{"case_path": "/root/Reseach/qwen-video/outputs/review/frames_64/sample_1_a1781_q0/case.json", "model_dir": "/model/ModelScope/Qwen/Qwen3-VL-4B-Instruct", "torch_version": "2.13.0+cu132", "transformers_version": "5.19.0", "status": "ok", "timing_scope": "cached_frames_to_answer", "question_id": "sample_1_a1781_q0", "question_timestamp_s": 4, "num_frames": 64, "input_frame_timestamps_s": [0.0, 0.04, 0.12, 0.16, 0.24, 0.28, 0.36, 0.44, 0.48, 0.56, 0.6, 0.68, 0.76, 0.8, 0.88, 0.92, 1.0, 1.04, 1.12, 1.2, 1.24, 1.32, 1.36, 1.44, 1.52, 1.56, 1.64, 1.68, 1.76, 1.84, 1.88, 1.96, 2.0, 2.08, 2.12, 2.2, 2.28, 2.32, 2.4, 2.44, 2.52, 2.6, 2.64, 2.72, 2.76, 2.84, 2.92, 2.96, 3.04, 3.08, 3.16, 3.2, 3.28, 3.36, 3.4, 3.48, 3.52, 3.6, 3.68, 3.72, 3.8, 3.84, 3.92, 4.0], "video_metadata": {"fps": 25.0, "total_num_frames": 4707, "frames_indices": [0, 1, 3, 4, 6, 7, 9, 11, 12, 14, 15, 17, 19, 20, 22, 23, 25, 26, 28, 30, 31, 33, 34, 36, 38, 39, 41, 42, 44, 46, 47, 49, 50, 52, 53, 55, 57, 58, 60, 61, 63, 65, 66, 68, 69, 71, 73, 74, 76, 77, 79, 80, 82, 84, 85, 87, 88, 90, 92, 93, 95, 96, 98, 100]}, "sampling": {"method": "uniform_time_floor", "history": "from_start_to_question", "max_n_frames": 64}, "video_preprocessing": {"size": {"shortest_edge": 1048576, "longest_edge": 4194304}, "do_sample_frames": false, "cap_pixels_per_frame": false}, "input_ids_shape": [1, 2259], "video_grid_thw": [[32, 12, 20]], "ground_truth": "D", "model_load_s": 0.00010246096644550562, "gpu": "NVIDIA GeForce RTX 4090", "inference_s": 0.28107347490731627, "peak_allocated_gib": 8.885725975036621, "peak_reserved_gib": 9.1015625, "generation": {"max_new_tokens": 32, "do_sample": false}, "warmup_performed": true, "raw_output": "D", "prediction": "D", "correct": 1, "total_s": 1.083931747940369}
{"case_path": "/root/Reseach/qwen-video/outputs/review/frames_64/sample_1_a1781_q1/case.json", "model_dir": "/model/ModelScope/Qwen/Qwen3-VL-4B-Instruct", "torch_version": "2.13.0+cu132", "transformers_version": "5.19.0", "status": "ok", "timing_scope": "cached_frames_to_answer", "question_id": "sample_1_a1781_q1", "question_timestamp_s": 53, "num_frames": 64, "input_frame_timestamps_s": [0.0, 0.84, 1.68, 2.52, 3.36, 4.2, 5.04, 5.88, 6.72, 7.56, 8.4, 9.24, 10.08, 10.92, 11.76, 12.6, 13.44, 14.28, 15.12, 15.96, 16.8, 17.64, 18.48, 19.32, 20.16, 21.0, 21.84, 22.68, 23.52, 24.36, 25.2, 26.04, 26.92, 27.76, 28.6, 29.44, 30.28, 31.12, 31.96, 32.8, 33.64, 34.48, 35.32, 36.16, 37.0, 37.84, 38.68, 39.52, 40.36, 41.2, 42.04, 42.88, 43.72, 44.56, 45.4, 46.24, 47.08, 47.92, 48.76, 49.6, 50.44, 51.28, 52.12, 53.0], "video_metadata": {"fps": 25.0, "total_num_frames": 4707, "frames_indices": [0, 21, 42, 63, 84, 105, 126, 147, 168, 189, 210, 231, 252, 273, 294, 315, 336, 357, 378, 399, 420, 441, 462, 483, 504, 525, 546, 567, 588, 609, 630, 651, 673, 694, 715, 736, 757, 778, 799, 820, 841, 862, 883, 904, 925, 946, 967, 988, 1009, 1030, 1051, 1072, 1093, 1114, 1135, 1156, 1177, 1198, 1219, 1240, 1261, 1282, 1303, 1325]}, "sampling": {"method": "uniform_time_floor", "history": "from_start_to_question", "max_n_frames": 64}, "video_preprocessing": {"size": {"shortest_edge": 1048576, "longest_edge": 4194304}, "do_sample_frames": false, "cap_pixels_per_frame": false}, "input_ids_shape": [1, 2282], "video_grid_thw": [[32, 12, 20]], "ground_truth": "C", "model_load_s": 0.00012618093751370907, "gpu": "NVIDIA GeForce RTX 4090", "inference_s": 0.2828223999822512, "peak_allocated_gib": 8.889913082122803, "peak_reserved_gib": 9.1015625, "generation": {"max_new_tokens": 32, "do_sample": false}, "warmup_performed": true, "raw_output": "C", "prediction": "C", "correct": 1, "total_s": 0.9282486820593476}
{"case_path": "/root/Reseach/qwen-video/outputs/review/frames_64/sample_1_a1782_q0/case.json", "model_dir": "/model/ModelScope/Qwen/Qwen3-VL-4B-Instruct", "torch_version": "2.13.0+cu132", "transformers_version": "5.19.0", "status": "ok", "timing_scope": "cached_frames_to_answer", "question_id": "sample_1_a1782_q0", "question_timestamp_s": 82, "num_frames": 64, "input_frame_timestamps_s": [0.0, 1.28, 2.6, 3.88, 5.2, 6.48, 7.8, 9.08, 10.4, 11.68, 13.0, 14.28, 15.6, 16.92, 18.2, 19.52, 20.8, 22.12, 23.4, 24.72, 26.0, 27.32, 28.6, 29.92, 31.2, 32.52, 33.84, 35.12, 36.44, 37.72, 39.04, 40.32, 41.64, 42.92, 44.24, 45.52, 46.84, 48.12, 49.44, 50.76, 52.04, 53.36, 54.64, 55.96, 57.24, 58.56, 59.84, 61.16, 62.44, 63.76, 65.04, 66.36, 67.68, 68.96, 70.28, 71.56, 72.88, 74.16, 75.48, 76.76, 78.08, 79.36, 80.68, 82.0], "video_metadata": {"fps": 25.0, "total_num_frames": 4707, "frames_indices": [0, 32, 65, 97, 130, 162, 195, 227, 260, 292, 325, 357, 390, 423, 455, 488, 520, 553, 585, 618, 650, 683, 715, 748, 780, 813, 846, 878, 911, 943, 976, 1008, 1041, 1073, 1106, 1138, 1171, 1203, 1236, 1269, 1301, 1334, 1366, 1399, 1431, 1464, 1496, 1529, 1561, 1594, 1626, 1659, 1692, 1724, 1757, 1789, 1822, 1854, 1887, 1919, 1952, 1984, 2017, 2050]}, "sampling": {"method": "uniform_time_floor", "history": "from_start_to_question", "max_n_frames": 64}, "video_preprocessing": {"size": {"shortest_edge": 1048576, "longest_edge": 4194304}, "do_sample_frames": false, "cap_pixels_per_frame": false}, "input_ids_shape": [1, 2271], "video_grid_thw": [[32, 12, 20]], "ground_truth": "D", "model_load_s": 0.00011407106649130583, "gpu": "NVIDIA GeForce RTX 4090", "inference_s": 0.2846061979653314, "peak_allocated_gib": 8.887662887573242, "peak_reserved_gib": 9.1015625, "generation": {"max_new_tokens": 32, "do_sample": false}, "warmup_performed": true, "raw_output": "D", "prediction": "D", "correct": 1, "total_s": 0.9438949470641091}
{"case_path": "/root/Reseach/qwen-video/outputs/review/frames_64/sample_1_a1782_q1/case.json", "model_dir": "/model/ModelScope/Qwen/Qwen3-VL-4B-Instruct", "torch_version": "2.13.0+cu132", "transformers_version": "5.19.0", "status": "ok", "timing_scope": "cached_frames_to_answer", "question_id": "sample_1_a1782_q1", "question_timestamp_s": 80, "num_frames": 64, "input_frame_timestamps_s": [0.0, 1.24, 2.52, 3.8, 5.04, 6.32, 7.6, 8.88, 10.12, 11.4, 12.68, 13.96, 15.2, 16.48, 17.76, 19.04, 20.28, 21.56, 22.84, 24.12, 25.36, 26.64, 27.92, 29.2, 30.44, 31.72, 33.0, 34.28, 35.52, 36.8, 38.08, 39.36, 40.6, 41.88, 43.16, 44.44, 45.68, 46.96, 48.24, 49.52, 50.76, 52.04, 53.32, 54.6, 55.84, 57.12, 58.4, 59.68, 60.92, 62.2, 63.48, 64.76, 66.0, 67.28, 68.56, 69.84, 71.08, 72.36, 73.64, 74.92, 76.16, 77.44, 78.72, 80.0], "video_metadata": {"fps": 25.0, "total_num_frames": 4707, "frames_indices": [0, 31, 63, 95, 126, 158, 190, 222, 253, 285, 317, 349, 380, 412, 444, 476, 507, 539, 571, 603, 634, 666, 698, 730, 761, 793, 825, 857, 888, 920, 952, 984, 1015, 1047, 1079, 1111, 1142, 1174, 1206, 1238, 1269, 1301, 1333, 1365, 1396, 1428, 1460, 1492, 1523, 1555, 1587, 1619, 1650, 1682, 1714, 1746, 1777, 1809, 1841, 1873, 1904, 1936, 1968, 2000]}, "sampling": {"method": "uniform_time_floor", "history": "from_start_to_question", "max_n_frames": 64}, "video_preprocessing": {"size": {"shortest_edge": 1048576, "longest_edge": 4194304}, "do_sample_frames": false, "cap_pixels_per_frame": false}, "input_ids_shape": [1, 2265], "video_grid_thw": [[32, 12, 20]], "ground_truth": "D", "model_load_s": 0.00023829296696931124, "gpu": "NVIDIA GeForce RTX 4090", "inference_s": 0.28449431399349123, "peak_allocated_gib": 8.886435508728027, "peak_reserved_gib": 9.1015625, "generation": {"max_new_tokens": 32, "do_sample": false}, "warmup_performed": true, "raw_output": "D", "prediction": "D", "correct": 1, "total_s": 0.9400815739063546}
{"case_path": "/root/Reseach/qwen-video/outputs/review/frames_64/sample_1_a1783_q0/case.json", "model_dir": "/model/ModelScope/Qwen/Qwen3-VL-4B-Instruct", "torch_version": "2.13.0+cu132", "transformers_version": "5.19.0", "status": "ok", "timing_scope": "cached_frames_to_answer", "question_id": "sample_1_a1783_q0", "question_timestamp_s": 128, "num_frames": 64, "input_frame_timestamps_s": [0.0, 2.0, 4.04, 6.08, 8.12, 10.12, 12.16, 14.2, 16.24, 18.28, 20.28, 22.32, 24.36, 26.4, 28.44, 30.44, 32.48, 34.52, 36.56, 38.6, 40.6, 42.64, 44.68, 46.72, 48.76, 50.76, 52.8, 54.84, 56.88, 58.92, 60.92, 62.96, 65.0, 67.04, 69.04, 71.08, 73.12, 75.16, 77.2, 79.2, 81.24, 83.28, 85.32, 87.36, 89.36, 91.4, 93.44, 95.48, 97.52, 99.52, 101.56, 103.6, 105.64, 107.68, 109.68, 111.72, 113.76, 115.8, 117.84, 119.84, 121.88, 123.92, 125.96, 128.0], "video_metadata": {"fps": 25.0, "total_num_frames": 4707, "frames_indices": [0, 50, 101, 152, 203, 253, 304, 355, 406, 457, 507, 558, 609, 660, 711, 761, 812, 863, 914, 965, 1015, 1066, 1117, 1168, 1219, 1269, 1320, 1371, 1422, 1473, 1523, 1574, 1625, 1676, 1726, 1777, 1828, 1879, 1930, 1980, 2031, 2082, 2133, 2184, 2234, 2285, 2336, 2387, 2438, 2488, 2539, 2590, 2641, 2692, 2742, 2793, 2844, 2895, 2946, 2996, 3047, 3098, 3149, 3200]}, "sampling": {"method": "uniform_time_floor", "history": "from_start_to_question", "max_n_frames": 64}, "video_preprocessing": {"size": {"shortest_edge": 1048576, "longest_edge": 4194304}, "do_sample_frames": false, "cap_pixels_per_frame": false}, "input_ids_shape": [1, 2282], "video_grid_thw": [[32, 12, 20]], "ground_truth": "D", "model_load_s": 0.00012469198554754257, "gpu": "NVIDIA GeForce RTX 4090", "inference_s": 0.28431306895799935, "peak_allocated_gib": 8.889913082122803, "peak_reserved_gib": 9.1015625, "generation": {"max_new_tokens": 32, "do_sample": false}, "warmup_performed": true, "raw_output": "D", "prediction": "D", "correct": 1, "total_s": 0.9293124679243192}
{"case_path": "/root/Reseach/qwen-video/outputs/review/frames_64/sample_2_a1807_q0/case.json", "model_dir": "/model/ModelScope/Qwen/Qwen3-VL-4B-Instruct", "torch_version": "2.13.0+cu132", "transformers_version": "5.19.0", "status": "ok", "timing_scope": "cached_frames_to_answer", "question_id": "sample_2_a1807_q0", "question_timestamp_s": 10, "num_frames": 64, "input_frame_timestamps_s": [0.0, 0.12, 0.28, 0.44, 0.6, 0.76, 0.92, 1.08, 1.24, 1.4, 1.56, 1.72, 1.88, 2.04, 2.2, 2.36, 2.52, 2.68, 2.84, 3.0, 3.16, 3.32, 3.48, 3.64, 3.8, 3.96, 4.12, 4.28, 4.44, 4.6, 4.76, 4.92, 5.04, 5.2, 5.36, 5.52, 5.68, 5.84, 6.0, 6.16, 6.32, 6.48, 6.64, 6.8, 6.96, 7.12, 7.28, 7.44, 7.6, 7.76, 7.92, 8.08, 8.24, 8.4, 8.56, 8.72, 8.88, 9.04, 9.2, 9.36, 9.52, 9.68, 9.84, 10.0], "video_metadata": {"fps": 25.0, "total_num_frames": 10164, "frames_indices": [0, 3, 7, 11, 15, 19, 23, 27, 31, 35, 39, 43, 47, 51, 55, 59, 63, 67, 71, 75, 79, 83, 87, 91, 95, 99, 103, 107, 111, 115, 119, 123, 126, 130, 134, 138, 142, 146, 150, 154, 158, 162, 166, 170, 174, 178, 182, 186, 190, 194, 198, 202, 206, 210, 214, 218, 222, 226, 230, 234, 238, 242, 246, 250]}, "sampling": {"method": "uniform_time_floor", "history": "from_start_to_question", "max_n_frames": 64}, "video_preprocessing": {"size": {"shortest_edge": 1048576, "longest_edge": 4194304}, "do_sample_frames": false, "cap_pixels_per_frame": false}, "input_ids_shape": [1, 2250], "video_grid_thw": [[32, 12, 20]], "ground_truth": "A", "model_load_s": 0.00013725098688155413, "gpu": "NVIDIA GeForce RTX 4090", "inference_s": 0.2850912739522755, "peak_allocated_gib": 8.884104251861572, "peak_reserved_gib": 9.1015625, "generation": {"max_new_tokens": 32, "do_sample": false}, "warmup_performed": true, "raw_output": "A", "prediction": "A", "correct": 1, "total_s": 0.9205594349186867}
{"case_path": "/root/Reseach/qwen-video/outputs/review/frames_64/sample_2_a1807_q1/case.json", "model_dir": "/model/ModelScope/Qwen/Qwen3-VL-4B-Instruct", "torch_version": "2.13.0+cu132", "transformers_version": "5.19.0", "status": "ok", "timing_scope": "cached_frames_to_answer", "question_id": "sample_2_a1807_q1", "question_timestamp_s": 19, "num_frames": 64, "input_frame_timestamps_s": [0.0, 0.28, 0.6, 0.88, 1.2, 1.48, 1.8, 2.08, 2.4, 2.68, 3.0, 3.28, 3.6, 3.92, 4.2, 4.52, 4.8, 5.12, 5.4, 5.72, 6.0, 6.32, 6.6, 6.92, 7.2, 7.52, 7.84, 8.12, 8.44, 8.72, 9.04, 9.32, 9.64, 9.92, 10.24, 10.52, 10.84, 11.12, 11.44, 11.76, 12.04, 12.36, 12.64, 12.96, 13.24, 13.56, 13.84, 14.16, 14.44, 14.76, 15.04, 15.36, 15.68, 15.96, 16.28, 16.56, 16.88, 17.16, 17.48, 17.76, 18.08, 18.36, 18.68, 19.0], "video_metadata": {"fps": 25.0, "total_num_frames": 10164, "frames_indices": [0, 7, 15, 22, 30, 37, 45, 52, 60, 67, 75, 82, 90, 98, 105, 113, 120, 128, 135, 143, 150, 158, 165, 173, 180, 188, 196, 203, 211, 218, 226, 233, 241, 248, 256, 263, 271, 278, 286, 294, 301, 309, 316, 324, 331, 339, 346, 354, 361, 369, 376, 384, 392, 399, 407, 414, 422, 429, 437, 444, 452, 459, 467, 475]}, "sampling": {"method": "uniform_time_floor", "history": "from_start_to_question", "max_n_frames": 64}, "video_preprocessing": {"size": {"shortest_edge": 1048576, "longest_edge": 4194304}, "do_sample_frames": false, "cap_pixels_per_frame": false}, "input_ids_shape": [1, 2258], "video_grid_thw": [[32, 12, 20]], "ground_truth": "A", "model_load_s": 0.00010623096022754908, "gpu": "NVIDIA GeForce RTX 4090", "inference_s": 0.28502247005235404, "peak_allocated_gib": 8.885306358337402, "peak_reserved_gib": 9.1015625, "generation": {"max_new_tokens": 32, "do_sample": false}, "warmup_performed": true, "raw_output": "A", "prediction": "A", "correct": 1, "total_s": 0.9259531199932098}
{"case_path": "/root/Reseach/qwen-video/outputs/review/frames_64/sample_2_a1808_q0/case.json", "model_dir": "/model/ModelScope/Qwen/Qwen3-VL-4B-Instruct", "torch_version": "2.13.0+cu132", "transformers_version": "5.19.0", "status": "ok", "timing_scope": "cached_frames_to_answer", "question_id": "sample_2_a1808_q0", "question_timestamp_s": 112, "num_frames": 64, "input_frame_timestamps_s": [0.0, 1.76, 3.52, 5.32, 7.08, 8.88, 10.64, 12.44, 14.2, 16.0, 17.76, 19.52, 21.32, 23.08, 24.88, 26.64, 28.44, 30.2, 32.0, 33.76, 35.52, 37.32, 39.08, 40.88, 42.64, 44.44, 46.2, 48.0, 49.76, 51.52, 53.32, 55.08, 56.88, 58.64, 60.44, 62.2, 64.0, 65.76, 67.52, 69.32, 71.08, 72.88, 74.64, 76.44, 78.2, 80.0, 81.76, 83.52, 85.32, 87.08, 88.88, 90.64, 92.44, 94.2, 96.0, 97.76, 99.52, 101.32, 103.08, 104.88, 106.64, 108.44, 110.2, 112.0], "video_metadata": {"fps": 25.0, "total_num_frames": 10164, "frames_indices": [0, 44, 88, 133, 177, 222, 266, 311, 355, 400, 444, 488, 533, 577, 622, 666, 711, 755, 800, 844, 888, 933, 977, 1022, 1066, 1111, 1155, 1200, 1244, 1288, 1333, 1377, 1422, 1466, 1511, 1555, 1600, 1644, 1688, 1733, 1777, 1822, 1866, 1911, 1955, 2000, 2044, 2088, 2133, 2177, 2222, 2266, 2311, 2355, 2400, 2444, 2488, 2533, 2577, 2622, 2666, 2711, 2755, 2800]}, "sampling": {"method": "uniform_time_floor", "history": "from_start_to_question", "max_n_frames": 64}, "video_preprocessing": {"size": {"shortest_edge": 1048576, "longest_edge": 4194304}, "do_sample_frames": false, "cap_pixels_per_frame": false}, "input_ids_shape": [1, 2286], "video_grid_thw": [[32, 12, 20]], "ground_truth": "B", "model_load_s": 0.00010509102139621973, "gpu": "NVIDIA GeForce RTX 4090", "inference_s": 0.2869199700653553, "peak_allocated_gib": 8.89073133468628, "peak_reserved_gib": 9.1015625, "generation": {"max_new_tokens": 32, "do_sample": false}, "warmup_performed": true, "raw_output": "B", "prediction": "B", "correct": 1, "total_s": 0.9414791560266167}
{"case_path": "/root/Reseach/qwen-video/outputs/review/frames_64/sample_3_a1290_q0/case.json", "model_dir": "/model/ModelScope/Qwen/Qwen3-VL-4B-Instruct", "torch_version": "2.13.0+cu132", "transformers_version": "5.19.0", "status": "ok", "timing_scope": "cached_frames_to_answer", "question_id": "sample_3_a1290_q0", "question_timestamp_s": 8, "num_frames": 64, "input_frame_timestamps_s": [0.0, 0.12, 0.24, 0.36, 0.48, 0.6, 0.76, 0.88, 1.0, 1.12, 1.24, 1.36, 1.52, 1.64, 1.76, 1.88, 2.0, 2.12, 2.28, 2.4, 2.52, 2.64, 2.76, 2.92, 3.04, 3.16, 3.28, 3.4, 3.52, 3.68, 3.8, 3.92, 4.04, 4.16, 4.28, 4.44, 4.56, 4.68, 4.8, 4.92, 5.04, 5.2, 5.32, 5.44, 5.56, 5.68, 5.84, 5.96, 6.08, 6.2, 6.32, 6.44, 6.6, 6.72, 6.84, 6.96, 7.08, 7.2, 7.36, 7.48, 7.6, 7.72, 7.84, 8.0], "video_metadata": {"fps": 25.0, "total_num_frames": 4829, "frames_indices": [0, 3, 6, 9, 12, 15, 19, 22, 25, 28, 31, 34, 38, 41, 44, 47, 50, 53, 57, 60, 63, 66, 69, 73, 76, 79, 82, 85, 88, 92, 95, 98, 101, 104, 107, 111, 114, 117, 120, 123, 126, 130, 133, 136, 139, 142, 146, 149, 152, 155, 158, 161, 165, 168, 171, 174, 177, 180, 184, 187, 190, 193, 196, 200]}, "sampling": {"method": "uniform_time_floor", "history": "from_start_to_question", "max_n_frames": 64}, "video_preprocessing": {"size": {"shortest_edge": 1048576, "longest_edge": 4194304}, "do_sample_frames": false, "cap_pixels_per_frame": false}, "input_ids_shape": [1, 2249], "video_grid_thw": [[32, 12, 20]], "ground_truth": "B", "model_load_s": 9.992101695388556e-05, "gpu": "NVIDIA GeForce RTX 4090", "inference_s": 0.281393118086271, "peak_allocated_gib": 8.883954524993896, "peak_reserved_gib": 9.1015625, "generation": {"max_new_tokens": 32, "do_sample": false}, "warmup_performed": true, "raw_output": "B", "prediction": "B", "correct": 1, "total_s": 0.8874699319712818}
{"case_path": "/root/Reseach/qwen-video/outputs/review/frames_64/sample_3_a1290_q1/case.json", "model_dir": "/model/ModelScope/Qwen/Qwen3-VL-4B-Instruct", "torch_version": "2.13.0+cu132", "transformers_version": "5.19.0", "status": "ok", "timing_scope": "cached_frames_to_answer", "question_id": "sample_3_a1290_q1", "question_timestamp_s": 20, "num_frames": 64, "input_frame_timestamps_s": [0.0, 0.28, 0.6, 0.92, 1.24, 1.56, 1.88, 2.2, 2.52, 2.84, 3.16, 3.48, 3.8, 4.12, 4.44, 4.76, 5.04, 5.36, 5.68, 6.0, 6.32, 6.64, 6.96, 7.28, 7.6, 7.92, 8.24, 8.56, 8.88, 9.2, 9.52, 9.84, 10.12, 10.44, 10.76, 11.08, 11.4, 11.72, 12.04, 12.36, 12.68, 13.0, 13.32, 13.64, 13.96, 14.28, 14.6, 14.92, 15.2, 15.52, 15.84, 16.16, 16.48, 16.8, 17.12, 17.44, 17.76, 18.08, 18.4, 18.72, 19.04, 19.36, 19.68, 20.0], "video_metadata": {"fps": 25.0, "total_num_frames": 4829, "frames_indices": [0, 7, 15, 23, 31, 39, 47, 55, 63, 71, 79, 87, 95, 103, 111, 119, 126, 134, 142, 150, 158, 166, 174, 182, 190, 198, 206, 214, 222, 230, 238, 246, 253, 261, 269, 277, 285, 293, 301, 309, 317, 325, 333, 341, 349, 357, 365, 373, 380, 388, 396, 404, 412, 420, 428, 436, 444, 452, 460, 468, 476, 484, 492, 500]}, "sampling": {"method": "uniform_time_floor", "history": "from_start_to_question", "max_n_frames": 64}, "video_preprocessing": {"size": {"shortest_edge": 1048576, "longest_edge": 4194304}, "do_sample_frames": false, "cap_pixels_per_frame": false}, "input_ids_shape": [1, 2293], "video_grid_thw": [[32, 12, 20]], "ground_truth": "A", "model_load_s": 9.843101724982262e-05, "gpu": "NVIDIA GeForce RTX 4090", "inference_s": 0.28070933499839157, "peak_allocated_gib": 8.89216423034668, "peak_reserved_gib": 9.1015625, "generation": {"max_new_tokens": 32, "do_sample": false}, "warmup_performed": true, "raw_output": "C", "prediction": "C", "correct": 0, "total_s": 0.907506401068531}
{"case_path": "/root/Reseach/qwen-video/outputs/review/frames_64/sample_3_a1291_q0/case.json", "model_dir": "/model/ModelScope/Qwen/Qwen3-VL-4B-Instruct", "torch_version": "2.13.0+cu132", "transformers_version": "5.19.0", "status": "ok", "timing_scope": "cached_frames_to_answer", "question_id": "sample_3_a1291_q0", "question_timestamp_s": 69, "num_frames": 64, "input_frame_timestamps_s": [0.0, 1.08, 2.16, 3.28, 4.36, 5.44, 6.56, 7.64, 8.76, 9.84, 10.92, 12.04, 13.12, 14.2, 15.32, 16.4, 17.52, 18.6, 19.68, 20.8, 21.88, 23.0, 24.08, 25.16, 26.28, 27.36, 28.44, 29.56, 30.64, 31.76, 32.84, 33.92, 35.04, 36.12, 37.2, 38.32, 39.4, 40.52, 41.6, 42.68, 43.8, 44.88, 46.0, 47.08, 48.16, 49.28, 50.36, 51.44, 52.56, 53.64, 54.76, 55.84, 56.92, 58.04, 59.12, 60.2, 61.32, 62.4, 63.52, 64.6, 65.68, 66.8, 67.88, 69.0], "video_metadata": {"fps": 25.0, "total_num_frames": 4829, "frames_indices": [0, 27, 54, 82, 109, 136, 164, 191, 219, 246, 273, 301, 328, 355, 383, 410, 438, 465, 492, 520, 547, 575, 602, 629, 657, 684, 711, 739, 766, 794, 821, 848, 876, 903, 930, 958, 985, 1013, 1040, 1067, 1095, 1122, 1150, 1177, 1204, 1232, 1259, 1286, 1314, 1341, 1369, 1396, 1423, 1451, 1478, 1505, 1533, 1560, 1588, 1615, 1642, 1670, 1697, 1725]}, "sampling": {"method": "uniform_time_floor", "history": "from_start_to_question", "max_n_frames": 64}, "video_preprocessing": {"size": {"shortest_edge": 1048576, "longest_edge": 4194304}, "do_sample_frames": false, "cap_pixels_per_frame": false}, "input_ids_shape": [1, 2272], "video_grid_thw": [[32, 12, 20]], "ground_truth": "C", "model_load_s": 0.00010990200098603964, "gpu": "NVIDIA GeForce RTX 4090", "inference_s": 0.28477869904600084, "peak_allocated_gib": 8.887866973876953, "peak_reserved_gib": 9.1015625, "generation": {"max_new_tokens": 32, "do_sample": false}, "warmup_performed": true, "raw_output": "C", "prediction": "C", "correct": 1, "total_s": 0.9166103199822828}
{"case_path": "/root/Reseach/qwen-video/outputs/review/frames_64/sample_3_a1292_q0/case.json", "model_dir": "/model/ModelScope/Qwen/Qwen3-VL-4B-Instruct", "torch_version": "2.13.0+cu132", "transformers_version": "5.19.0", "status": "ok", "timing_scope": "cached_frames_to_answer", "question_id": "sample_3_a1292_q0", "question_timestamp_s": 123, "num_frames": 64, "input_frame_timestamps_s": [0.0, 1.92, 3.88, 5.84, 7.8, 9.76, 11.68, 13.64, 15.6, 17.56, 19.52, 21.44, 23.4, 25.36, 27.32, 29.28, 31.2, 33.16, 35.12, 37.08, 39.04, 41.0, 42.92, 44.88, 46.84, 48.8, 50.76, 52.68, 54.64, 56.6, 58.56, 60.52, 62.44, 64.4, 66.36, 68.32, 70.28, 72.2, 74.16, 76.12, 78.08, 80.04, 82.0, 83.92, 85.88, 87.84, 89.8, 91.76, 93.68, 95.64, 97.6, 99.56, 101.52, 103.44, 105.4, 107.36, 109.32, 111.28, 113.2, 115.16, 117.12, 119.08, 121.04, 123.0], "video_metadata": {"fps": 25.0, "total_num_frames": 4829, "frames_indices": [0, 48, 97, 146, 195, 244, 292, 341, 390, 439, 488, 536, 585, 634, 683, 732, 780, 829, 878, 927, 976, 1025, 1073, 1122, 1171, 1220, 1269, 1317, 1366, 1415, 1464, 1513, 1561, 1610, 1659, 1708, 1757, 1805, 1854, 1903, 1952, 2001, 2050, 2098, 2147, 2196, 2245, 2294, 2342, 2391, 2440, 2489, 2538, 2586, 2635, 2684, 2733, 2782, 2830, 2879, 2928, 2977, 3026, 3075]}, "sampling": {"method": "uniform_time_floor", "history": "from_start_to_question", "max_n_frames": 64}, "video_preprocessing": {"size": {"shortest_edge": 1048576, "longest_edge": 4194304}, "do_sample_frames": false, "cap_pixels_per_frame": false}, "input_ids_shape": [1, 2288], "video_grid_thw": [[32, 12, 20]], "ground_truth": "C", "model_load_s": 0.00011609098874032497, "gpu": "NVIDIA GeForce RTX 4090", "inference_s": 0.2851396190235391, "peak_allocated_gib": 8.891140460968018, "peak_reserved_gib": 9.1015625, "generation": {"max_new_tokens": 32, "do_sample": false}, "warmup_performed": true, "raw_output": "C", "prediction": "C", "correct": 1, "total_s": 0.926617382094264}
{"case_path": "/root/Reseach/qwen-video/outputs/review/frames_64/sample_3_a1292_q1/case.json", "model_dir": "/model/ModelScope/Qwen/Qwen3-VL-4B-Instruct", "torch_version": "2.13.0+cu132", "transformers_version": "5.19.0", "status": "ok", "timing_scope": "cached_frames_to_answer", "question_id": "sample_3_a1292_q1", "question_timestamp_s": 171, "num_frames": 64, "input_frame_timestamps_s": [0.0, 2.68, 5.4, 8.12, 10.84, 13.56, 16.28, 19.0, 21.68, 24.4, 27.12, 29.84, 32.56, 35.28, 38.0, 40.68, 43.4, 46.12, 48.84, 51.56, 54.28, 57.0, 59.68, 62.4, 65.12, 67.84, 70.56, 73.28, 76.0, 78.68, 81.4, 84.12, 86.84, 89.56, 92.28, 95.0, 97.68, 100.4, 103.12, 105.84, 108.56, 111.28, 114.0, 116.68, 119.4, 122.12, 124.84, 127.56, 130.28, 133.0, 135.68, 138.4, 141.12, 143.84, 146.56, 149.28, 152.0, 154.68, 157.4, 160.12, 162.84, 165.56, 168.28, 171.0], "video_metadata": {"fps": 25.0, "total_num_frames": 4829, "frames_indices": [0, 67, 135, 203, 271, 339, 407, 475, 542, 610, 678, 746, 814, 882, 950, 1017, 1085, 1153, 1221, 1289, 1357, 1425, 1492, 1560, 1628, 1696, 1764, 1832, 1900, 1967, 2035, 2103, 2171, 2239, 2307, 2375, 2442, 2510, 2578, 2646, 2714, 2782, 2850, 2917, 2985, 3053, 3121, 3189, 3257, 3325, 3392, 3460, 3528, 3596, 3664, 3732, 3800, 3867, 3935, 4003, 4071, 4139, 4207, 4275]}, "sampling": {"method": "uniform_time_floor", "history": "from_start_to_question", "max_n_frames": 64}, "video_preprocessing": {"size": {"shortest_edge": 1048576, "longest_edge": 4194304}, "do_sample_frames": false, "cap_pixels_per_frame": false}, "input_ids_shape": [1, 2291], "video_grid_thw": [[32, 12, 20]], "ground_truth": "A", "model_load_s": 0.00010696100071072578, "gpu": "NVIDIA GeForce RTX 4090", "inference_s": 0.2849091140087694, "peak_allocated_gib": 8.891755104064941, "peak_reserved_gib": 9.1015625, "generation": {"max_new_tokens": 32, "do_sample": false}, "warmup_performed": true, "raw_output": "A", "prediction": "A", "correct": 1, "total_s": 0.9077046050224453}
{"case_path": "/root/Reseach/qwen-video/outputs/review/frames_64/sample_4_a610_q0/case.json", "model_dir": "/model/ModelScope/Qwen/Qwen3-VL-4B-Instruct", "torch_version": "2.13.0+cu132", "transformers_version": "5.19.0", "status": "ok", "timing_scope": "cached_frames_to_answer", "question_id": "sample_4_a610_q0", "question_timestamp_s": 20, "num_frames": 64, "input_frame_timestamps_s": [0.0, 0.28, 0.6, 0.92, 1.24, 1.56, 1.88, 2.2, 2.52, 2.84, 3.16, 3.48, 3.8, 4.12, 4.44, 4.76, 5.04, 5.36, 5.68, 6.0, 6.32, 6.64, 6.96, 7.28, 7.6, 7.92, 8.24, 8.56, 8.88, 9.2, 9.52, 9.84, 10.12, 10.44, 10.76, 11.08, 11.4, 11.72, 12.04, 12.36, 12.68, 13.0, 13.32, 13.64, 13.96, 14.28, 14.6, 14.92, 15.2, 15.52, 15.84, 16.16, 16.48, 16.8, 17.12, 17.44, 17.76, 18.08, 18.4, 18.72, 19.04, 19.36, 19.68, 20.0], "video_metadata": {"fps": 25.0, "total_num_frames": 5775, "frames_indices": [0, 7, 15, 23, 31, 39, 47, 55, 63, 71, 79, 87, 95, 103, 111, 119, 126, 134, 142, 150, 158, 166, 174, 182, 190, 198, 206, 214, 222, 230, 238, 246, 253, 261, 269, 277, 285, 293, 301, 309, 317, 325, 333, 341, 349, 357, 365, 373, 380, 388, 396, 404, 412, 420, 428, 436, 444, 452, 460, 468, 476, 484, 492, 500]}, "sampling": {"method": "uniform_time_floor", "history": "from_start_to_question", "max_n_frames": 64}, "video_preprocessing": {"size": {"shortest_edge": 1048576, "longest_edge": 4194304}, "do_sample_frames": false, "cap_pixels_per_frame": false}, "input_ids_shape": [1, 2264], "video_grid_thw": [[32, 12, 20]], "ground_truth": "B", "model_load_s": 0.00010248098988085985, "gpu": "NVIDIA GeForce RTX 4090", "inference_s": 0.28220175695605576, "peak_allocated_gib": 8.88623046875, "peak_reserved_gib": 9.1015625, "generation": {"max_new_tokens": 32, "do_sample": false}, "warmup_performed": true, "raw_output": "B", "prediction": "B", "correct": 1, "total_s": 0.9137490279972553}
{"case_path": "/root/Reseach/qwen-video/outputs/review/frames_64/sample_4_a610_q1/case.json", "model_dir": "/model/ModelScope/Qwen/Qwen3-VL-4B-Instruct", "torch_version": "2.13.0+cu132", "transformers_version": "5.19.0", "status": "ok", "timing_scope": "cached_frames_to_answer", "question_id": "sample_4_a610_q1", "question_timestamp_s": 19, "num_frames": 64, "input_frame_timestamps_s": [0.0, 0.28, 0.6, 0.88, 1.2, 1.48, 1.8, 2.08, 2.4, 2.68, 3.0, 3.28, 3.6, 3.92, 4.2, 4.52, 4.8, 5.12, 5.4, 5.72, 6.0, 6.32, 6.6, 6.92, 7.2, 7.52, 7.84, 8.12, 8.44, 8.72, 9.04, 9.32, 9.64, 9.92, 10.24, 10.52, 10.84, 11.12, 11.44, 11.76, 12.04, 12.36, 12.64, 12.96, 13.24, 13.56, 13.84, 14.16, 14.44, 14.76, 15.04, 15.36, 15.68, 15.96, 16.28, 16.56, 16.88, 17.16, 17.48, 17.76, 18.08, 18.36, 18.68, 19.0], "video_metadata": {"fps": 25.0, "total_num_frames": 5775, "frames_indices": [0, 7, 15, 22, 30, 37, 45, 52, 60, 67, 75, 82, 90, 98, 105, 113, 120, 128, 135, 143, 150, 158, 165, 173, 180, 188, 196, 203, 211, 218, 226, 233, 241, 248, 256, 263, 271, 278, 286, 294, 301, 309, 316, 324, 331, 339, 346, 354, 361, 369, 376, 384, 392, 399, 407, 414, 422, 429, 437, 444, 452, 459, 467, 475]}, "sampling": {"method": "uniform_time_floor", "history": "from_start_to_question", "max_n_frames": 64}, "video_preprocessing": {"size": {"shortest_edge": 1048576, "longest_edge": 4194304}, "do_sample_frames": false, "cap_pixels_per_frame": false}, "input_ids_shape": [1, 2284], "video_grid_thw": [[32, 12, 20]], "ground_truth": "C", "model_load_s": 0.00011380109935998917, "gpu": "NVIDIA GeForce RTX 4090", "inference_s": 0.2836042509879917, "peak_allocated_gib": 8.890322208404541, "peak_reserved_gib": 9.1015625, "generation": {"max_new_tokens": 32, "do_sample": false}, "warmup_performed": true, "raw_output": "C", "prediction": "C", "correct": 1, "total_s": 0.9067716029239818}
{"case_path": "/root/Reseach/qwen-video/outputs/review/frames_64/sample_4_a611_q0/case.json", "model_dir": "/model/ModelScope/Qwen/Qwen3-VL-4B-Instruct", "torch_version": "2.13.0+cu132", "transformers_version": "5.19.0", "status": "ok", "timing_scope": "cached_frames_to_answer", "question_id": "sample_4_a611_q0", "question_timestamp_s": 64, "num_frames": 64, "input_frame_timestamps_s": [0.0, 1.0, 2.0, 3.04, 4.04, 5.04, 6.08, 7.08, 8.12, 9.12, 10.12, 11.16, 12.16, 13.2, 14.2, 15.2, 16.24, 17.24, 18.28, 19.28, 20.28, 21.32, 22.32, 23.36, 24.36, 25.36, 26.4, 27.4, 28.44, 29.44, 30.44, 31.48, 32.48, 33.52, 34.52, 35.52, 36.56, 37.56, 38.6, 39.6, 40.6, 41.64, 42.64, 43.68, 44.68, 45.68, 46.72, 47.72, 48.76, 49.76, 50.76, 51.8, 52.8, 53.84, 54.84, 55.84, 56.88, 57.88, 58.92, 59.92, 60.92, 61.96, 62.96, 64.0], "video_metadata": {"fps": 25.0, "total_num_frames": 5775, "frames_indices": [0, 25, 50, 76, 101, 126, 152, 177, 203, 228, 253, 279, 304, 330, 355, 380, 406, 431, 457, 482, 507, 533, 558, 584, 609, 634, 660, 685, 711, 736, 761, 787, 812, 838, 863, 888, 914, 939, 965, 990, 1015, 1041, 1066, 1092, 1117, 1142, 1168, 1193, 1219, 1244, 1269, 1295, 1320, 1346, 1371, 1396, 1422, 1447, 1473, 1498, 1523, 1549, 1574, 1600]}, "sampling": {"method": "uniform_time_floor", "history": "from_start_to_question", "max_n_frames": 64}, "video_preprocessing": {"size": {"shortest_edge": 1048576, "longest_edge": 4194304}, "do_sample_frames": false, "cap_pixels_per_frame": false}, "input_ids_shape": [1, 2311], "video_grid_thw": [[32, 12, 20]], "ground_truth": "C", "model_load_s": 0.00012034201063215733, "gpu": "NVIDIA GeForce RTX 4090", "inference_s": 0.2838775709969923, "peak_allocated_gib": 8.895808696746826, "peak_reserved_gib": 9.1015625, "generation": {"max_new_tokens": 32, "do_sample": false}, "warmup_performed": true, "raw_output": "A", "prediction": "A", "correct": 0, "total_s": 0.9044579539913684}
{"case_path": "/root/Reseach/qwen-video/outputs/review/frames_64/sample_4_a613_q0/case.json", "model_dir": "/model/ModelScope/Qwen/Qwen3-VL-4B-Instruct", "torch_version": "2.13.0+cu132", "transformers_version": "5.19.0", "status": "ok", "timing_scope": "cached_frames_to_answer", "question_id": "sample_4_a613_q0", "question_timestamp_s": 220, "num_frames": 64, "input_frame_timestamps_s": [0.0, 3.48, 6.96, 10.44, 13.96, 17.44, 20.92, 24.44, 27.92, 31.4, 34.92, 38.4, 41.88, 45.36, 48.88, 52.36, 55.84, 59.36, 62.84, 66.32, 69.84, 73.32, 76.8, 80.28, 83.8, 87.28, 90.76, 94.28, 97.76, 101.24, 104.76, 108.24, 111.72, 115.2, 118.72, 122.2, 125.68, 129.2, 132.68, 136.16, 139.68, 143.16, 146.64, 150.12, 153.64, 157.12, 160.6, 164.12, 167.6, 171.08, 174.6, 178.08, 181.56, 185.04, 188.56, 192.04, 195.52, 199.04, 202.52, 206.0, 209.52, 213.0, 216.48, 220.0], "video_metadata": {"fps": 25.0, "total_num_frames": 5775, "frames_indices": [0, 87, 174, 261, 349, 436, 523, 611, 698, 785, 873, 960, 1047, 1134, 1222, 1309, 1396, 1484, 1571, 1658, 1746, 1833, 1920, 2007, 2095, 2182, 2269, 2357, 2444, 2531, 2619, 2706, 2793, 2880, 2968, 3055, 3142, 3230, 3317, 3404, 3492, 3579, 3666, 3753, 3841, 3928, 4015, 4103, 4190, 4277, 4365, 4452, 4539, 4626, 4714, 4801, 4888, 4976, 5063, 5150, 5238, 5325, 5412, 5500]}, "sampling": {"method": "uniform_time_floor", "history": "from_start_to_question", "max_n_frames": 64}, "video_preprocessing": {"size": {"shortest_edge": 1048576, "longest_edge": 4194304}, "do_sample_frames": false, "cap_pixels_per_frame": false}, "input_ids_shape": [1, 2300], "video_grid_thw": [[32, 12, 20]], "ground_truth": "B", "model_load_s": 0.00011098210234194994, "gpu": "NVIDIA GeForce RTX 4090", "inference_s": 0.284608667017892, "peak_allocated_gib": 8.893595695495605, "peak_reserved_gib": 9.1015625, "generation": {"max_new_tokens": 32, "do_sample": false}, "warmup_performed": true, "raw_output": "B", "prediction": "B", "correct": 1, "total_s": 0.9345362680032849}
{"case_path": "/root/Reseach/qwen-video/outputs/review/frames_64/sample_5_a880_q0/case.json", "model_dir": "/model/ModelScope/Qwen/Qwen3-VL-4B-Instruct", "torch_version": "2.13.0+cu132", "transformers_version": "5.19.0", "status": "ok", "timing_scope": "cached_frames_to_answer", "question_id": "sample_5_a880_q0", "question_timestamp_s": 14, "num_frames": 64, "input_frame_timestamps_s": [0.0, 0.22, 0.44, 0.66, 0.88, 1.1, 1.32, 1.54, 1.76, 2.0, 2.22, 2.44, 2.66, 2.88, 3.1, 3.32, 3.54, 3.76, 4.0, 4.22, 4.44, 4.66, 4.88, 5.1, 5.32, 5.54, 5.76, 6.0, 6.22, 6.44, 6.66, 6.88, 7.1, 7.32, 7.54, 7.76, 8.0, 8.22, 8.44, 8.66, 8.88, 9.1, 9.32, 9.54, 9.76, 10.0, 10.22, 10.44, 10.66, 10.88, 11.1, 11.32, 11.54, 11.76, 12.0, 12.22, 12.44, 12.66, 12.88, 13.1, 13.32, 13.54, 13.76, 14.0], "video_metadata": {"fps": 50.0, "total_num_frames": 25607, "frames_indices": [0, 11, 22, 33, 44, 55, 66, 77, 88, 100, 111, 122, 133, 144, 155, 166, 177, 188, 200, 211, 222, 233, 244, 255, 266, 277, 288, 300, 311, 322, 333, 344, 355, 366, 377, 388, 400, 411, 422, 433, 444, 455, 466, 477, 488, 500, 511, 522, 533, 544, 555, 566, 577, 588, 600, 611, 622, 633, 644, 655, 666, 677, 688, 700]}, "sampling": {"method": "uniform_time_floor", "history": "from_start_to_question", "max_n_frames": 64}, "video_preprocessing": {"size": {"shortest_edge": 1048576, "longest_edge": 4194304}, "do_sample_frames": false, "cap_pixels_per_frame": false}, "input_ids_shape": [1, 2258], "video_grid_thw": [[32, 12, 20]], "ground_truth": "B", "model_load_s": 0.00012671202421188354, "gpu": "NVIDIA GeForce RTX 4090", "inference_s": 0.28550935408566147, "peak_allocated_gib": 8.885306358337402, "peak_reserved_gib": 9.1015625, "generation": {"max_new_tokens": 32, "do_sample": false}, "warmup_performed": true, "raw_output": "B", "prediction": "B", "correct": 1, "total_s": 0.9191208010306582}
{"case_path": "/root/Reseach/qwen-video/outputs/review/frames_64/sample_5_a883_q0/case.json", "model_dir": "/model/ModelScope/Qwen/Qwen3-VL-4B-Instruct", "torch_version": "2.13.0+cu132", "transformers_version": "5.19.0", "status": "ok", "timing_scope": "cached_frames_to_answer", "question_id": "sample_5_a883_q0", "question_timestamp_s": 443, "num_frames": 64, "input_frame_timestamps_s": [0.0, 7.02, 14.06, 21.08, 28.12, 35.14, 42.18, 49.22, 56.24, 63.28, 70.3, 77.34, 84.38, 91.4, 98.44, 105.46, 112.5, 119.52, 126.56, 133.6, 140.62, 147.66, 154.68, 161.72, 168.76, 175.78, 182.82, 189.84, 196.88, 203.92, 210.94, 217.98, 225.0, 232.04, 239.06, 246.1, 253.14, 260.16, 267.2, 274.22, 281.26, 288.3, 295.32, 302.36, 309.38, 316.42, 323.46, 330.48, 337.52, 344.54, 351.58, 358.6, 365.64, 372.68, 379.7, 386.74, 393.76, 400.8, 407.84, 414.86, 421.9, 428.92, 435.96, 443.0], "video_metadata": {"fps": 50.0, "total_num_frames": 25607, "frames_indices": [0, 351, 703, 1054, 1406, 1757, 2109, 2461, 2812, 3164, 3515, 3867, 4219, 4570, 4922, 5273, 5625, 5976, 6328, 6680, 7031, 7383, 7734, 8086, 8438, 8789, 9141, 9492, 9844, 10196, 10547, 10899, 11250, 11602, 11953, 12305, 12657, 13008, 13360, 13711, 14063, 14415, 14766, 15118, 15469, 15821, 16173, 16524, 16876, 17227, 17579, 17930, 18282, 18634, 18985, 19337, 19688, 20040, 20392, 20743, 21095, 21446, 21798, 22150]}, "sampling": {"method": "uniform_time_floor", "history": "from_start_to_question", "max_n_frames": 64}, "video_preprocessing": {"size": {"shortest_edge": 1048576, "longest_edge": 4194304}, "do_sample_frames": false, "cap_pixels_per_frame": false}, "input_ids_shape": [1, 2311], "video_grid_thw": [[32, 12, 20]], "ground_truth": "D", "model_load_s": 0.00011103099677711725, "gpu": "NVIDIA GeForce RTX 4090", "inference_s": 0.2875939660007134, "peak_allocated_gib": 8.895808696746826, "peak_reserved_gib": 9.1015625, "generation": {"max_new_tokens": 32, "do_sample": false}, "warmup_performed": true, "raw_output": "D", "prediction": "D", "correct": 1, "total_s": 0.9784921529935673}
{"case_path": "/root/Reseach/qwen-video/outputs/review/frames_64/sample_6_a1421_q0/case.json", "model_dir": "/model/ModelScope/Qwen/Qwen3-VL-4B-Instruct", "torch_version": "2.13.0+cu132", "transformers_version": "5.19.0", "status": "ok", "timing_scope": "cached_frames_to_answer", "question_id": "sample_6_a1421_q0", "question_timestamp_s": 14, "num_frames": 64, "input_frame_timestamps_s": [0.0, 0.2, 0.44, 0.64, 0.88, 1.08, 1.32, 1.52, 1.76, 2.0, 2.2, 2.44, 2.64, 2.88, 3.08, 3.32, 3.52, 3.76, 4.0, 4.2, 4.44, 4.64, 4.88, 5.08, 5.32, 5.52, 5.76, 6.0, 6.2, 6.44, 6.64, 6.88, 7.08, 7.32, 7.52, 7.76, 8.0, 8.2, 8.44, 8.64, 8.88, 9.08, 9.32, 9.52, 9.76, 10.0, 10.2, 10.44, 10.64, 10.88, 11.08, 11.32, 11.52, 11.76, 12.0, 12.2, 12.44, 12.64, 12.88, 13.08, 13.32, 13.52, 13.76, 14.0], "video_metadata": {"fps": 25.0, "total_num_frames": 14415, "frames_indices": [0, 5, 11, 16, 22, 27, 33, 38, 44, 50, 55, 61, 66, 72, 77, 83, 88, 94, 100, 105, 111, 116, 122, 127, 133, 138, 144, 150, 155, 161, 166, 172, 177, 183, 188, 194, 200, 205, 211, 216, 222, 227, 233, 238, 244, 250, 255, 261, 266, 272, 277, 283, 288, 294, 300, 305, 311, 316, 322, 327, 333, 338, 344, 350]}, "sampling": {"method": "uniform_time_floor", "history": "from_start_to_question", "max_n_frames": 64}, "video_preprocessing": {"size": {"shortest_edge": 1048576, "longest_edge": 4194304}, "do_sample_frames": false, "cap_pixels_per_frame": false}, "input_ids_shape": [1, 2271], "video_grid_thw": [[32, 12, 20]], "ground_truth": "D", "model_load_s": 0.00010372092947363853, "gpu": "NVIDIA GeForce RTX 4090", "inference_s": 0.2811800540657714, "peak_allocated_gib": 8.887662887573242, "peak_reserved_gib": 9.1015625, "generation": {"max_new_tokens": 32, "do_sample": false}, "warmup_performed": true, "raw_output": "D", "prediction": "D", "correct": 1, "total_s": 0.957515359041281}
{"case_path": "/root/Reseach/qwen-video/outputs/review/frames_64/sample_6_a1422_q0/case.json", "model_dir": "/model/ModelScope/Qwen/Qwen3-VL-4B-Instruct", "torch_version": "2.13.0+cu132", "transformers_version": "5.19.0", "status": "ok", "timing_scope": "cached_frames_to_answer", "question_id": "sample_6_a1422_q0", "question_timestamp_s": 145, "num_frames": 64, "input_frame_timestamps_s": [0.0, 2.28, 4.6, 6.88, 9.2, 11.48, 13.8, 16.08, 18.4, 20.68, 23.0, 25.28, 27.6, 29.92, 32.2, 34.52, 36.8, 39.12, 41.4, 43.72, 46.0, 48.32, 50.6, 52.92, 55.2, 57.52, 59.84, 62.12, 64.44, 66.72, 69.04, 71.32, 73.64, 75.92, 78.24, 80.52, 82.84, 85.12, 87.44, 89.76, 92.04, 94.36, 96.64, 98.96, 101.24, 103.56, 105.84, 108.16, 110.44, 112.76, 115.04, 117.36, 119.68, 121.96, 124.28, 126.56, 128.88, 131.16, 133.48, 135.76, 138.08, 140.36, 142.68, 145.0], "video_metadata": {"fps": 25.0, "total_num_frames": 14415, "frames_indices": [0, 57, 115, 172, 230, 287, 345, 402, 460, 517, 575, 632, 690, 748, 805, 863, 920, 978, 1035, 1093, 1150, 1208, 1265, 1323, 1380, 1438, 1496, 1553, 1611, 1668, 1726, 1783, 1841, 1898, 1956, 2013, 2071, 2128, 2186, 2244, 2301, 2359, 2416, 2474, 2531, 2589, 2646, 2704, 2761, 2819, 2876, 2934, 2992, 3049, 3107, 3164, 3222, 3279, 3337, 3394, 3452, 3509, 3567, 3625]}, "sampling": {"method": "uniform_time_floor", "history": "from_start_to_question", "max_n_frames": 64}, "video_preprocessing": {"size": {"shortest_edge": 1048576, "longest_edge": 4194304}, "do_sample_frames": false, "cap_pixels_per_frame": false}, "input_ids_shape": [1, 2281], "video_grid_thw": [[32, 12, 20]], "ground_truth": "D", "model_load_s": 0.00010325200855731964, "gpu": "NVIDIA GeForce RTX 4090", "inference_s": 0.2853179411031306, "peak_allocated_gib": 8.889708995819092, "peak_reserved_gib": 9.1015625, "generation": {"max_new_tokens": 32, "do_sample": false}, "warmup_performed": true, "raw_output": "D", "prediction": "D", "correct": 1, "total_s": 0.9987080139108002}
{"case_path": "/root/Reseach/qwen-video/outputs/review/frames_64/sample_6_a1422_q1/case.json", "model_dir": "/model/ModelScope/Qwen/Qwen3-VL-4B-Instruct", "torch_version": "2.13.0+cu132", "transformers_version": "5.19.0", "status": "ok", "timing_scope": "cached_frames_to_answer", "question_id": "sample_6_a1422_q1", "question_timestamp_s": 153, "num_frames": 64, "input_frame_timestamps_s": [0.0, 2.4, 4.84, 7.28, 9.68, 12.12, 14.56, 17.0, 19.4, 21.84, 24.28, 26.68, 29.12, 31.56, 34.0, 36.4, 38.84, 41.28, 43.68, 46.12, 48.56, 51.0, 53.4, 55.84, 58.28, 60.68, 63.12, 65.56, 68.0, 70.4, 72.84, 75.28, 77.68, 80.12, 82.56, 85.0, 87.4, 89.84, 92.28, 94.68, 97.12, 99.56, 102.0, 104.4, 106.84, 109.28, 111.68, 114.12, 116.56, 119.0, 121.4, 123.84, 126.28, 128.68, 131.12, 133.56, 136.0, 138.4, 140.84, 143.28, 145.68, 148.12, 150.56, 153.0], "video_metadata": {"fps": 25.0, "total_num_frames": 14415, "frames_indices": [0, 60, 121, 182, 242, 303, 364, 425, 485, 546, 607, 667, 728, 789, 850, 910, 971, 1032, 1092, 1153, 1214, 1275, 1335, 1396, 1457, 1517, 1578, 1639, 1700, 1760, 1821, 1882, 1942, 2003, 2064, 2125, 2185, 2246, 2307, 2367, 2428, 2489, 2550, 2610, 2671, 2732, 2792, 2853, 2914, 2975, 3035, 3096, 3157, 3217, 3278, 3339, 3400, 3460, 3521, 3582, 3642, 3703, 3764, 3825]}, "sampling": {"method": "uniform_time_floor", "history": "from_start_to_question", "max_n_frames": 64}, "video_preprocessing": {"size": {"shortest_edge": 1048576, "longest_edge": 4194304}, "do_sample_frames": false, "cap_pixels_per_frame": false}, "input_ids_shape": [1, 2299], "video_grid_thw": [[32, 12, 20]], "ground_truth": "D", "model_load_s": 0.00011879100929945707, "gpu": "NVIDIA GeForce RTX 4090", "inference_s": 0.2858261839719489, "peak_allocated_gib": 8.893391609191895, "peak_reserved_gib": 9.1015625, "generation": {"max_new_tokens": 32, "do_sample": false}, "warmup_performed": true, "raw_output": "D", "prediction": "D", "correct": 1, "total_s": 1.009115494089201}
{"case_path": "/root/Reseach/qwen-video/outputs/review/frames_64/sample_6_a1423_q0/case.json", "model_dir": "/model/ModelScope/Qwen/Qwen3-VL-4B-Instruct", "torch_version": "2.13.0+cu132", "transformers_version": "5.19.0", "status": "ok", "timing_scope": "cached_frames_to_answer", "question_id": "sample_6_a1423_q0", "question_timestamp_s": 293, "num_frames": 64, "input_frame_timestamps_s": [0.0, 4.64, 9.28, 13.92, 18.6, 23.24, 27.88, 32.52, 37.2, 41.84, 46.48, 51.12, 55.8, 60.44, 65.08, 69.76, 74.4, 79.04, 83.68, 88.36, 93.0, 97.64, 102.28, 106.96, 111.6, 116.24, 120.92, 125.56, 130.2, 134.84, 139.52, 144.16, 148.8, 153.44, 158.12, 162.76, 167.4, 172.04, 176.72, 181.36, 186.0, 190.68, 195.32, 199.96, 204.6, 209.28, 213.92, 218.56, 223.2, 227.88, 232.52, 237.16, 241.84, 246.48, 251.12, 255.76, 260.44, 265.08, 269.72, 274.36, 279.04, 283.68, 288.32, 293.0], "video_metadata": {"fps": 25.0, "total_num_frames": 14415, "frames_indices": [0, 116, 232, 348, 465, 581, 697, 813, 930, 1046, 1162, 1278, 1395, 1511, 1627, 1744, 1860, 1976, 2092, 2209, 2325, 2441, 2557, 2674, 2790, 2906, 3023, 3139, 3255, 3371, 3488, 3604, 3720, 3836, 3953, 4069, 4185, 4301, 4418, 4534, 4650, 4767, 4883, 4999, 5115, 5232, 5348, 5464, 5580, 5697, 5813, 5929, 6046, 6162, 6278, 6394, 6511, 6627, 6743, 6859, 6976, 7092, 7208, 7325]}, "sampling": {"method": "uniform_time_floor", "history": "from_start_to_question", "max_n_frames": 64}, "video_preprocessing": {"size": {"shortest_edge": 1048576, "longest_edge": 4194304}, "do_sample_frames": false, "cap_pixels_per_frame": false}, "input_ids_shape": [1, 2309], "video_grid_thw": [[32, 12, 20]], "ground_truth": "D", "model_load_s": 9.593099821358919e-05, "gpu": "NVIDIA GeForce RTX 4090", "inference_s": 0.2853524449747056, "peak_allocated_gib": 8.895411014556885, "peak_reserved_gib": 9.1015625, "generation": {"max_new_tokens": 32, "do_sample": false}, "warmup_performed": true, "raw_output": "A", "prediction": "A", "correct": 0, "total_s": 1.0041216759709641}
{"case_path": "/root/Reseach/qwen-video/outputs/review/frames_64/sample_6_a1424_q0/case.json", "model_dir": "/model/ModelScope/Qwen/Qwen3-VL-4B-Instruct", "torch_version": "2.13.0+cu132", "transformers_version": "5.19.0", "status": "ok", "timing_scope": "cached_frames_to_answer", "question_id": "sample_6_a1424_q0", "question_timestamp_s": 425, "num_frames": 64, "input_frame_timestamps_s": [0.0, 6.72, 13.48, 20.2, 26.96, 33.72, 40.44, 47.2, 53.96, 60.68, 67.44, 74.2, 80.92, 87.68, 94.44, 101.16, 107.92, 114.68, 121.4, 128.16, 134.92, 141.64, 148.4, 155.12, 161.88, 168.64, 175.36, 182.12, 188.88, 195.6, 202.36, 209.12, 215.84, 222.6, 229.36, 236.08, 242.84, 249.6, 256.32, 263.08, 269.84, 276.56, 283.32, 290.04, 296.8, 303.56, 310.28, 317.04, 323.8, 330.52, 337.28, 344.04, 350.76, 357.52, 364.28, 371.0, 377.76, 384.52, 391.24, 398.0, 404.76, 411.48, 418.24, 425.0], "video_metadata": {"fps": 25.0, "total_num_frames": 14415, "frames_indices": [0, 168, 337, 505, 674, 843, 1011, 1180, 1349, 1517, 1686, 1855, 2023, 2192, 2361, 2529, 2698, 2867, 3035, 3204, 3373, 3541, 3710, 3878, 4047, 4216, 4384, 4553, 4722, 4890, 5059, 5228, 5396, 5565, 5734, 5902, 6071, 6240, 6408, 6577, 6746, 6914, 7083, 7251, 7420, 7589, 7757, 7926, 8095, 8263, 8432, 8601, 8769, 8938, 9107, 9275, 9444, 9613, 9781, 9950, 10119, 10287, 10456, 10625]}, "sampling": {"method": "uniform_time_floor", "history": "from_start_to_question", "max_n_frames": 64}, "video_preprocessing": {"size": {"shortest_edge": 1048576, "longest_edge": 4194304}, "do_sample_frames": false, "cap_pixels_per_frame": false}, "input_ids_shape": [1, 2312], "video_grid_thw": [[32, 12, 20]], "ground_truth": "C", "model_load_s": 0.00010504201054573059, "gpu": "NVIDIA GeForce RTX 4090", "inference_s": 0.2915490090381354, "peak_allocated_gib": 8.896007061004639, "peak_reserved_gib": 9.1015625, "generation": {"max_new_tokens": 32, "do_sample": false}, "warmup_performed": true, "raw_output": "C", "prediction": "C", "correct": 1, "total_s": 0.9968228109646589}
{"case_path": "/root/Reseach/qwen-video/outputs/review/frames_64/sample_7_a2136_q0/case.json", "model_dir": "/model/ModelScope/Qwen/Qwen3-VL-4B-Instruct", "torch_version": "2.13.0+cu132", "transformers_version": "5.19.0", "status": "ok", "timing_scope": "cached_frames_to_answer", "question_id": "sample_7_a2136_q0", "question_timestamp_s": 16, "num_frames": 64, "input_frame_timestamps_s": [0.0, 0.24, 0.48, 0.76, 1.0, 1.24, 1.52, 1.76, 2.0, 2.28, 2.52, 2.76, 3.04, 3.28, 3.52, 3.8, 4.04, 4.28, 4.56, 4.8, 5.04, 5.32, 5.56, 5.84, 6.08, 6.32, 6.6, 6.84, 7.08, 7.36, 7.6, 7.84, 8.12, 8.36, 8.6, 8.88, 9.12, 9.36, 9.64, 9.88, 10.12, 10.4, 10.64, 10.92, 11.16, 11.4, 11.68, 11.92, 12.16, 12.44, 12.68, 12.92, 13.2, 13.44, 13.68, 13.96, 14.2, 14.44, 14.72, 14.96, 15.2, 15.48, 15.72, 16.0], "video_metadata": {"fps": 25.0, "total_num_frames": 11475, "frames_indices": [0, 6, 12, 19, 25, 31, 38, 44, 50, 57, 63, 69, 76, 82, 88, 95, 101, 107, 114, 120, 126, 133, 139, 146, 152, 158, 165, 171, 177, 184, 190, 196, 203, 209, 215, 222, 228, 234, 241, 247, 253, 260, 266, 273, 279, 285, 292, 298, 304, 311, 317, 323, 330, 336, 342, 349, 355, 361, 368, 374, 380, 387, 393, 400]}, "sampling": {"method": "uniform_time_floor", "history": "from_start_to_question", "max_n_frames": 64}, "video_preprocessing": {"size": {"shortest_edge": 1048576, "longest_edge": 4194304}, "do_sample_frames": false, "cap_pixels_per_frame": false}, "input_ids_shape": [1, 2275], "video_grid_thw": [[32, 12, 20]], "ground_truth": "A", "model_load_s": 9.327102452516556e-05, "gpu": "NVIDIA GeForce RTX 4090", "inference_s": 0.28419327398296446, "peak_allocated_gib": 8.888481616973877, "peak_reserved_gib": 9.1015625, "generation": {"max_new_tokens": 32, "do_sample": false}, "warmup_performed": true, "raw_output": "A", "prediction": "A", "correct": 1, "total_s": 0.9719389469828457}
{"case_path": "/root/Reseach/qwen-video/outputs/review/frames_64/sample_7_a2138_q0/case.json", "model_dir": "/model/ModelScope/Qwen/Qwen3-VL-4B-Instruct", "torch_version": "2.13.0+cu132", "transformers_version": "5.19.0", "status": "ok", "timing_scope": "cached_frames_to_answer", "question_id": "sample_7_a2138_q0", "question_timestamp_s": 260, "num_frames": 64, "input_frame_timestamps_s": [0.0, 4.12, 8.24, 12.36, 16.48, 20.6, 24.76, 28.88, 33.0, 37.12, 41.24, 45.36, 49.52, 53.64, 57.76, 61.88, 66.0, 70.12, 74.28, 78.4, 82.52, 86.64, 90.76, 94.92, 99.04, 103.16, 107.28, 111.4, 115.52, 119.68, 123.8, 127.92, 132.04, 136.16, 140.28, 144.44, 148.56, 152.68, 156.8, 160.92, 165.04, 169.2, 173.32, 177.44, 181.56, 185.68, 189.84, 193.96, 198.08, 202.2, 206.32, 210.44, 214.6, 218.72, 222.84, 226.96, 231.08, 235.2, 239.36, 243.48, 247.6, 251.72, 255.84, 260.0], "video_metadata": {"fps": 25.0, "total_num_frames": 11475, "frames_indices": [0, 103, 206, 309, 412, 515, 619, 722, 825, 928, 1031, 1134, 1238, 1341, 1444, 1547, 1650, 1753, 1857, 1960, 2063, 2166, 2269, 2373, 2476, 2579, 2682, 2785, 2888, 2992, 3095, 3198, 3301, 3404, 3507, 3611, 3714, 3817, 3920, 4023, 4126, 4230, 4333, 4436, 4539, 4642, 4746, 4849, 4952, 5055, 5158, 5261, 5365, 5468, 5571, 5674, 5777, 5880, 5984, 6087, 6190, 6293, 6396, 6500]}, "sampling": {"method": "uniform_time_floor", "history": "from_start_to_question", "max_n_frames": 64}, "video_preprocessing": {"size": {"shortest_edge": 1048576, "longest_edge": 4194304}, "do_sample_frames": false, "cap_pixels_per_frame": false}, "input_ids_shape": [1, 2308], "video_grid_thw": [[32, 12, 20]], "ground_truth": "D", "model_load_s": 0.00010331196244806051, "gpu": "NVIDIA GeForce RTX 4090", "inference_s": 0.2840067590586841, "peak_allocated_gib": 8.895211696624756, "peak_reserved_gib": 9.1015625, "generation": {"max_new_tokens": 32, "do_sample": false}, "warmup_performed": true, "raw_output": "D", "prediction": "D", "correct": 1, "total_s": 0.9817599399248138}
{"case_path": "/root/Reseach/qwen-video/outputs/review/frames_64/sample_7_a2139_q0/case.json", "model_dir": "/model/ModelScope/Qwen/Qwen3-VL-4B-Instruct", "torch_version": "2.13.0+cu132", "transformers_version": "5.19.0", "status": "ok", "timing_scope": "cached_frames_to_answer", "question_id": "sample_7_a2139_q0", "question_timestamp_s": 396, "num_frames": 64, "input_frame_timestamps_s": [0.0, 6.28, 12.56, 18.84, 25.12, 31.4, 37.68, 44.0, 50.28, 56.56, 62.84, 69.12, 75.4, 81.68, 88.0, 94.28, 100.56, 106.84, 113.12, 119.4, 125.68, 132.0, 138.28, 144.56, 150.84, 157.12, 163.4, 169.68, 176.0, 182.28, 188.56, 194.84, 201.12, 207.4, 213.68, 220.0, 226.28, 232.56, 238.84, 245.12, 251.4, 257.68, 264.0, 270.28, 276.56, 282.84, 289.12, 295.4, 301.68, 308.0, 314.28, 320.56, 326.84, 333.12, 339.4, 345.68, 352.0, 358.28, 364.56, 370.84, 377.12, 383.4, 389.68, 396.0], "video_metadata": {"fps": 25.0, "total_num_frames": 11475, "frames_indices": [0, 157, 314, 471, 628, 785, 942, 1100, 1257, 1414, 1571, 1728, 1885, 2042, 2200, 2357, 2514, 2671, 2828, 2985, 3142, 3300, 3457, 3614, 3771, 3928, 4085, 4242, 4400, 4557, 4714, 4871, 5028, 5185, 5342, 5500, 5657, 5814, 5971, 6128, 6285, 6442, 6600, 6757, 6914, 7071, 7228, 7385, 7542, 7700, 7857, 8014, 8171, 8328, 8485, 8642, 8800, 8957, 9114, 9271, 9428, 9585, 9742, 9900]}, "sampling": {"method": "uniform_time_floor", "history": "from_start_to_question", "max_n_frames": 64}, "video_preprocessing": {"size": {"shortest_edge": 1048576, "longest_edge": 4194304}, "do_sample_frames": false, "cap_pixels_per_frame": false}, "input_ids_shape": [1, 2299], "video_grid_thw": [[32, 12, 20]], "ground_truth": "A", "model_load_s": 0.00011225102934986353, "gpu": "NVIDIA GeForce RTX 4090", "inference_s": 0.28503977798391134, "peak_allocated_gib": 8.893391609191895, "peak_reserved_gib": 9.1015625, "generation": {"max_new_tokens": 32, "do_sample": false}, "warmup_performed": true, "raw_output": "A", "prediction": "A", "correct": 1, "total_s": 0.974115936900489}
{"case_path": "/root/Reseach/qwen-video/outputs/review/frames_64/sample_8_a1403_q0/case.json", "model_dir": "/model/ModelScope/Qwen/Qwen3-VL-4B-Instruct", "torch_version": "2.13.0+cu132", "transformers_version": "5.19.0", "status": "ok", "timing_scope": "cached_frames_to_answer", "question_id": "sample_8_a1403_q0", "question_timestamp_s": 5, "num_frames": 64, "input_frame_timestamps_s": [0.0, 0.06673333333333334, 0.15015, 0.23356666666666667, 0.3003, 0.38371666666666665, 0.46713333333333334, 0.55055, 0.6172833333333333, 0.7007, 0.7841166666666667, 0.8675333333333334, 0.9342666666666667, 1.0176833333333333, 1.1011, 1.1845166666666667, 1.25125, 1.3346666666666667, 1.4180833333333334, 1.5015, 1.5682333333333334, 1.65165, 1.7350666666666668, 1.8184833333333332, 1.8852166666666668, 1.9686333333333332, 2.05205, 2.135466666666667, 2.2022, 2.285616666666667, 2.3690333333333333, 2.45245, 2.5191833333333333, 2.6026, 2.6860166666666667, 2.769433333333333, 2.8361666666666667, 2.919583333333333, 3.003, 3.0864166666666666, 3.15315, 3.2365666666666666, 3.3199833333333335, 3.4034, 3.4701333333333335, 3.55355, 3.6369666666666665, 3.7203833333333334, 3.7871166666666665, 3.8705333333333334, 3.95395, 4.037366666666666, 4.1041, 4.187516666666666, 4.270933333333334, 4.35435, 4.421083333333334, 4.5045, 4.587916666666667, 4.671333333333333, 4.738066666666667, 4.821483333333333, 4.9049, 4.988316666666667], "video_metadata": {"fps": 59.94005994005994, "total_num_frames": 18800, "frames_indices": [0, 4, 9, 14, 18, 23, 28, 33, 37, 42, 47, 52, 56, 61, 66, 71, 75, 80, 85, 90, 94, 99, 104, 109, 113, 118, 123, 128, 132, 137, 142, 147, 151, 156, 161, 166, 170, 175, 180, 185, 189, 194, 199, 204, 208, 213, 218, 223, 227, 232, 237, 242, 246, 251, 256, 261, 265, 270, 275, 280, 284, 289, 294, 299]}, "sampling": {"method": "uniform_time_floor", "history": "from_start_to_question", "max_n_frames": 64}, "video_preprocessing": {"size": {"shortest_edge": 1048576, "longest_edge": 4194304}, "do_sample_frames": false, "cap_pixels_per_frame": false}, "input_ids_shape": [1, 2265], "video_grid_thw": [[32, 12, 20]], "ground_truth": "D", "model_load_s": 0.00010379101149737835, "gpu": "NVIDIA GeForce RTX 4090", "inference_s": 0.28460588899906725, "peak_allocated_gib": 8.886435508728027, "peak_reserved_gib": 9.1015625, "generation": {"max_new_tokens": 32, "do_sample": false}, "warmup_performed": true, "raw_output": "D", "prediction": "D", "correct": 1, "total_s": 0.9173228170257062}
{"case_path": "/root/Reseach/qwen-video/outputs/review/frames_64/sample_8_a1403_q1/case.json", "model_dir": "/model/ModelScope/Qwen/Qwen3-VL-4B-Instruct", "torch_version": "2.13.0+cu132", "transformers_version": "5.19.0", "status": "ok", "timing_scope": "cached_frames_to_answer", "question_id": "sample_8_a1403_q1", "question_timestamp_s": 14, "num_frames": 64, "input_frame_timestamps_s": [0.0, 0.21688333333333334, 0.4337666666666667, 0.65065, 0.8842166666666667, 1.1011, 1.3179833333333333, 1.55155, 1.7684333333333333, 1.9853166666666666, 2.2188833333333333, 2.435766666666667, 2.65265, 2.886216666666667, 3.1031, 3.3199833333333335, 3.55355, 3.7704333333333335, 3.9873166666666666, 4.220883333333333, 4.437766666666667, 4.65465, 4.871533333333334, 5.1051, 5.321983333333334, 5.538866666666666, 5.772433333333334, 5.989316666666666, 6.2062, 6.439766666666666, 6.65665, 6.8735333333333335, 7.1071, 7.3239833333333335, 7.540866666666667, 7.7744333333333335, 7.991316666666667, 8.2082, 8.441766666666666, 8.65865, 8.875533333333333, 9.1091, 9.325983333333333, 9.542866666666667, 9.75975, 9.993316666666667, 10.2102, 10.427083333333334, 10.66065, 10.877533333333334, 11.094416666666667, 11.327983333333334, 11.544866666666667, 11.76175, 11.995316666666668, 12.2122, 12.429083333333333, 12.66265, 12.879533333333333, 13.096416666666666, 13.329983333333333, 13.546866666666666, 13.76375, 13.997316666666666], "video_metadata": {"fps": 59.94005994005994, "total_num_frames": 18800, "frames_indices": [0, 13, 26, 39, 53, 66, 79, 93, 106, 119, 133, 146, 159, 173, 186, 199, 213, 226, 239, 253, 266, 279, 292, 306, 319, 332, 346, 359, 372, 386, 399, 412, 426, 439, 452, 466, 479, 492, 506, 519, 532, 546, 559, 572, 585, 599, 612, 625, 639, 652, 665, 679, 692, 705, 719, 732, 745, 759, 772, 785, 799, 812, 825, 839]}, "sampling": {"method": "uniform_time_floor", "history": "from_start_to_question", "max_n_frames": 64}, "video_preprocessing": {"size": {"shortest_edge": 1048576, "longest_edge": 4194304}, "do_sample_frames": false, "cap_pixels_per_frame": false}, "input_ids_shape": [1, 2277], "video_grid_thw": [[32, 12, 20]], "ground_truth": "D", "model_load_s": 0.00011302100028842688, "gpu": "NVIDIA GeForce RTX 4090", "inference_s": 0.28533468290697783, "peak_allocated_gib": 8.888890743255615, "peak_reserved_gib": 9.1015625, "generation": {"max_new_tokens": 32, "do_sample": false}, "warmup_performed": true, "raw_output": "D", "prediction": "D", "correct": 1, "total_s": 0.9463465349981561}
{"case_path": "/root/Reseach/qwen-video/outputs/review/frames_64/sample_8_a1404_q0/case.json", "model_dir": "/model/ModelScope/Qwen/Qwen3-VL-4B-Instruct", "torch_version": "2.13.0+cu132", "transformers_version": "5.19.0", "status": "ok", "timing_scope": "cached_frames_to_answer", "question_id": "sample_8_a1404_q0", "question_timestamp_s": 85, "num_frames": 64, "input_frame_timestamps_s": [0.0, 1.3346666666666667, 2.6860166666666667, 4.037366666666666, 5.388716666666666, 6.740066666666666, 8.091416666666667, 9.426083333333333, 10.777433333333333, 12.128783333333333, 13.480133333333333, 14.831483333333333, 16.182833333333335, 17.534183333333335, 18.86885, 20.2202, 21.57155, 22.9229, 24.27425, 25.6256, 26.97695, 28.3283, 29.662966666666666, 31.014316666666666, 32.36566666666667, 33.717016666666666, 35.06836666666667, 36.419716666666666, 37.75438333333334, 39.10573333333333, 40.45708333333334, 41.80843333333333, 43.15978333333333, 44.51113333333333, 45.86248333333333, 47.19715, 48.5485, 49.89985, 51.2512, 52.60255, 53.9539, 55.30525, 56.6566, 57.99126666666667, 59.342616666666665, 60.69396666666667, 62.045316666666665, 63.39666666666667, 64.74801666666667, 66.08268333333334, 67.43403333333333, 68.78538333333333, 70.13673333333334, 71.48808333333334, 72.83943333333333, 74.19078333333333, 75.52545, 76.8768, 78.22815, 79.5795, 80.93085, 82.2822, 83.63355, 84.9849], "video_metadata": {"fps": 59.94005994005994, "total_num_frames": 18800, "frames_indices": [0, 80, 161, 242, 323, 404, 485, 565, 646, 727, 808, 889, 970, 1051, 1131, 1212, 1293, 1374, 1455, 1536, 1617, 1698, 1778, 1859, 1940, 2021, 2102, 2183, 2263, 2344, 2425, 2506, 2587, 2668, 2749, 2829, 2910, 2991, 3072, 3153, 3234, 3315, 3396, 3476, 3557, 3638, 3719, 3800, 3881, 3961, 4042, 4123, 4204, 4285, 4366, 4447, 4527, 4608, 4689, 4770, 4851, 4932, 5013, 5094]}, "sampling": {"method": "uniform_time_floor", "history": "from_start_to_question", "max_n_frames": 64}, "video_preprocessing": {"size": {"shortest_edge": 1048576, "longest_edge": 4194304}, "do_sample_frames": false, "cap_pixels_per_frame": false}, "input_ids_shape": [1, 2278], "video_grid_thw": [[32, 12, 20]], "ground_truth": "C", "model_load_s": 0.0001230410998687148, "gpu": "NVIDIA GeForce RTX 4090", "inference_s": 0.2838136489735916, "peak_allocated_gib": 8.889094829559326, "peak_reserved_gib": 9.1015625, "generation": {"max_new_tokens": 32, "do_sample": false}, "warmup_performed": true, "raw_output": "C", "prediction": "C", "correct": 1, "total_s": 0.9355213310336694}