Voyager 3

onnxruntime.capi.onnxruntime_pybind11_state.Fail: [ONNXRuntimeError] : 1 : FAIL : Load model from +all-intermediate.onnx failed:/onnxruntime_src/onnxruntime/core/graph/graph.cc:1170 void

6

(py36) $ onnx_opcounter --calculate-macs dd3d_b1_cv22.onnx Number of parameters in the model: 6985714 Conv_0 -> Conv_0 index 5 Relu_2 -> Relu_2 index 6 MaxPool_3 -> MaxPool_3 index 7 Conv_4 -> Conv_4...

Using tensorflow/tensorflow:1.12-gpu Although this docker image doesn't installed cudnn, the msraner_task baseline still work. the only problem is it's too slow. run on my 1080Ti 11GB GPU. But the speed...

python3.x is not supported by the tensorflow baseline models

1

when I use tensorflow/tensorflow1.12-gpu-py3 docker images to run the ner_task, it crashed with following messages: nvidia-smi check pass Tue Jan 7 07:06:25 UTC 2020 File "run_ner.py", line 40, in reload(sys)...

What's the difference between local and global attention maps?

2

Thanks for great work! pytorch version is really readable after reading the paper, I confused what's the difference between local and global attention maps? Anyone can help me

/home/csx/remote_workdir/Bert_people_paper_2014_pytorch/preprocessing/data_processor.py line 244-247 ` ## Notes: label_id中不包括[CLS]和[SEP]` ` label_id = [label_map[l] for l in labels]` ` label_padding = [-1] * (max_seq_length-len(label_id))` ` label_id += label_padding` 考虑到input第一个字符是CLS，最后一个字符是SEP，label在padding时应该和input对齐，开头结尾分别是CLS，SEP，置为-1 `label_id = [-1]+label_id +[-1]...

Is there any weight larger than 1 to strengthen the addative noise?

1

when transfer to my own images (safe vision system), the noise could not be elimate by addative real noise from noise pool. So Is there any weight to strengthen the...

中文年份含〇转换失败：二〇〇二年的第一场雪

2

**Describe the bug** 中文年份含〇转换失败〇 To Reproduce Steps to reproduce the behavior: 1. run following code: ``` python import cn2an result = cn2an.transform(input_sentence) print(result) ``` 2. See error **Expected behavior** 2002年的第1场雪...

练习时长两年半

1

**Describe the bug** 练习时长两年半 转换结果不符合预期 To Reproduce Steps to reproduce the behavior: 1. run following code: ``` python import cn2an result = cn2an.transform(input_sentence) print(result) ``` 2. See error **Expected behavior**...

**Describe the bug** 转换错误：一个大西瓜切两半->1个大西瓜切20.5 To Reproduce Steps to reproduce the behavior: 1. run following code: ``` python import cn2an result = cn2an.transform(input_sentence) print(result) ``` 2. See error **Expected behavior** 1个大西瓜切2半...

fix the #190 Issue The mysql and qanything_local containers use an option `privileged: true` in both `docker-compose-linux.yaml` and `docker-compose-windows.yaml`. However, the default namespace of docker container is user, which is...