Terry Wang

使用similarity.py进行推理，限定32C的情况下，速度耗时800毫秒

2

使用similarity.py进行推理，限定32C的情况下，速度耗时800毫秒。 而标准版本的bert速度大概在300毫秒。

I have train almost 80thousand examples within 2000 labels，valid acc almost 92%，but test result all example prob is blew 0.01.

6

I have trained almost 80thousand examples within 2000 labels，valid acc almost 92%，but test result all example prob is blew 0.01. I have tried tranning examples to predict.

Can provide links of Albert's model and sentence word model

2

Can provide links of Albert's model and sentence word model ,Thks.

train and crash on jupyter

1

``` from graphvite.dataset import Dataset import numpy as np import graphvite as gv import graphvite.application as gap import networkx as nx G = nx.barabasi_albert_graph(100,3) nx.write_edgelist(G, 'data/test.edge_list', data=False, delimiter='\t') app =...

[ 13%] Performing build step for 'jemalloc' cp include/jemalloc/internal/private_namespace.gen.h include/jemalloc/internal/private_namespace.gen.h [ 13%] Performing install step for 'jemalloc' Skipping install step. [ 13%] Completed 'jemalloc' [ 13%] Built target jemalloc

thks

生成的摘要基本不可用

2

1、数据集用自己得 2、调整了训练的steps变短些 结果惨不忍睹，是不是哪里设置错了。