| Commit message (Collapse) | Author | Age |
|
|
|
|
|
|
|
|
|
|
|
| |
This is useful for limiting the amount of debug tensor data dumped and/or focusing the debugging on specific parts of the graph.
For example:
tfdbg> run --node_name_filter hidden
tfdbg> run --op_type_filter Variable.*
tfdbg> run --tensor_dtype_filter int.*
RELNOTES: The "run" command of tfdbg's command-line interface now supports filtering of tensors by node name, op type and tensor dtype.
Change: 151405236
|
|
|
|
| |
Change: 151403279
|
|
|
|
|
|
| |
Fixes #7877
Fixes #6926
Change: 151400818
|
|
|
|
| |
Change: 151399581
|
|
|
|
| |
Change: 151396171
|
|
|
|
| |
Change: 151395519
|
|
|
|
| |
Change: 151394592
|
|
|
|
|
|
| |
bucket and bucket_by_sequence_length.
Change: 151393852
|
|
|
|
|
| |
package.
Change: 151393424
|
|
|
|
|
|
| |
Not enabled by default, but enable it with construction flag
attention_history=True.
Change: 151391741
|
|
|
|
| |
Change: 151391358
|
|
|
|
| |
Change: 151390706
|
|
|
|
|
| |
processed
Change: 151383375
|
|
|
|
| |
Change: 151382854
|
|
|
|
| |
Change: 151380006
|
|
|
|
| |
Change: 151375546
|
|
|
|
| |
Change: 151374696
|
|
|
|
| |
Change: 151374177
|
|
|
|
|
| |
issue was reported by a customer that uses XLA:CPU in production.
Change: 151370263
|
|
|
|
| |
Change: 151366403
|
|
|
|
| |
Change: 151366249
|
|
|
|
| |
Change: 151366175
|
|
|
|
|
| |
constant folding from wiping out a large part of the test network
Change: 151354452
|
|
|
|
| |
Change: 151352935
|
|
|
|
| |
Change: 151351473
|
|
|
|
| |
Change: 151349771
|
|
|
|
| |
Change: 151345424
|
|
|
|
| |
Change: 151344092
|
|
|
|
| |
Change: 151341806
|
|
|
|
| |
Change: 151341608
|
|
|
|
|
|
| |
Trainability in Recurrent Neural Networks
Change: 151339833
|
|
|
|
|
| |
output_attention is False.
Change: 151337979
|
|
|
|
| |
Change: 151335029
|
|
|
|
|
| |
Missed an incref the first time.
Change: 151332384
|
|
|
|
| |
Change: 151332329
|
|
|
|
| |
Change: 151331999
|
|
|
|
| |
Change: 151331705
|
|
|
|
|
|
| |
roughly equivalent to np.percentile
Change: 151317484
|
|
|
|
|
| |
requested_port_ is only used within a single function, and thus would be better as a local variable.
Change: 151285269
|
|
|
|
| |
Change: 151282782
|
|
|
|
| |
Change: 151267767
|
|
|
|
|
|
| |
- an approximate kernel mapper for the RBF kernel (based on Random Fourier Features)
- a (canned) tf.learn kernel-based classifier.
Change: 151237967
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
State in workers is currently splayed across graph_mgr, rendezvous_mgr, and
additional components. This has resulted in it being difficult to ensure proper
cleanup and shut down of the worker components.
In addition to paving the way for a more reliable shut down, this CL also sets
up the beginnings of ClusterSpec propagation.
ClusterSpec propagation is a capability upgrade for TensorFlow that should make
it much easier to (1) build distributed TensorFlow clusters, and (2) handle
node failures. After the ClusterSpec propagation capability is fully
implemented, the TensorFlow workers can be booted independently of each other,
and with no knowledge about others. A client can then query a central cluster
scheduler or other API to find all of the workers, and then send the
ClusterDef (ClusterSpec) to the TF master, which then propagates that along to
all of the workers.
This change is only the first of a sequence to fully implement ClusterSpec
propagation in TensorFlow.
Change: 151229111
|
|
|
|
| |
Change: 151201476
|
|
|
|
| |
Change: 151191936
|
|
|
|
| |
Change: 151188187
|
|
|
|
| |
Change: 151185292
|
|
|
|
|
|
| |
tensorflow/contrib/. (Leave shims in the old location for now.)
Change: 151184823
|
|
|
|
|
|
|
| |
Extend HLO rematerialization to rematerialize subcomputations in addition to the entry computations. Outer nesting levels of computations are rematerialized before inner nesting levels because inner subcomputations may be while bodies where rematerialization is more expensive. Also
Also fix latent bug in call_graph dealing with fusion instructions, and extend HloInstruction::Clone to accept a string suffix (eg, "remat") for the clone name.
Change: 151179956
|
|
|
|
|
|
| |
- Useful for allocators which want to override only some functionality of
another allocator.
Change: 151179248
|