1. 18 12月, 2019 4 次提交
  2. 17 12月, 2019 6 次提交
  3. 16 12月, 2019 1 次提交
  4. 15 12月, 2019 3 次提交
  5. 14 12月, 2019 4 次提交
  6. 13 12月, 2019 5 次提交
  7. 12 12月, 2019 1 次提交
  8. 11 12月, 2019 4 次提交
  9. 10 12月, 2019 1 次提交
  10. 08 12月, 2019 1 次提交
  11. 07 12月, 2019 4 次提交
    • C
      Internal Change · 440f31ab
      Chen Chen 提交于
      PiperOrigin-RevId: 284278191
      440f31ab
    • S
      Push CLs 283653640, 284011539 (#7921) · 22e20f84
      Skirmantas Kligys 提交于
      * Manually quantize selected inputs before summing them up.
      
      PiperOrigin-RevId: 283653640
      
      * Fix messy variable paths in LSTD head.
      
      With a VariableScope passed as parameter into tf.variable_scope(),
      TF gets all confused and starts placing variables into scopes different
      from operations (variable_scope != name_scope).  That leads to very messy
      TF graphs.
      
      Before CL:
        FeatureExtractor/LSTM/conv_lstm_cell/bottleneck_0/depthwise_weights
        FeatureExtractor/LSTM/LSTM/conv_lstm_cell/bottleneck_0/separable_conv2d/depthwise
      
      After CL:
        FeatureExtractor/LSTM/conv_lstm_cell/bottleneck_0/depthwise_weights
        FeatureExtractor/LSTM/conv_lstm_cell/bottleneck_0/separable_conv2d/depthwise
      
      PiperOrigin-RevId: 284011539
      22e20f84
    • Y
      Internal cleanup. · 8a8baded
      Yeqing Li 提交于
      PiperOrigin-RevId: 284259997
      8a8baded
    • C
      Internal change · 6e847fbd
      Chen Chen 提交于
      PiperOrigin-RevId: 284203261
      6e847fbd
  12. 06 12月, 2019 4 次提交
  13. 05 12月, 2019 2 次提交
    • J
      Internal change · c115444f
      Jaehong Kim 提交于
      PiperOrigin-RevId: 283962490
      c115444f
    • H
      Remove tweaked benchmark for no_dist_strat in function mode. · ef8aed79
      Haoyu Zhang 提交于
      No need to test explicit_gpu_placement + no_dist_strat + function case, since the performance issue is already fixed by unifying the Keras codepaths. However, there is still performance gap using the explicit_gpu_placement flag when run_eagerly.
      
      PiperOrigin-RevId: 283890008
      ef8aed79