wip

2020-12-11 14:23:07 -08:00 · 2020-12-11 14:20:02 -08:00 · 2020-12-11 14:17:17 -08:00 · 2020-12-11 13:53:53 -08:00 · 2020-12-11 13:53:12 -08:00 · 2020-12-11 12:58:01 -08:00
8 changed files with 89 additions and 400 deletions
@@ -3,47 +3,6 @@ orbs:
    gcp-gke: circleci/gcp-gke@1.0.4
    go: circleci/go@1.3.0

-commands:
-  skip-job-on-doc-only-changes:
-    description: "Do not continue this job and exit with success for PRs with only doc changes"
-    steps:
-
-      - run:
-          name: docs-only changes skip check
-          command: |
-            if test -n "$CIRCLE_PR_NUMBER"
-            then
-                echo $CIRCLE_PR_NUMBER
-                resp=$(curl -Ls https://api.github.com/repos/huggingface/transformers/pulls/${CIRCLE_PR_NUMBER})
-                user=$(jq -r .user.login \<<< $resp)   # PR creator username
-                head_ref=$(jq -r .head.ref \<<< $resp) # PR user's branch name
-                echo head_ref=$head_ref, user=$user
-            fi
-            
-            if test -n "$user" && test -n "$head_ref"
-            then
-                git clone https://github.com/$user/transformers user-clone
-                cd user-clone
-                git checkout $head_ref
-                fork_point_sha=$(git merge-base --fork-point master)
-                cd -
-            fi
-
-            if test -n "$fork_point_sha" && test -n "$(git diff --name-only $fork_point_sha)"
-            then
-                git --no-pager diff --name-only $fork_point_sha
-                if git diff --name-only $fork_point_sha | egrep -qv '\.(md|rst)$'
-                then
-                    echo "Non-docs were modified in this PR, proceeding normally"
-                else
-                    echo "Only docs were modified in this PR, quitting this job"
-                    # enable skipping once we get this sorted out
-                    # circleci step halt
-                fi
-            else
-                echo "Not enough data to perform a skipping check - continuing the job"
-            fi
-

 # TPU REFERENCES
 references:
@@ -104,278 +63,6 @@ references:


 jobs:
-    run_tests_torch_and_tf:
-        working_directory: ~/transformers
-        docker:
-            - image: circleci/python:3.6
-        environment:
-            OMP_NUM_THREADS: 1
-        resource_class: xlarge
-        parallelism: 1
-        steps:
-            - checkout
-            - skip-job-on-doc-only-changes
-            - restore_cache:
-                  keys:
-                      - v0.4-torch_and_tf-{{ checksum "setup.py" }}
-                      - v0.4-{{ checksum "setup.py" }}
-            - run: pip install --upgrade pip
-            - run: pip install .[sklearn,tf-cpu,torch,testing,sentencepiece]
-            - save_cache:
-                key: v0.4-{{ checksum "setup.py" }}
-                paths:
-                    - '~/.cache/pip'
-            - run: RUN_PT_TF_CROSS_TESTS=1 python -m pytest -n 8 --dist=loadfile -rA -s --make-reports=tests_torch_and_tf ./tests/ -m is_pt_tf_cross_test --durations=0 | tee tests_output.txt
-            - store_artifacts:
-                  path: ~/transformers/tests_output.txt
-            - store_artifacts:
-                  path: ~/transformers/reports
-
-    run_tests_torch:
-        working_directory: ~/transformers
-        docker:
-            - image: circleci/python:3.7
-        environment:
-            OMP_NUM_THREADS: 1
-        resource_class: xlarge
-        parallelism: 1
-        steps:
-            - checkout
-            - skip-job-on-doc-only-changes
-            - restore_cache:
-                  keys:
-                      - v0.4-torch-{{ checksum "setup.py" }}
-                      - v0.4-{{ checksum "setup.py" }}
-            - run: pip install --upgrade pip
-            - run: pip install .[sklearn,torch,testing,sentencepiece]
-            - save_cache:
-                  key: v0.4-torch-{{ checksum "setup.py" }}
-                  paths:
-                      - '~/.cache/pip'
-            - run: python -m pytest -n 8 --dist=loadfile -s --make-reports=tests_torch ./tests/ | tee tests_output.txt
-            - store_artifacts:
-                  path: ~/transformers/tests_output.txt
-            - store_artifacts:
-                  path: ~/transformers/reports
-
-    run_tests_tf:
-        working_directory: ~/transformers
-        docker:
-            - image: circleci/python:3.7
-        environment:
-            OMP_NUM_THREADS: 1
-        resource_class: xlarge
-        parallelism: 1
-        steps:
-            - checkout
-            - skip-job-on-doc-only-changes
-            - restore_cache:
-                  keys:
-                      - v0.4-tf-{{ checksum "setup.py" }}
-                      - v0.4-{{ checksum "setup.py" }}
-            - run: pip install --upgrade pip
-            - run: pip install .[sklearn,tf-cpu,testing,sentencepiece]
-            - save_cache:
-                  key: v0.4-tf-{{ checksum "setup.py" }}
-                  paths:
-                      - '~/.cache/pip'
-            - run: python -m pytest -n 8 --dist=loadfile -rA -s --make-reports=tests_tf ./tests/ | tee tests_output.txt
-            - store_artifacts:
-                  path: ~/transformers/tests_output.txt
-            - store_artifacts:
-                  path: ~/transformers/reports
-
-    run_tests_flax:
-        working_directory: ~/transformers
-        docker:
-            - image: circleci/python:3.7
-        environment:
-            OMP_NUM_THREADS: 1
-        resource_class: xlarge
-        parallelism: 1
-        steps:
-            - checkout
-            - skip-job-on-doc-only-changes
-            - restore_cache:
-                keys:
-                    - v0.4-flax-{{ checksum "setup.py" }}
-                    - v0.4-{{ checksum "setup.py" }}
-            - run: pip install --upgrade pip
-            - run: sudo pip install .[flax,sklearn,torch,testing,sentencepiece]
-            - save_cache:
-                  key: v0.4-flax-{{ checksum "setup.py" }}
-                  paths:
-                      - '~/.cache/pip'
-            - run: python -m pytest -n 8 --dist=loadfile -rA -s --make-reports=tests_flax ./tests/ | tee tests_output.txt
-            - store_artifacts:
-                  path: ~/transformers/tests_output.txt
-            - store_artifacts:
-                  path: ~/transformers/reports
-
-    run_tests_pipelines_torch:
-        working_directory: ~/transformers
-        docker:
-            - image: circleci/python:3.7
-        environment:
-            OMP_NUM_THREADS: 1
-        resource_class: xlarge
-        parallelism: 1
-        steps:
-            - checkout
-            - skip-job-on-doc-only-changes
-            - restore_cache:
-                  keys:
-                      - v0.4-torch-{{ checksum "setup.py" }}
-                      - v0.4-{{ checksum "setup.py" }}
-            - run: pip install --upgrade pip
-            - run: pip install .[sklearn,torch,testing,sentencepiece]
-            - save_cache:
-                  key: v0.4-torch-{{ checksum "setup.py" }}
-                  paths:
-                      - '~/.cache/pip'
-            - run: RUN_PIPELINE_TESTS=1 python -m pytest -n 8 --dist=loadfile -rA -s --make-reports=tests_pipelines_torch -m is_pipeline_test ./tests/ | tee tests_output.txt
-            - store_artifacts:
-                  path: ~/transformers/tests_output.txt
-            - store_artifacts:
-                  path: ~/transformers/reports
-
-    run_tests_pipelines_tf:
-        working_directory: ~/transformers
-        docker:
-            - image: circleci/python:3.7
-        environment:
-            OMP_NUM_THREADS: 1
-        resource_class: xlarge
-        parallelism: 1
-        steps:
-            - checkout
-            - skip-job-on-doc-only-changes
-            - restore_cache:
-                  keys:
-                      - v0.4-tf-{{ checksum "setup.py" }}
-                      - v0.4-{{ checksum "setup.py" }}
-            - run: pip install --upgrade pip
-            - run: pip install .[sklearn,tf-cpu,testing,sentencepiece]
-            - save_cache:
-                  key: v0.4-tf-{{ checksum "setup.py" }}
-                  paths:
-                      - '~/.cache/pip'
-            - run: RUN_PIPELINE_TESTS=1 python -m pytest -n 8 --dist=loadfile -rA -s --make-reports=tests_pipelines_tf ./tests/ -m is_pipeline_test | tee tests_output.txt
-            - store_artifacts:
-                  path: ~/transformers/tests_output.txt
-            - store_artifacts:
-                  path: ~/transformers/reports
-
-    run_tests_custom_tokenizers:
-        working_directory: ~/transformers
-        docker:
-            - image: circleci/python:3.7
-        environment:
-            RUN_CUSTOM_TOKENIZERS: yes
-        steps:
-            - checkout
-            - skip-job-on-doc-only-changes
-            - restore_cache:
-                  keys:
-                      - v0.4-custom_tokenizers-{{ checksum "setup.py" }}
-                      - v0.4-{{ checksum "setup.py" }}
-            - run: pip install --upgrade pip
-            - run: pip install .[ja,testing,sentencepiece]
-            - run: python -m unidic download
-            - save_cache:
-                  key: v0.4-custom_tokenizers-{{ checksum "setup.py" }}
-                  paths:
-                      - '~/.cache/pip'
-            - run: python -m pytest -s --make-reports=tests_custom_tokenizers ./tests/test_tokenization_bert_japanese.py | tee tests_output.txt
-            - store_artifacts:
-                  path: ~/transformers/tests_output.txt
-            - store_artifacts:
-                  path: ~/transformers/reports
-
-    run_examples_torch:
-        working_directory: ~/transformers
-        docker:
-            - image: circleci/python:3.6
-        environment:
-            OMP_NUM_THREADS: 1
-        resource_class: xlarge
-        parallelism: 1
-        steps:
-            - checkout
-            - skip-job-on-doc-only-changes
-            - restore_cache:
-                  keys:
-                      - v0.4-torch_examples-{{ checksum "setup.py" }}
-                      - v0.4-{{ checksum "setup.py" }}
-            - run: pip install --upgrade pip
-            - run: pip install .[sklearn,torch,sentencepiece,testing]
-            - run: pip install -r examples/_tests_requirements.txt
-            - save_cache:
-                  key: v0.4-torch_examples-{{ checksum "setup.py" }}
-                  paths:
-                      - '~/.cache/pip'
-            - run: python -m pytest -n 8 --dist=loadfile -s --make-reports=examples_torch ./examples/ | tee examples_output.txt
-            - store_artifacts:
-                  path: ~/transformers/examples_output.txt
-            - store_artifacts:
-                  path: ~/transformers/reports
-
-    run_tests_git_lfs:
-        working_directory: ~/transformers
-        docker:
-            - image: circleci/python:3.7
-        resource_class: xlarge
-        parallelism: 1
-        steps:
-            - checkout
-            - run: sudo apt-get install git-lfs
-            - run: |
-                git config --global user.email "ci@dummy.com"
-                git config --global user.name "ci"
-            - run: pip install --upgrade pip
-            - run: pip install .[testing]
-            - run: RUN_GIT_LFS_TESTS=1 python -m pytest -sv ./tests/test_hf_api.py -k "HfLargefilesTest"
-
-    build_doc:
-        working_directory: ~/transformers
-        docker:
-            - image: circleci/python:3.6
-        steps:
-            - checkout
-            - restore_cache:
-                  keys:
-                      - v0.4-build_doc-{{ checksum "setup.py" }}
-                      - v0.4-{{ checksum "setup.py" }}
-            - run: pip install --upgrade pip
-            - run: pip install ."[all, docs]"
-            - save_cache:
-                  key: v0.4-build_doc-{{ checksum "setup.py" }}
-                  paths:
-                      - '~/.cache/pip'
-            - run: cd docs && make html SPHINXOPTS="-W"
-            - store_artifacts:
-                path: ./docs/_build
-
-    deploy_doc:
-        working_directory: ~/transformers
-        docker:
-            - image: circleci/python:3.6
-        steps:
-            - add_ssh_keys:
-                fingerprints:
-                    - "5b:7a:95:18:07:8c:aa:76:4c:60:35:88:ad:60:56:71"
-            - checkout
-            - restore_cache:
-                  keys:
-                      - v0.4-deploy_doc-{{ checksum "setup.py" }}
-                      - v0.4-{{ checksum "setup.py" }}
-            - run: pip install ."[all,docs]"
-            - save_cache:
-                  key: v0.4-deploy_doc-{{ checksum "setup.py" }}
-                  paths:
-                      - '~/.cache/pip'
-            - run: ./.circleci/deploy.sh
-
    check_code_quality:
        working_directory: ~/transformers
        docker:
@@ -389,20 +76,22 @@ jobs:
                      - v0.4-code_quality-{{ checksum "setup.py" }}
                      - v0.4-{{ checksum "setup.py" }}
            - run: pip install --upgrade pip
-            - run: pip install isort
-            - run: pip install .[all,quality]
-            - save_cache:
-                  key: v0.4-code_quality-{{ checksum "setup.py" }}
-                  paths:
-                      - '~/.cache/pip'
-            - run: black --check examples tests src utils
-            - run: isort --check-only examples tests src utils
-            - run: flake8 examples tests src utils
-            - run: python utils/style_doc.py src/transformers docs/source --max_len 119 --check_only
-            - run: python utils/check_copies.py
-            - run: python utils/check_table.py
-            - run: python utils/check_dummies.py
-            - run: python utils/check_repo.py
+            - run:
+                command: |
+                    set +e
+                    echo "my experiment is here"
+                    # emulate failure
+                    false
+                    echo "it can safely fail"
+                    some non existing command
+                    echo "should still reach here"
+                    some non existing command again
+                    echo "should still reach here too"
+                    false
+            - run:
+                when: always
+                command: |
+                    echo "forcing success for this experiment"

    check_repository_consistency:
        working_directory: ~/transformers
@@ -414,37 +103,7 @@ jobs:
            - checkout
            - run: pip install requests
            - run: python ./utils/link_tester.py
-
-# TPU JOBS
-    run_examples_tpu:
-        docker:
-            - image: circleci/python:3.6
-        environment:
-            OMP_NUM_THREADS: 1
-        resource_class: xlarge
-        parallelism: 1
-        steps:
-            - checkout
-            - go/install
-            - *checkout_ml_testing
-            - gcp-gke/install
-            - gcp-gke/update-kubeconfig-with-credentials:
-                  cluster: $GKE_CLUSTER
-                  perform-login: true
-            - setup_remote_docker
-            - *build_push_docker
-            - *deploy_cluster
-
-    cleanup-gke-jobs:
-        docker:
-            - image: circleci/python:3.6
-        steps:
-            - gcp-gke/install
-            - gcp-gke/update-kubeconfig-with-credentials:
-                  cluster: $GKE_CLUSTER
-                  perform-login: true
-            - *delete_gke_jobs
-
+            
 workflow_filters: &workflow_filters
    filters:
        branches:
@@ -456,26 +115,5 @@ workflows:
        jobs:
            - check_code_quality
            - check_repository_consistency
-            - run_examples_torch
-            - run_tests_custom_tokenizers
-            - run_tests_torch_and_tf
-            - run_tests_torch
-            - run_tests_tf
-            - run_tests_flax
-            - run_tests_pipelines_torch
-            - run_tests_pipelines_tf
-            - run_tests_git_lfs
-            - build_doc
-            - deploy_doc: *workflow_filters
-    tpu_testing_jobs:
-        triggers:
-            - schedule:
-                # Set to run at the first minute of every hour.
-                cron: "0 8 * * *"
-                filters:
-                    branches:
-                        only:
-                            - master
-        jobs:
-            - cleanup-gke-jobs
-            - run_examples_tpu
+
+            
@@ -47,7 +47,7 @@ nbclient==0.5.0
 nbconvert==6.0.1
 nbformat==5.0.7
 nest-asyncio==1.4.0
-notebook==6.1.4
+notebook==6.1.5
 numpy==1.19.2
 opencv-python==4.4.0.42
 packaging==20.3
@@ -1886,7 +1886,7 @@ class ProphetNetForCausalLM(ProphetNetPreTrainedModel):
        config = copy.deepcopy(config)
        config.is_decoder = True
        config.is_encoder_decoder = False
-        self.decoder = ProphetNetDecoder(config)
+        self.prophetnet = ProphetNetDecoderWrapper(config)

        self.padding_idx = config.pad_token_id
        self.disable_ngram_loss = config.disable_ngram_loss
@@ -1896,10 +1896,10 @@ class ProphetNetForCausalLM(ProphetNetPreTrainedModel):
        self.init_weights()

    def get_input_embeddings(self):
-        return self.decoder.word_embeddings
+        return self.prophetnet.decoder.word_embeddings

    def set_input_embeddings(self, value):
-        self.decoder.word_embeddings = value
+        self.prophetnet.decoder.word_embeddings = value

    def get_output_embeddings(self):
        return self.lm_head
@@ -1907,6 +1907,12 @@ class ProphetNetForCausalLM(ProphetNetPreTrainedModel):
    def set_output_embeddings(self, new_embeddings):
        self.lm_head = new_embeddings

+    def set_decoder(self, decoder):
+        self.prophetnet.decoder = decoder
+
+    def get_decoder(self):
+        return self.prophetnet.decoder
+
    @add_start_docstrings_to_model_forward(PROPHETNET_STANDALONE_INPUTS_DOCSTRING)
    @replace_return_docstrings(output_type=ProphetNetDecoderLMOutput, config_class=_CONFIG_FOR_DOC)
    def forward(
@@ -1956,7 +1962,7 @@ class ProphetNetForCausalLM(ProphetNetPreTrainedModel):
            >>> import torch

            >>> tokenizer = ProphetNetTokenizer.from_pretrained('microsoft/prophetnet-large-uncased')
-            >>> model = ProphetNetForCausalLM.from_pretrained('patrickvonplaten/prophetnet-decoder-clm-large-uncased')
+            >>> model = ProphetNetForCausalLM.from_pretrained('microsoft/prophetnet-large-uncased')
            >>> assert model.config.is_decoder, f"{model.__class__} has to be configured as a decoder."
            >>> inputs = tokenizer("Hello, my dog is cute", return_tensors="pt")
            >>> outputs = model(**inputs)
@@ -1969,7 +1975,7 @@ class ProphetNetForCausalLM(ProphetNetPreTrainedModel):

            >>> tokenizer_enc = BertTokenizer.from_pretrained('bert-large-uncased')
            >>> tokenizer_dec = ProphetNetTokenizer.from_pretrained('microsoft/prophetnet-large-uncased')
-            >>> model = EncoderDecoderModel.from_encoder_decoder_pretrained("bert-large-uncased", "patrickvonplaten/prophetnet-decoder-clm-large-uncased")
+            >>> model = EncoderDecoderModel.from_encoder_decoder_pretrained("bert-large-uncased", "microsoft/prophetnet-large-uncased")

            >>> ARTICLE = (
            ... "the us state department said wednesday it had received no "
@@ -1985,7 +1991,7 @@ class ProphetNetForCausalLM(ProphetNetPreTrainedModel):
        return_dict = return_dict if return_dict is not None else self.config.use_return_dict

        # decoder outputs consists of (dec_features, layer_state, dec_hidden, dec_attn)
-        outputs = self.decoder(
+        outputs = self.prophetnet.decoder(
            input_ids=input_ids,
            attention_mask=attention_mask,
            encoder_hidden_states=encoder_hidden_states,
@@ -2086,8 +2092,16 @@ class ProphetNetForCausalLM(ProphetNetPreTrainedModel):
            reordered_past.append(layer_past_new)
        return reordered_past

-    def set_decoder(self, decoder):
-        self.decoder = decoder

-    def get_decoder(self):
-        return self.decoder
+class ProphetNetDecoderWrapper(ProphetNetPreTrainedModel):
+    """
+    This is a wrapper class, so that :class:`~transformers.ProphetNetForCausalLM` can correctly be loaded from
+    pretrained prophetnet classes.
+    """
+
+    def __init__(self, config):
+        super().__init__(config)
+        self.decoder = ProphetNetDecoder(config)
+
+    def forward(self, *args, **kwargs):
+        return self.decoder(*args, **kwargs)
@@ -136,7 +136,7 @@ class XLMProphetNetForCausalLM(ProphetNetForCausalLM):
        >>> import torch

        >>> tokenizer = XLMProphetNetTokenizer.from_pretrained('microsoft/xprophetnet-large-wiki100-cased')
-        >>> model = XLMProphetNetForCausalLM.from_pretrained('patrickvonplaten/xprophetnet-decoder-clm-large-uncased')
+        >>> model = XLMProphetNetForCausalLM.from_pretrained('microsoft/xprophetnet-large-wiki100-cased')
        >>> assert model.config.is_decoder, f"{model.__class__} has to be configured as a decoder."
        >>> inputs = tokenizer("Hello, my dog is cute", return_tensors="pt")
        >>> outputs = model(**inputs)
@@ -149,7 +149,7 @@ class XLMProphetNetForCausalLM(ProphetNetForCausalLM):

        >>> tokenizer_enc = XLMRobertaTokenizer.from_pretrained('xlm-roberta-large')
        >>> tokenizer_dec = XLMProphetNetTokenizer.from_pretrained('microsoft/xprophetnet-large-wiki100-cased')
-        >>> model = EncoderDecoderModel.from_encoder_decoder_pretrained("xlm-roberta-large", "patrickvonplaten/xprophetnet-decoder-clm-large-uncased")
+        >>> model = EncoderDecoderModel.from_encoder_decoder_pretrained("xlm-roberta-large", 'microsoft/xprophetnet-large-wiki100-cased')

        >>> ARTICLE = (
        ... "the us state department said wednesday it had received no "
@@ -2664,10 +2664,17 @@ class PreTrainedTokenizerBase(SpecialTokensMixin):
        # If we have PyTorch/TF/NumPy tensors/arrays as inputs, we cast them as python objects
        # and rebuild them afterwards if no return_tensors is specified
        # Note that we lose the specific device the tensor may be on for PyTorch
+
        first_element = encoded_inputs["input_ids"][0]
-        if isinstance(first_element, (list, tuple)) and first_element:
-            first_element = first_element[0]
-        if not isinstance(first_element, int):
+        if isinstance(first_element, (list, tuple)):
+            # first_element might be an empty list/tuple in some edge cases so we grab the first non empty element.
+            index = 0
+            while len(encoded_inputs["input_ids"][index]) == 0:
+                index += 1
+            if index < len(encoded_inputs["input_ids"]):
+                first_element = encoded_inputs["input_ids"][index][0]
+        # At this state, if `first_element` is still a list/tuple, it's an empty one so there is nothing to do.
+        if not isinstance(first_element, (int, list, tuple)):
            if is_tf_available() and isinstance(first_element, tf.Tensor):
                return_tensors = "tf" if return_tensors is None else return_tensors
            elif is_torch_available() and isinstance(first_element, torch.Tensor):
@@ -802,9 +802,7 @@ class ProphetNetEncoderDecoderModelTest(EncoderDecoderMixin, unittest.TestCase):
        }

    def get_pretrained_model(self):
-        return EncoderDecoderModel.from_encoder_decoder_pretrained(
-            "bert-large-uncased", "patrickvonplaten/prophetnet-decoder-clm-large-uncased"
-        )
+        return EncoderDecoderModel.from_encoder_decoder_pretrained("bert-large-uncased", "prophetnet-large-uncased")

    def test_encoder_decoder_model_shared_weights(self):
        pass
@@ -38,6 +38,7 @@ if is_torch_available():
        ProphetNetModel,
        ProphetNetTokenizer,
    )
+    from transformers.modeling_outputs import BaseModelOutput


 class ProphetNetModelTester:
@@ -467,6 +468,31 @@ class ProphetNetModelTester:
            )
        )

+    def check_causal_lm_from_pretrained(
+        self, config, input_ids, decoder_input_ids, attention_mask, decoder_attention_mask, *args
+    ):
+        model = ProphetNetForConditionalGeneration(config).to(torch_device).eval()
+
+        with tempfile.TemporaryDirectory() as tmp_dirname:
+            model.save_pretrained(tmp_dirname)
+            decoder = ProphetNetForCausalLM.from_pretrained(tmp_dirname).to(torch_device)
+
+        encoder_hidden_states = model.prophetnet.encoder(input_ids).last_hidden_state
+
+        model_outputs = model(
+            encoder_outputs=BaseModelOutput(last_hidden_state=encoder_hidden_states),
+            decoder_input_ids=decoder_input_ids,
+        )
+        dec_outputs = decoder(encoder_hidden_states=encoder_hidden_states, input_ids=decoder_input_ids)
+
+        self.parent.assertTrue(
+            torch.allclose(
+                model_outputs.logits[0, :5],
+                dec_outputs.logits[0, :5],
+                atol=1e-3,
+            )
+        )
+
    def prepare_config_and_inputs_for_common(self):
        config_and_inputs = self.prepare_config_and_inputs()
        (
@@ -898,6 +924,10 @@ class ProphetNetModelTest(ModelTesterMixin, GenerationTesterMixin, unittest.Test

        self.assertFalse(config.add_cross_attention)

+    def test_causal_lm_from_pretrained(self):
+        config_and_inputs = self.model_tester.prepare_config_and_inputs()
+        self.model_tester.check_causal_lm_from_pretrained(*config_and_inputs)
+
    @unittest.skipIf(torch_device == "cpu", "Cant do half precision")
    def test_fp16_forward(self):
        config_and_inputs = self.model_tester.prepare_config_and_inputs()
@@ -34,6 +34,7 @@ IGNORE_NON_TESTED = [
    "BertLMHeadModel",  # Needs to be setup as decoder.
    "DPREncoder",  # Building part of bigger (tested) model.
    "DPRSpanPredictor",  # Building part of bigger (tested) model.
+    "ProphetNetDecoderWrapper",  # Building part of bigger (tested) model.
    "ReformerForMaskedLM",  # Needs to be setup as decoder.
    "T5Stack",  # Building part of bigger (tested) model.
    "TFDPREncoder",  # Building part of bigger (tested) model.
@@ -74,6 +75,7 @@ IGNORE_NON_AUTO_CONFIGURED = [
    "OpenAIGPTDoubleHeadsModel",
    "ProphetNetDecoder",
    "ProphetNetEncoder",
+    "ProphetNetDecoderWrapper",
    "RagModel",
    "RagSequenceForGeneration",
    "RagTokenForGeneration",
Author	SHA1	Message	Date
Stas Bekman	5ea787a2d8	wip	2020-12-11 14:23:07 -08:00
Stas Bekman	a019424ff0	wip	2020-12-11 14:20:02 -08:00
Stas Bekman	49f7744aa3	wip	2020-12-11 14:17:17 -08:00
Stas Bekman	5455191606	shorten job	2020-12-11 13:53:53 -08:00
Stas Bekman	8507bb7352	set +e	2020-12-11 13:53:12 -08:00
Stas Bekman	f339c16a38	wip	2020-12-11 12:58:01 -08:00
Stas Bekman	40a6227fe1	workflows	2020-12-11 12:56:33 -08:00
Stas Bekman	09cd2464db	workflow restore	2020-12-11 12:55:15 -08:00
Stas Bekman	6a2d0b4e50	test with just 2 jobs	2020-12-11 12:53:54 -08:00
Stas Bekman	eacbd879c9	[wip] [ci] experiment for documentation please ignore	2020-12-11 12:49:31 -08:00
Patrick von Platen	9cc9f4122e	Make ProphetNetModel really compatible with EncoderDecoder (#9033 ) * improve * finish * upload model * fix lm head * fix test	2020-12-11 16:59:54 +01:00
dependabot[bot]	24f6cdeab6	Bump notebook in /examples/research_projects/movement-pruning/lxmert (#9062 ) Bumps [notebook](https://github.com/jupyter/jupyterhub) from 6.1.4 to 6.1.5. - [Release notes](https://github.com/jupyter/jupyterhub/releases) - [Changelog](https://github.com/jupyterhub/jupyterhub/blob/master/CHECKLIST-Release.md) - [Commits](https://github.com/jupyter/jupyterhub/commits) Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>	2020-12-11 10:32:43 -05:00
Lysandre Debut	91fa707217	Remove docs only check (#9065 )	2020-12-11 10:27:31 -05:00
Sylvain Gugger	70527ba694	Fix PreTrainedTokenizer.pad when first inputs are empty (#9018 ) * Fix PreTrainedTokenizer.pad when first inputs are empty * Handle empty inputs case	2020-12-11 10:25:00 -05:00