From 215d0c78c40bcaa5bec9bc795c07df55be928d2f Mon Sep 17 00:00:00 2001 From: Arthur Zucker Date: Wed, 4 Jan 2023 08:56:03 +0000 Subject: [PATCH 1/3] Remove random output --- .../roberta_prelayernorm/modeling_roberta_prelayernorm.py | 3 --- 1 file changed, 3 deletions(-) diff --git a/src/transformers/models/roberta_prelayernorm/modeling_roberta_prelayernorm.py b/src/transformers/models/roberta_prelayernorm/modeling_roberta_prelayernorm.py index d3aece98dc4f..e7a0128aa883 100644 --- a/src/transformers/models/roberta_prelayernorm/modeling_roberta_prelayernorm.py +++ b/src/transformers/models/roberta_prelayernorm/modeling_roberta_prelayernorm.py @@ -1078,9 +1078,6 @@ def set_output_embeddings(self, new_embeddings): checkpoint=_CHECKPOINT_FOR_DOC, output_type=MaskedLMOutput, config_class=_CONFIG_FOR_DOC, - mask="", - expected_output="' Paris'", - expected_loss=0.1, ) def forward( self, From 6c36965c060655eb87da884d413dc5823434e604 Mon Sep 17 00:00:00 2001 From: Arthur Zucker Date: Wed, 4 Jan 2023 08:59:39 +0000 Subject: [PATCH 2/3] remove values --- .../roberta_prelayernorm/modeling_roberta_prelayernorm.py | 6 ------ 1 file changed, 6 deletions(-) diff --git a/src/transformers/models/roberta_prelayernorm/modeling_roberta_prelayernorm.py b/src/transformers/models/roberta_prelayernorm/modeling_roberta_prelayernorm.py index e7a0128aa883..aa4fe48d8553 100644 --- a/src/transformers/models/roberta_prelayernorm/modeling_roberta_prelayernorm.py +++ b/src/transformers/models/roberta_prelayernorm/modeling_roberta_prelayernorm.py @@ -1196,8 +1196,6 @@ def __init__(self, config): checkpoint=_CHECKPOINT_FOR_DOC, output_type=SequenceClassifierOutput, config_class=_CONFIG_FOR_DOC, - expected_output="'optimism'", - expected_loss=0.08, ) # Copied from transformers.models.roberta.modeling_roberta.RobertaForSequenceClassification.forward with roberta->roberta_prelayernorm def forward( @@ -1397,8 +1395,6 @@ def __init__(self, config): checkpoint=_CHECKPOINT_FOR_DOC, output_type=TokenClassifierOutput, config_class=_CONFIG_FOR_DOC, - expected_output="['O', 'ORG', 'ORG', 'O', 'O', 'O', 'O', 'O', 'LOC', 'O', 'LOC', 'LOC']", - expected_loss=0.01, ) # Copied from transformers.models.roberta.modeling_roberta.RobertaForTokenClassification.forward with roberta->roberta_prelayernorm def forward( @@ -1504,8 +1500,6 @@ def __init__(self, config): checkpoint=_CHECKPOINT_FOR_DOC, output_type=QuestionAnsweringModelOutput, config_class=_CONFIG_FOR_DOC, - expected_output="' puppet'", - expected_loss=0.86, ) # Copied from transformers.models.roberta.modeling_roberta.RobertaForQuestionAnswering.forward with roberta->roberta_prelayernorm def forward( From 0f2e446154e74bfe3968efa3239cbf67f5d3d260 Mon Sep 17 00:00:00 2001 From: Arthur Zucker Date: Wed, 4 Jan 2023 09:07:29 +0000 Subject: [PATCH 3/3] fix copy statements --- .../roberta_prelayernorm/modeling_roberta_prelayernorm.py | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/src/transformers/models/roberta_prelayernorm/modeling_roberta_prelayernorm.py b/src/transformers/models/roberta_prelayernorm/modeling_roberta_prelayernorm.py index aa4fe48d8553..d3640e2232a6 100644 --- a/src/transformers/models/roberta_prelayernorm/modeling_roberta_prelayernorm.py +++ b/src/transformers/models/roberta_prelayernorm/modeling_roberta_prelayernorm.py @@ -1042,12 +1042,12 @@ def _reorder_cache(self, past, beam_idx): @add_start_docstrings( """RoBERTa-PreLayerNorm Model with a `language modeling` head on top.""", ROBERTA_PRELAYERNORM_START_DOCSTRING ) -# Copied from transformers.models.roberta.modeling_roberta.RobertaForMaskedLM with ROBERTA->ROBERTA_PRELAYERNORM,Roberta->RobertaPreLayerNorm,roberta->roberta_prelayernorm class RobertaPreLayerNormForMaskedLM(RobertaPreLayerNormPreTrainedModel): _keys_to_ignore_on_save = [r"lm_head.decoder.weight", r"lm_head.decoder.bias"] _keys_to_ignore_on_load_missing = [r"position_ids", r"lm_head.decoder.weight", r"lm_head.decoder.bias"] _keys_to_ignore_on_load_unexpected = [r"pooler"] + # Copied from transformers.models.roberta.modeling_roberta.RobertaForMaskedLM.__init__ with ROBERTA->ROBERTA_PRELAYERNORM,Roberta->RobertaPreLayerNorm,roberta->roberta_prelayernorm def __init__(self, config): super().__init__(config) @@ -1079,6 +1079,7 @@ def set_output_embeddings(self, new_embeddings): output_type=MaskedLMOutput, config_class=_CONFIG_FOR_DOC, ) + # Copied from transformers.models.roberta.modeling_roberta.RobertaForMaskedLM.forward with ROBERTA->ROBERTA_PRELAYERNORM,Roberta->RobertaPreLayerNorm,roberta->roberta_prelayernorm def forward( self, input_ids: Optional[torch.LongTensor] = None,