Tao He

# ❓ Questions & Help Thanks for your great works on ASR. I am wondering if we have training result data for provided model? I think it is important for...

RNNT training on CPU

1

Appreciate for the job on supporting RNN-T training on CPU (models/language_modeling/pytorch/rnnt/training/cpu), just quick evaluated the training code and found that WER would keep in 1.00 after even training 10+ epoches....

segfault with large tensor for pytorch

1

Hi, I am facing segmentation fault issue with large input tensor. Test code: ``` import torch from warprnnt_pytorch import RNNTLoss rnnt_loss = RNNTLoss() acts = torch.ones(128,256,129,1024, requires_grad=True) labels = torch.ones(128,128)...

Spark3 tpcds setup

1

Noticed that spark 3.0.0 does not have method org.apache.spark.sql.SQLContext.createExternalTable, updated the TPC-DS setup part to support spark 3.0.0 by using SparkSession to create table. Thanks for your reviewing.

alltoall performance regression after upgrading from 2021.1-beta07-1 to 1.10

1

Hi, We upgraded torch-ccl from 2021.1-beta07-1 to 1.10 and noticed some performance regression for all_to_all. overall, ccl 1.10 is 2x worse than 2021.1-beta07-1. **system config:** - single node, 2 proc_per_node,...