0.34, 0.22], "data/scannet/posed_images/scene0599_02/02310.jpg": [0.61, 0.04, 0.75, 0.16]}, "gt_response": [-0.24786773324012756, 1.877582311630249, 0.43661763751879334, 0.3189149498939514, 0.4453005790710449, 0.8842277033254504], "model_id": "llavanext-qwen-uniform-try5", "question_type": "small_object"}
+{"dataset": "SCANNET", "sample_id": 14, "scan_id": "scannet/scene0599_02", "pred_sentence": "12: [0.61, 0.15, 0.71, 0.39], Frame-23: [0.87, 0.43, 0.99, 0.58], Frame-24: [0.02, 0.25, 0.12, 0.33].", "prompt": "The video captures 3D spatial information of a scene. Please focus on the spatial relationships in the video and answer the following questions.\n<image> A pair of black and white shoes.", "pred_response": {"data/scannet/posed_images/scene0599_02/00870.jpg": [0.61, 0.15, 0.71, 0.39], "data/scannet/posed_images/scene0599_02/01750.jpg": [0.87, 0.43, 0.99, 0.58], "data/scannet/posed_images/scene0599_02/01830.jpg": [0.02, 0.25, 0.12, 0.33]}, "gt_response": [1.6486455202102661, -0.3962952196598053, 0.22468932019546628, 0.30964159965515137, 0.45015662908554077, 0.44347682502120733], "model_id": "llavanext-qwen-uniform-try5", "question_type": "small_object"}
+{"dataset": "SCANNET", "sample_id": 10, "scan_id": "scannet/scene0599_02", "pred_sentence": "5: [0.73, 0.4, 0.84, 0.5], Frame-6: [0.28, 0.34, 0.4, 0.42], Frame-13: [0.79, 0.6, 0.91, 0.7], Frame-14: [0.49, 0.19, 0.6, 0.28], Frame-22: [0.76, 0.38, 0.92, 0.5], Frame-23: [0.76, 0.04, 0.91, 0.12], Frame-24: [0.38, 0.04, 0.49, 0.15].", "prompt": "The video captures 3D spatial information of a scene. Please focus on the spatial relationships in the video and answer the following questions.\n<image> The back of a person's hand.", "pred_response": {"data/scannet/posed_images/scene0599_02/00310.jpg": [0.73, 0.4, 0.84, 0.5], "data/scannet/posed_images/scene0599_02/00390.jpg": [0.28, 0.34, 0.4, 0.42], "data/scannet/posed_images/scene0599_02/00950.jpg": [0.79, 0.6, 0.91, 0.7], "data/scannet/posed_images/scene0599_02/01030.jpg": [0.49, 0.19, 0.6, 0.28], "data/scannet/posed_images/scene0599_02/01670.jpg": [0.76, 0.38, 0.92, 0.5], "data/scannet/posed_images/scene0599_02/01750.jpg": [0.76, 0.04, 0.91, 0.12], "data/scannet/posed_images/scene0599_02/01830.jpg": [0.38, 0.04, 0.49, 0.15]}, "gt_response": [-0.3589605391025543, 2.507765531539917, 0.4581643361598253, 0.2793652415275574, 0.4195270538330078, 0.8811715804040432], "model_id": "llavanext-qwen-uniform-try5", "question_type": "small_object"}
+{"dataset": "SCANNET", "sample_id": 5, "scan_id": "scannet/scene0599_02", "pred_sentence": "2: [0.39, 0.52, 1.0, 1.0], Frame-3: [0.02, 0.23, 1.0, 1.0], Frame-4: [0.27, 0.2, 1.0, 1.0], Frame-5: [0.18, 0.36, 1.0, 1.0], Frame-6: [0.01, 0.39, 1.0, 1.0], Frame-7: [0.02, 0.39, 0.97, 1.0], Frame-10: [0.85, 0.61, 1.0, 1.0], Frame-11: [0.61, 0.53, 1.0, 1.0], Frame-12: [0.01, 0.0, 0.82, 1.0], Frame-13: [0.01, 0.43, 0.53, 1.0], Frame-14: [0.01, 0.18, 0.5, 1.0], Frame-15: [0.02, 0.41, 0.54, 1.0], Frame-16: [0.02, 0.49, 0.71, 1.0], Frame-20: [0.8, 0.57, 1.0, 1.0], Frame-21: [0.39, 0.25, 1.0, 1.0], Frame-22: [0.14, 0.3, 1.0, 1.0], Frame-23: [0.2, 0.0, 0.99, 1.0], Frame-24: [0.0, 0.0, 0.46, 1.0], Frame-29: [0.97, 0.71, 1.0, 1.0], Frame-30: [0.", "prompt": "The video captures 3D spatial information of a scene. Please focus on the spatial relationships in the video and answer the following questions.\n<image> A wooden table.", "pred_response": {"data/scannet/posed_images/scene0599_02/00070.jpg": [0.39, 0.52, 1.0, 1.0], "data/scannet/posed_images/scene0599_02/00150.jpg": [0.02, 0.23, 1.0, 1.0], "data/scannet/posed_images/scene0599_02/00230.jpg": [0.27, 0.2, 1.0, 1.0], "data/scannet/posed_images/scene0599_02/00310.jpg": [0.18, 0.36, 1.0, 1.0], "data/scannet/posed_images/scene0599_02/00390.jpg": [0.01, 0.39, 1.0, 1.0], "data/scannet/posed_images/scene0599_02/00470.jpg": [0.02, 0.39, 0.97, 1.0], "data/scannet/posed_images/scene0599_02/00710.jpg": [0.85, 0.61, 1.0, 1.0], "data/scannet/posed_images/scene0599_02/00790.jpg": [0.61, 0.53, 1.0, 1.0], "data/scannet/posed_images/scene0599_02/00870.jpg": [0.01, 0.0, 0.82, 1.0], "data/scannet/posed_images/scene0599_02/00950.jpg": [0.01, 0.43, 0.53, 1.0], "data/scannet/posed_images/scene0599_02/01030.jpg": [0.01, 0.18, 0.5, 1.0], "data/scannet/posed_images/scene0599_02/01110.jpg": [0.02, 0.41, 0.54, 1.0], "data/scannet/posed_images/scene0599_02/01190.jpg": [0.02, 0.49, 0.71, 1.0], "data/scannet/posed_images/scene0599_02/01510.jpg": [0.8, 0.57, 1.0, 1.0], "data/scannet/posed_images/scene0599_02/01590.jpg": [0.39, 0.25, 1.0, 1.0], "data/scannet/posed_images/scene0599_02/01670.jpg": [0.14, 0.3, 1.0, 1.0], "data/scannet/posed_images/scene0599_02/01750.jpg": [0.2, 0.0, 0.99, 1.0], "data/scannet/posed_images/scene0599_02/01830.jpg": [0.0, 0.0, 0.46, 1.0], "data/scannet/posed_images/scene0599_02/02230.jpg": [0.97, 0.71, 1.0, 1.0]}, "gt_response": [-0.12867745757102966, 0.22920931223779917, 0.420411067083478, 0.5767800211906433, 0.4797020982950926, 0.8741179369390011], "model_id": "llavanext-qwen-uniform-try5", "question_type": "small_object"}
+{"dataset": "SCANNET", "sample_id": 11, "scan_id": "scannet/scene0599_02", "pred_sentence": "12: [0.79, 0.01, 1.0, 0.38], Frame-14: [0.23, 0.09, 0.33, 0.38], Frame-22: [0.51, 0.28, 0.6, 0.38], Frame-23: [0.55, 0.01, 0.64, 0.06].", "prompt": "The video captures 3D spatial information of a scene. Please focus on the spatial relationships in the video and answer the following questions.\n<image> The back of a person's head.", "pred_response": {"data/scannet/posed_images/scene0599_02/00870.jpg": [0.79, 0.01, 1.0, 0.38], "data/scannet/posed_images/scene0599_02/01030.jpg": [0.23, 0.09, 0.33, 0.38], "data/scannet/posed_images/scene0599_02/01670.jpg": [0.51, 0.28, 0.6, 0.38], "data/scannet/posed_images/scene0599_02/01750.jpg": [0.55, 0.01, 0.64, 0.06]}, "gt_response": [0.6974533200263977, 1.8360600471496582, 0.4532242347486317, 0.3970503807067871, 0.47058916091918945, 0.8757983585819602], "model_id": "llavanext-qwen-uniform-try5", "question_type": "small_object"}
+{"dataset": "SCANNET", "sample_id": 13, "scan_id": "scannet/scene0599_02", "pred_sentence": "1: [0.02, 0.0, 0.12, 0.69], Frame-8: [0.01, 0.0, 0.61, 0.35], Frame-9: [0.02, 0.0, 0.18, 0.2], Frame-17: [0.01, 0.0, 0.51, 0.26], Frame-18: [0.03, 0.0, 0.13, 0.43], Frame-31: [0.01, 0.0, 0.09, 0.76], Frame-32: [0.02, 0.0, 0.24, 0.53].", "prompt": "The video captures 3D spatial information of a scene. Please focus on the spatial relationships in the video and answer the following questions.\n<image> The wall is white.", "pred_response": {"data/scannet/posed_images/scene0599_02/00000.jpg": [0.02, 0.0, 0.12, 0.69], "data/scannet/posed_images/scene0599_02/00550.jpg": [0.01, 0.0, 0.61, 0.35], "data/scannet/posed_images/scene0599_02/00630.jpg": [0.02, 0.0, 0.18, 0.2], "data/scannet/posed_images/scene0599_02/01270.jpg": [0.01, 0.0, 0.51, 0.26], "data/scannet/posed_images/scene0599_02/01350.jpg": [0.03, 0.0, 0.13, 0.43], "data/scannet/posed_images/scene0599_02/02390.jpg": [0.01, 0.0, 0.09, 0.76], "data/scannet/posed_images/scene0599_02/02470.jpg": [0.02, 0.0, 0.24, 0.53]}, "gt_response": [1.9768236875534058, 0.5828660726547241, 1.1121440827846527, 0.24813342094421387, 1.7795336246490479, 0.8373213410377502], "model_id": "llavanext-qwen-uniform-try5", "question_type": "small_object"}