From 1a6d7f5acd50ed23c38f14e11e563f771a596656 Mon Sep 17 00:00:00 2001 From: Tiezhen WANG Date: Mon, 20 Aug 2018 11:45:04 -0700 Subject: Doc: Update xla example instruction. Otherwise it fails with 2018-08-20 08:28:40.998261: F tensorflow/compiler/xla/service/hlo_graph_dumper.cc:1431] Check failed: graph_renderer != nullptr No registered renderer for the HLO graph. Use --xla_hlo_graph_path=PATH to export to local file system Aborted PiperOrigin-RevId: 209456948 --- tensorflow/docs_src/performance/xla/jit.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'tensorflow/docs_src') diff --git a/tensorflow/docs_src/performance/xla/jit.md b/tensorflow/docs_src/performance/xla/jit.md index 7202ef47f7..83b3e71566 100644 --- a/tensorflow/docs_src/performance/xla/jit.md +++ b/tensorflow/docs_src/performance/xla/jit.md @@ -133,7 +133,7 @@ Execute the python script to train the model with XLA and turn on a debugging feature of XLA via an environmental variable that outputs the XLA graph. ```shell -TF_XLA_FLAGS=--xla_generate_hlo_graph=.* python mnist_softmax_xla.py +TF_XLA_FLAGS="--xla_hlo_graph_path=/tmp --xla_generate_hlo_graph=.*" python mnist_softmax_xla.py ``` Open the timeline file created (`timeline.ctf.json`). The rendered timeline -- cgit v1.2.3