| Commit message (Collapse) | Author | Age | |
|---|---|---|---|
| * | Does a deep copy of the tensors output from GraphRunner::Run(...) |  Vinu Rajashekhar | 2017-04-10 |
| | | | | | | | | | | | | The current GraphRunner::Run(...) outputs tensors produced from running the Executor on the graph, but these tensors are actually owned by the allocator from the device created for the Run(...), which could be deleted along with the device. The deep copy allows the ownership to be transferred to the global static cpu_allocator(). Before, the allocator was always a global cpu_allocator(), but with a recent change there is an option to tie allocations to a memory limited allocator per-session. Change: 152756520 | ||
| * | Makes GraphRunner a class to explicitly control it's lifetime. | Vinu Rajashekhar | 2017-03-31 |
| | | | | | Change: 151853846 | ||
| * | Cleanup and consolidation: Add a GraphRunner::Run function runs a |  Vijay Vasudevan | 2016-09-15 |
| graph. This is similar to what is already being done in ConstantFolding, but allows us to feed inputs too. This function replaces some of the code in ConstantFolding, and will also be used for the purposes of running constant graphs for shape inference (I have another CL that will use this, once checked in. The other CL was using DirectSession as a proof of concept, but we need something simpler to avoid layering violations). I have documented the class / function as not super general and only really useful for constant evaluations. Change: 133331175 | |||
 |
index : tensorflow | |
| machine learning framework |
| aboutsummaryrefslogtreecommitdiffhomepage |