AssertionError: Given weight dir /home/b308/hao/models-develop/dygraph/bmn/checkpoint/bmn.pdparams not exist.

[haohaichao]

(czk) b308@b308:~/hao/models-develop/dygraph/bmn$ python eval.py --weights=checkpoint/bmn.pdparams [INFO: config_utils.py: 80]: ---------------- Test Arguments ---------------- [INFO: config_utils.py: 82]: MODEL: [INFO: config_utils.py: 84]: name:BMN [INFO: config_utils.py: 84]: tscale:100 [INFO: config_utils.py: 84]: dscale:100 [INFO: config_utils.py: 84]: feat_dim:400 [INFO: config_utils.py: 84]: prop_boundary_ratio:0.5 [INFO: config_utils.py: 84]: num_sample:32 [INFO: config_utils.py: 84]: num_sample_perbin:3 [INFO: config_utils.py: 84]: anno_file:./activitynet_1.3_annotations.json [INFO: config_utils.py: 84]: feat_path:./fix_feat_100 [INFO: config_utils.py: 82]: TRAIN: [INFO: config_utils.py: 84]: subset:train [INFO: config_utils.py: 84]: epoch:9 [INFO: config_utils.py: 84]: batch_size:16 [INFO: config_utils.py: 84]: num_threads:8 [INFO: config_utils.py: 84]: use_gpu:True [INFO: config_utils.py: 84]: num_gpus:4 [INFO: config_utils.py: 84]: learning_rate:0.001 [INFO: config_utils.py: 84]: learning_rate_decay:0.1 [INFO: config_utils.py: 84]: lr_decay_iter:4200 [INFO: config_utils.py: 84]: l2_weight_decay:0.0001 [INFO: config_utils.py: 82]: VALID: [INFO: config_utils.py: 84]: subset:validation [INFO: config_utils.py: 84]: batch_size:16 [INFO: config_utils.py: 84]: num_threads:8 [INFO: config_utils.py: 84]: use_gpu:True [INFO: config_utils.py: 84]: num_gpus:4 [INFO: config_utils.py: 82]: TEST: [INFO: config_utils.py: 84]: subset:validation [INFO: config_utils.py: 84]: batch_size:1 [INFO: config_utils.py: 84]: num_threads:1 [INFO: config_utils.py: 84]: snms_alpha:0.001 [INFO: config_utils.py: 84]: snms_t1:0.5 [INFO: config_utils.py: 84]: snms_t2:0.9 [INFO: config_utils.py: 84]: output_path:output/EVAL/BMN_results [INFO: config_utils.py: 84]: result_path:evaluate_results [INFO: config_utils.py: 82]: INFER: [INFO: config_utils.py: 84]: subset:test [INFO: config_utils.py: 84]: batch_size:1 [INFO: config_utils.py: 84]: num_threads:1 [INFO: config_utils.py: 84]: snms_alpha:0.4 [INFO: config_utils.py: 84]: snms_t1:0.5 [INFO: config_utils.py: 84]: snms_t2:0.9 [INFO: config_utils.py: 84]: filelist:./infer.list [INFO: config_utils.py: 84]: output_path:output/INFER/BMN_results [INFO: config_utils.py: 84]: result_path:predict_results [INFO: config_utils.py: 85]: ------------------------------------------------- W0427 14:50:08.213558 6388 device_context.cc:447] Please NOTE: device: 0, GPU Compute Capability: 7.5, Driver API Version: 11.2, Runtime API Version: 10.2 W0427 14:50:08.214612 6388 device_context.cc:465] device: 0, cuDNN Version: 8.0. Traceback (most recent call last): File "eval.py", line 220, in test_bmn(args) File "eval.py", line 143, in test_bmn args.weights) AssertionError: Given weight dir checkpoint/bmn.pdparams not exist.

使用的是动态图进行验证，我可以确定权重文件夹存在，权重文件夹放在/models-develop/dygraph/bmn/checkpoint中，但运行之后系统报给的文件夹不存在。想询问下我这个是哪里除了问题？

[D-DanielYang]

BMN模型的一个更好的实现，可以试试PaddleVideo里面的维护的