search

KyungMinJin / HANet

Kinematic-aware Hierarchical Attention Network for Human Pose Estimation in Videos (WACV 2023)
50 stars 4 forks source link

RuntimeError: "baddbmm_cuda" not implemented for 'Int' #1

Open fj6833 opened 1 year ago

fj6833 commented 1 year ago

Namespace(body_representation='2D', cfg='configs/config_jhmdb_simplebaseline_2D.yaml', dataset_name='jhmdb', estimator='simplebaseline')

Seed value for the experiment is 4321 GPU name -> NVIDIA GeForce RTX 3060 Laptop GPU GPU feat -> _CudaDeviceProperties(name='NVIDIA GeForce RTX 3060 Laptop GPU', major=8, minor=6, total_memory=6143MB, multi_processor_count=30) {'BODY_REPRESENTATION': '2D', 'CUDNN': CfgNode({'BENCHMARK': True, 'DETERMINISTIC': False, 'ENABLED': True}), 'DATASET': {'AIST': {'DETECTED_PATH': './data\detected_poses/aist', 'GROUND_TRUTH_PATH': './data\groundtruth_poses/aist', 'KEYPOINT_NUM': 14, 'KEYPOINT_ROOT': [2, 3]}, 'H36M': {'DETECTED_PATH': './data\detected_poses/h36m', 'GROUND_TRUTH_PATH': './data\groundtruth_poses/h36m', 'KEYPOINT_NUM': 17, 'KEYPOINT_ROOT': [0]}, 'JHMDB': {'DETECTED_PATH': './data\detected_poses/jhmdb', 'GROUND_TRUTH_PATH': './data\groundtruth_poses/jhmdb', 'KEYPOINT_NUM': 15, 'KEYPOINT_ROOT': [2]}, 'PW3D': {'DETECTED_PATH': './data\detected_poses/pw3d', 'GROUND_TRUTH_PATH': './data\groundtruth_poses/pw3d', 'KEYPOINT_NUM': 14, 'KEYPOINT_ROOT': [2, 3]}}, 'DATASET_NAME': 'jhmdb', 'DEBUG': True, 'DEVICE': 'cuda', 'ESTIMATOR': 'simplebaseline', 'EVALUATE': {'DECODER': False, 'INTERP': 'linear', 'PRETRAINED': 'results/30-08-2022_16-06-59_jhmdb_simplebaseline_N10_10/EP68_PCK@0.05_0.89_checkpoint.pth.tar', 'RELATIVE_IMPROVEMENT': False, 'ROOT_RELATIVE': True, 'SLIDE_WINDOW_STEP_Q': 1, 'SLIDE_WINDOW_STEP_SIZE': 10}, 'EXP_NAME': 'jhmdb_simplebaseline_N10_1_256', 'GPUS': ['0'], 'LOG': CfgNode({'NAME': ''}), 'LOGDIR': 'results\29-10-2022_17-15-29_jhmdb_simplebaseline_N10_1_256', 'LOSS': CfgNode({'LAMADA': 5.0, 'W_DECODER': 1.0}), 'MODEL': {'DECODER': 'transformer', 'DECODER_EMBEDDING_DIMENSION': 256, 'DECODER_HEAD': 4, 'DECODER_INTERP': 'linear', 'DECODER_RESIDUAL': True, 'DECODER_TOKEN_WINDOW': 5, 'DECODER_TRANSFORMER_BLOCK': 5, 'DROPOUT': 0.1, 'ENCODER_EMBEDDING_DIMENSION': 256, 'ENCODER_HEAD': 4, 'ENCODER_RESIDUAL': True, 'ENCODER_TRANSFORMER_BLOCK': 5, 'INTERVAL_N': 10, 'NAME': '', 'SAMPLE_TYPE': 'uniform', 'SLIDE_WINDOW': True, 'SLIDE_WINDOW_Q': 1, 'SLIDE_WINDOW_SIZE': 11, 'TYPE': 'network'}, 'OUTPUT_DIR': 'results', 'SAMPLE_INTERVAL': 10, 'SEED_VALUE': 4321, 'SMPL_MODEL_DIR': 'data/smpl/', 'TRAIN': {'BATCH_SIZE': 16, 'EPOCH': 70, 'LR': 0.001, 'LRDECAY': 0.95, 'PRE_NORM': False, 'RESUME': None, 'USE_6D_SMPL': False, 'USE_SMPL_LOSS': False, 'VALIDATE': True, 'WORKERS_NUM': 0}, 'VIS': {'END': 100, 'INPUT_VIDEO_NUMBER': 160, 'INPUT_VIDEO_PATH': 'data/videos/', 'OUTPUT_VIDEO_PATH': 'demo/', 'START': 0}} ############################################################# You are loading the [training set] of dataset [jhmdb] You are using pose esimator [simplebaseline] The type of the data is [2D] The frame number is [24372] The sequence number is [687] ############################################################# ############################################################# You are loading the [testing set] of dataset [jhmdb] You are using pose esimator [simplebaseline] The type of the data is [2D] The frame number is [9228] The sequence number is [261] ############################################################# Slide window: 11 Sample interval: 10

Traceback (most recent call last): File "train.py", line 109, in main(cfg) File "train.py", line 96, in main Trainer(train_dataloader=train_loader, File "D:\GitLoadWareHouse\HANet\lib\core\trainer.py", line 71, in run self.train() File "D:\GitLoadWareHouse\HANet\lib\core\trainer.py", line 124, in train predicted_3d_pos, decoderd_3d_pos = self.model( File "D:\Environment\anaconda3\envs\pytorch\lib\site-packages\torch\nn\modules\module.py", line 1130, in _call_impl return forward_call(*input, kwargs) File "D:\GitLoadWareHouse\HANet\lib\models\HANet.py", line 175, in forward self.hierarchical_encoder, self.decoder = self.transformer.forward( File "D:\GitLoadWareHouse\HANet\lib\models\HANet.py", line 335, in forward output = self.decode(mem, encoder_mask, encoder_pos_embed[0], trans_tgt, File "D:\GitLoadWareHouse\HANet\lib\models\HANet.py", line 373, in decode hs = self.decoder(tgt, File "D:\Environment\anaconda3\envs\pytorch\lib\site-packages\torch\nn\modules\module.py", line 1130, in _call_impl return forward_call(*input, *kwargs) File "D:\GitLoadWareHouse\HANet\lib\models\HANet.py", line 429, in forward output = layer(output, File "D:\Environment\anaconda3\envs\pytorch\lib\site-packages\torch\nn\modules\module.py", line 1130, in _call_impl return forward_call(input, kwargs) File "D:\GitLoadWareHouse\HANet\lib\models\HANet.py", line 633, in forward return self.forward_post(tgt, memory, tgt_mask, memory_mask, File "D:\GitLoadWareHouse\HANet\lib\models\HANet.py", line 570, in forward_post tgt2 = self.self_attn(q, File "D:\Environment\anaconda3\envs\pytorch\lib\site-packages\torch\nn\modules\module.py", line 1130, in _call_impl return forward_call(*input, **kwargs) File "D:\Environment\anaconda3\envs\pytorch\lib\site-packages\torch\nn\modules\activation.py", line 1153, in forward attn_output, attn_output_weights = F.multi_head_attention_forward( File "D:\Environment\anaconda3\envs\pytorch\lib\site-packages\torch\nn\functional.py", line 5179, in multi_head_attention_forward attn_output, attn_output_weights = _scaled_dot_product_attention(q, k, v, attn_mask, dropout_p) File "D:\Environment\anaconda3\envs\pytorch\lib\site-packages\torch\nn\functional.py", line 4852, in _scaled_dot_product_attention attn = torch.baddbmm(attn_mask, q, k.transpose(-2, -1)) RuntimeError: "baddbmm_cuda" not implemented for 'Int' [?25h

Hello, the author. I encountered this problem during training. How can I solve it?

KyungMinJin commented 1 year ago

Hi, @fj6833! Thanks for letting me know! Try with python3.6 and let me know if it doesn't work.