diff options
author | 2016-09-16 11:22:05 -0800 | |
---|---|---|
committer | 2016-09-16 12:32:32 -0700 | |
commit | 3505673a75b23c5f48983cf0b0230f0e7004d05e (patch) | |
tree | 933a54ba37eeb0ff4d728720cba63975aa98a87f | |
parent | a738a3bcafcd66fd241ad6e09f429dd6294f1669 (diff) |
Update generated Python Op docs.
Change: 133413914
-rw-r--r-- | tensorflow/g3doc/api_docs/python/functions_and_classes/shard4/tf.nn.nce_loss.md | 5 | ||||
-rw-r--r-- | tensorflow/g3doc/api_docs/python/nn.md | 5 |
2 files changed, 10 insertions, 0 deletions
diff --git a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard4/tf.nn.nce_loss.md b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard4/tf.nn.nce_loss.md index 2fc7ab6b65..3de6d1ae3f 100644 --- a/tensorflow/g3doc/api_docs/python/functions_and_classes/shard4/tf.nn.nce_loss.md +++ b/tensorflow/g3doc/api_docs/python/functions_and_classes/shard4/tf.nn.nce_loss.md @@ -8,6 +8,11 @@ unnormalized statistical models] Also see our [Candidate Sampling Algorithms Reference] (../../extras/candidate_sampling.pdf) +Note: By default this uses a log-uniform (Zipfian) distribution for sampling, +so your labels must be sorted in order of decreasing frequency to achieve +good results. For more details, see +[log_uniform_candidate_sampler](#log_uniform_candidate_sampler). + Note: In the case where `num_true` > 1, we assign to each target class the target probability 1 / `num_true` so that the target probabilities sum to 1 per-example. diff --git a/tensorflow/g3doc/api_docs/python/nn.md b/tensorflow/g3doc/api_docs/python/nn.md index a163cc8ec7..724476efa8 100644 --- a/tensorflow/g3doc/api_docs/python/nn.md +++ b/tensorflow/g3doc/api_docs/python/nn.md @@ -2604,6 +2604,11 @@ unnormalized statistical models] Also see our [Candidate Sampling Algorithms Reference] (../../extras/candidate_sampling.pdf) +Note: By default this uses a log-uniform (Zipfian) distribution for sampling, +so your labels must be sorted in order of decreasing frequency to achieve +good results. For more details, see +[log_uniform_candidate_sampler](#log_uniform_candidate_sampler). + Note: In the case where `num_true` > 1, we assign to each target class the target probability 1 / `num_true` so that the target probabilities sum to 1 per-example. |