1073521013

‘answer_coordinates’ and ‘answers’ predicted as '[]'

6

Hello, I would like to ask a question. After the model fineturning is completed, answer_coordinates and answers are predicted [], but I can get the value directly using the pretrained...

Inefficiency Decoder Training with Multi-GPU on Several Nodes (Machines)

2

Hello everyone, training with multi-GPU on one node (machine) is normal, but multi-GPU on several nodes (machines) training is Inefficiency. By sorting out the problem, it is found that it...

A merge layer should be called on a list of inputs

2

**Describe the bug(问题描述)** 保存模型model.save，然后加载的时候，tf.keras.models.load_model报错"A merge layer should be called on a list of inputs",官方找到类似问题说更新keras,现在最新2.4.3还是报错 **To Reproduce(复现步骤)** DeepFM模型完整保存，model.save() **Operating environment(运行环境):** - python version 3.7 - tensorflow version 2.2.0 - keras version...

麻烦问两个问题，感谢 1 请问loss在预训练的下降情况是怎么样的，各自初始值是多少呢，最终各自收敛到什么情况呢，finetune大概是什么情况呢 2 lr一般和batch关系比较紧密，那请问finetune过程中你们的batch和lr大概多少呢

预训练中构建全词mask问题

4

提问时请尽可能提供如下信息： ### 基本信息 - 你使用的**操作系统**: - 你使用的**Python**版本: - 你使用的**Tensorflow**版本: - 你使用的**Keras**版本: - 你使用的**bert4keras**版本: - 你使用纯**keras**还是**tf.keras**: - 你加载的**预训练模型**: ### 核心代码 ```python # 请在此处贴上你的核心代码。 # 请尽量只保留关键部分，不要无脑贴全部代码。 ``` def sentence_process(self, text): """单个文本的处理函数 流程：分词，然后转id，按照mask_rate构建全词mask的序列...

The upsample model generates more regular noise points

11

Hello everyone, I would like to ask if you have encountered when training upsample (64-->256), the generated images always show the following question: 1 In the same batch, some images...

您好，请问Multilingual-Cultural- 18这个数据可以开源吗

Github: https://github.com/OPPO-Mente-Lab/X2I Huggingface: https://huggingface.co/OPPOer/X2I