Ihor Indyk

> I don't see where the attributes prefetch_buffer_size and parser_num_threads documented in TensorFlowDatasetOptions are being used. Perhaps this is intentional? They are not used in that particular util, but are...

In tfx 1.13 we [introduced](https://github.com/tensorflow/tfx-bsl/commit/ce2e12b4a78ecced09e0e5e30e34d310656166d7) a new batching mode that tries to deserialize data in batches of ~ 100MB. It can be enabled with [tfxio_use_byte_size_batching](https://github.com/tensorflow/tfx-bsl/blob/b566c3970bc36d94fbf27e989df510d1f81e7ce0/tfx_bsl/coders/batch_util.py#L35) flag. Could you try updating...

Depending on how exactly you use transform and BulkInferrer you may also be able to set data (tfxio) source [batch size](https://github.com/tensorflow/tfx-bsl/blob/b566c3970bc36d94fbf27e989df510d1f81e7ce0/tfx_bsl/tfxio/tfxio.py#L183). Or, if you use the instance dict format with...

Yes, you're right, the component itself does not expose the parameter. Even if we were to add it, it would be available at an even later tfx version than the...

You could try creating a [custom component](https://www.tensorflow.org/tfx/guide/custom_component) based on transform that overrides the parameter, but that may be pretty involving.

since the OOM happens when applying the model and setting the `tfxio_use_byte_size_batching value` did not help it could be the case that input batch is small enough (batching happens on...