Matt 96d833b211 Return scalar losses instead of per-sample means (#18013) ...

* Return scalar losses instead of per-sample means

* Make loss shape (1,) instead of scalar

* Allow scalar losses in test_loss_computation

* Allow scalar losses in test_loss_computation

* Allow scalar losses in test_loss_computation

* Remove XLA loss function for RAG

2022-07-04 17:26:19 +01:00

..

benchmark

[Test refactor 1/5] Per-folder tests reorganization (#15725)

2022-02-23 15:46:28 -05:00

deepspeed

deprecate is_torch_bf16_available (#17738)

2022-06-20 08:40:11 -04:00

extended

Update self-push workflow (#17177)

2022-05-13 16:28:00 +02:00

fixtures

add a warning in SpmConverter for sentencepiece's model using the byte fallback feature (#16629)

2022-04-11 11:06:10 +02:00

generation

Update expected values in constrained beam search tests (#17887)

2022-06-28 08:53:53 +02:00

models

Return scalar losses instead of per-sample means (#18013)

2022-07-04 17:26:19 +01:00

onnx

add ONNX support for BLOOM (#17961)

2022-07-01 10:44:42 -04:00

optimization

[Test refactor 1/5] Per-folder tests reorganization (#15725)

2022-02-23 15:46:28 -05:00

pipelines

Restore original task in test_warning_logs (#17985)

2022-07-01 20:44:27 +02:00

sagemaker

Black preview (#17217)

2022-05-12 16:25:55 -04:00

tokenization

fix train_new_from_iterator in the case of byte-level tokenizers (#17549)

2022-06-08 15:30:41 +02:00

trainer

higher atol to avoid flaky trainer test failure (#17979)

2022-07-01 17:53:16 +02:00

utils

XLA train step fixes (#17973)

2022-07-01 19:11:14 +01:00

__init__.py

GPU text generation: mMoved the encoded_prompt to correct device

2020-01-06 15:11:12 +01:00

test_configuration_common.py

XLA train step fixes (#17973)

2022-07-01 19:11:14 +01:00

test_feature_extraction_common.py

Compute min_resolution in prepare_image_inputs (#17915)

2022-06-29 10:30:20 +02:00

test_modeling_common.py

Offload fixes (#17810)

2022-06-22 12:23:07 -04:00

test_modeling_flax_common.py

[Flax] Add remat (gradient checkpointing) (#17843)

2022-07-01 18:33:54 +01:00

test_modeling_tf_common.py

Return scalar losses instead of per-sample means (#18013)

2022-07-04 17:26:19 +01:00

test_sequence_feature_extraction_common.py

Some tests misusing assertTrue for comparisons fix (#16771)

2022-04-19 14:44:08 +02:00

test_tokenization_common.py

Fix properties of unset special tokens in non verbose mode (#17797)

2022-06-23 14:40:13 +02:00