Added support for Tapas Model (#520)

* Added support for Tapas Model * Added support for Tapas Model * reformatted files with black * Update tests/bettertransformer/test_bettertransformer_encoder.py test_better_encoder Co-authored-by: Michael Benayoun <[email protected]> * Update optimum/bettertransformer/models/encoder_models.py Co-authored-by: Michael Benayoun <[email protected]> * Update tests/bettertransformer/test_bettertransformer_encoder.py Co-authored-by: Sylvain Gugger <[email protected]> * Styled optimum files * Update optimum/bettertransformer/models/encoder_models.py Co-authored-by: fxmarty <[email protected]> * Update optimum/bettertransformer/models/encoder_models.py Call super() in the _init() to inherit from BetterTransformerBaseLayer Co-authored-by: fxmarty <[email protected]> * Update tests/bettertransformer/test_bettertransformer_encoder.py Co-authored-by: Younes Belkada <[email protected]> * Moved Tapas Encoder model to Encoder * change mapping in __init_.py * deleted * Update __init__.py * Update __init__.py * refactor doc + remove class + styling Co-authored-by: Michael Benayoun <[email protected]> Co-authored-by: Sylvain Gugger <[email protected]> Co-authored-by: fxmarty <[email protected]> Co-authored-by: Younes Belkada <[email protected]> Co-authored-by: younesbelkada <[email protected]>
huggingface · Nov 30, 2022 · d33801d · d33801d
1 parent 6ee424b
commit d33801d
Show file tree

Hide file tree

Showing 3 changed files with 4 additions and 1 deletion.
diff --git a/docs/source/bettertransformer/overview.mdx b/docs/source/bettertransformer/overview.mdx
@@ -14,7 +14,6 @@ specific language governing permissions and limitations under the License.
 
 🤗 Optimum provides an integration with `BetterTransformer`, a stable API from PyTorch to benefit from interesting speedups on CPU & GPU through sparsity and fused kernels.
 
-
 ## Quickstart
 
 Since its 1.13 version, PyTorch released the stable version of `BetterTransformer` in its library. You can benefit from interesting speedup on most consumer-type devices, including CPUs, older and newer versions of NIVIDIA GPUs.
@@ -23,6 +22,7 @@ You can now use this feature in 🤗 Optimum together with Transformers and use
 ### Supported models
 
 The list of supported model below:
+
 - [AlBERT](https://arxiv.org/abs/1909.11942)
 - [BART](https://arxiv.org/abs/1910.13461)
 - [BERT](https://arxiv.org/abs/1810.04805)
@@ -41,6 +41,7 @@ The list of supported model below:
 - [M2M100](https://arxiv.org/abs/2010.11125)
 - [RoBERTa](https://arxiv.org/abs/1907.11692)
 - [Splinter](https://arxiv.org/abs/2101.00438)
+- [Tapas](https://arxiv.org/abs/2211.06550)
 - [ViLT](https://arxiv.org/abs/2102.03334)
 - [ViT](https://arxiv.org/abs/2010.11929)
 - [ViT-MAE](https://arxiv.org/abs/2111.06377)

diff --git a/optimum/bettertransformer/models/__init__.py b/optimum/bettertransformer/models/__init__.py
@@ -29,6 +29,7 @@
 
 BETTER_TRANFORMER_LAYERS_MAPPING_DICT = {
     # Bert Family
+    "TapasLayer": BertLayerBetterTransformer,
     "BertLayer": BertLayerBetterTransformer,
     "ElectraLayer": BertLayerBetterTransformer,
     "Data2VecTextLayer": BertLayerBetterTransformer,

diff --git a/tests/bettertransformer/test_bettertransformer_encoder.py b/tests/bettertransformer/test_bettertransformer_encoder.py
@@ -44,6 +44,7 @@
     "hf-internal-testing/tiny-random-MarkupLMModel",
     "hf-internal-testing/tiny-random-BertModel",
     "ybelkada/random-tiny-BertGenerationModel",
+    "hf-internal-testing/tiny-random-TapasModel",
 ]
 
 ALL_ENCODER_DECODER_MODELS_TO_TEST = [