Yuxiang Zhou
Yuxiang Zhou
Got the same issue!
> Hey @whatdhack @ft3020997, what versions of CMake are you using? Hi @tgaddair the version of cmake is 3.18.2 and the command used is as follows: HOROVOD_NCCL_HOME=/usr/local/nccl_2.7.6 HOROVOD_GPU_OPERATIONS=NCCL HOROVOD_WITH_TENSORFLOW=1 HOROVOD_WITH_MPI=1...
> Hey @whatdhack @ft3020997, what versions of CMake are you using? Hi ,i fixed this. By "conda install openmpi-mpicc" and it works well.
> Hey @ft3020997, thanks for the update! So seems that the MPI package that ships with Conda by default does not contain compilers ([conda-forge/openmpi-feedstock#34](https://github.com/conda-forge/openmpi-feedstock/issues/34)). What package did you use to...
> @brendanxwhitaker > > I have added “print(inp_k) print(inp_k.shape)” in the two_stream_loss function，but it didn't print out. > And I got this information: > I0730 15:41:08.266821 8780 tf_logging.py:115] input_k: Tensor("Cast_9:0",...
> As @Toilal commented use jsonnet-binary. For me, `pip install jsonnet-binary` fixed the issue. It works for my issue on Windows Envs. Thx!
你好，想问一下训练文本输入的事情。 假设我有2000条长文本，每条文本包含约10-20句话，每句话包含若干单词。 如何做到 “word-sent-doc”的向量化输入，因为看到feature_content的单条数据就是每个doc分词后的结果。
@RandolphVI 那模型知道该如何进行单个doc句子的截断吗？因为看到feature_content 的内容，都是[word1,word2....wordN]的格式，如果需要做成“word-sent-doc”形式的，是否需要doc1[sent1[word,word],sent2[word,word]....sentN]的输入？
@RandolphVI 十分感谢大神的解答，我先尝试一下，谢谢。