Search results

Nan loss when replace pytorch LSTM with your LSTM or LayerNormLSTM

Status: Open.

#50 In lmnt-com/haste;

· andkot opened on Oct 9, 2022

Supporting RWKV (a RNN that can match transformer LM & zero-shot performance at 1B+ params)

Status: Open.

#48 In lmnt-com/haste;

· BlinkDL opened on Jul 18, 2022

haste_tf compilation fails with "‘bfloat16’ in namespace ‘Eigen’ does not name a type"

Status: Open.

#45 In lmnt-com/haste;

· alethes opened on Feb 9, 2022

Can Bidirectional Rnn and multi-layer Rnn be supported?

Status: Open.

#38 In lmnt-com/haste;

· zhang662817 opened on Jul 2, 2021

return_state_sequence for tf version

Status: Open.

#36 In lmnt-com/haste;

· thegodone opened on Jun 16, 2021

Can't run haste layers in Keras

Status: Open.

#33 In lmnt-com/haste;

· mahenning opened on Jun 5, 2021

Support zoneout on lstm cell state and add recurrent dropout

Status: Open.

#29 In lmnt-com/haste;

· azraelkuan opened on Nov 12, 2020

Install on pip on systems without cuda

Status: Open.

#27 In lmnt-com/haste;

· bratao opened on Sep 16, 2020

haste_tf\libhaste_tf.so not found

Status: Open.

#23 In lmnt-com/haste;

· MichaelJanz opened on Jul 13, 2020

IndRNNs

Status: Open.

#7 In lmnt-com/haste;

· OverLordGoldDragon opened on Mar 16, 2020