SUMIN/HuggingFace_transformer
1
0
Fork 0
Code Issues Pull Requests Actions 7 Packages Projects Releases Wiki Activity

Remove ConversationalPipeline and Conversation object (#31165)

Browse Source 
* Remove ConversationalPipeline and Conversation object, as they have been deprecated for some time and are due for removal

* Update not-doctested.txt

* Fix JA and ZH docs

* Fix JA and ZH docs some more

* Fix JA and ZH docs some more
This commit is contained in:
Matt
2024-06-07 17:50:18 +01:00
committed by GitHub
parent 3a10058201
commit 065729a692
46 changed files with 30 additions and 914 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
tests/models/bart/test_modeling_bart.py
View File

@@ -430,7 +430,6 @@ class BartModelTest(ModelTesterMixin, GenerationTesterMixin, PipelineTesterMixin
all_generative_model_classes = (BartForConditionalGeneration,) if is_torch_available() else ()
pipeline_model_mapping = (
{
"conversational": BartForConditionalGeneration,
"feature-extraction": BartModel,
"fill-mask": BartForConditionalGeneration,
"question-answering": BartForQuestionAnswering,
@@ -513,10 +512,6 @@ class BartModelTest(ModelTesterMixin, GenerationTesterMixin, PipelineTesterMixin
model.generate(input_ids, attention_mask=attention_mask)
model.generate(num_beams=4, do_sample=True, early_stopping=False, num_return_sequences=3)
@unittest.skip("Does not support conversations.")
def test_pipeline_conversational(self):
pass
def assert_tensors_close(a, b, atol=1e-12, prefix=""):
"""If tensors have different shapes, different values or a and b are not both tensors, raise a nice Assertion error."""

5
tests/models/bart/test_modeling_tf_bart.py
View File

@@ -198,7 +198,6 @@ class TFBartModelTest(TFModelTesterMixin, TFCoreModelTesterMixin, PipelineTester
all_generative_model_classes = (TFBartForConditionalGeneration,) if is_tf_available() else ()
pipeline_model_mapping = (
{
"conversational": TFBartForConditionalGeneration,
"feature-extraction": TFBartModel,
"summarization": TFBartForConditionalGeneration,
"text-classification": TFBartForSequenceClassification,
@@ -343,10 +342,6 @@ class TFBartModelTest(TFModelTesterMixin, TFCoreModelTesterMixin, PipelineTester
# check that the output for the restored model is the same
self.assert_outputs_same(restored_model_outputs, outputs)
@unittest.skip("Does not support conversations.")
def test_pipeline_conversational(self):
pass
def _long_tensor(tok_lst):
return tf.constant(tok_lst, dtype=tf.int32)

1
tests/models/bigbird_pegasus/test_modeling_bigbird_pegasus.py
View File

@@ -253,7 +253,6 @@ class BigBirdPegasusModelTest(ModelTesterMixin, GenerationTesterMixin, PipelineT
all_generative_model_classes = (BigBirdPegasusForConditionalGeneration,) if is_torch_available() else ()
pipeline_model_mapping = (
{
"conversational": BigBirdPegasusForConditionalGeneration,
"feature-extraction": BigBirdPegasusModel,
"question-answering": BigBirdPegasusForQuestionAnswering,
"summarization": BigBirdPegasusForConditionalGeneration,

1
tests/models/blenderbot/test_modeling_blenderbot.py
View File

@@ -237,7 +237,6 @@ class BlenderbotModelTest(ModelTesterMixin, GenerationTesterMixin, PipelineTeste
all_generative_model_classes = (BlenderbotForConditionalGeneration,) if is_torch_available() else ()
pipeline_model_mapping = (
{
"conversational": BlenderbotForConditionalGeneration,
"feature-extraction": BlenderbotModel,
"summarization": BlenderbotForConditionalGeneration,
"text-generation": BlenderbotForCausalLM,

1
tests/models/blenderbot/test_modeling_tf_blenderbot.py
View File

@@ -183,7 +183,6 @@ class TFBlenderbotModelTest(TFModelTesterMixin, PipelineTesterMixin, unittest.Te
all_generative_model_classes = (TFBlenderbotForConditionalGeneration,) if is_tf_available() else ()
pipeline_model_mapping = (
{
"conversational": TFBlenderbotForConditionalGeneration,
"feature-extraction": TFBlenderbotModel,
"summarization": TFBlenderbotForConditionalGeneration,
"text2text-generation": TFBlenderbotForConditionalGeneration,

3
tests/models/blenderbot_small/test_modeling_blenderbot_small.py
View File

@@ -228,7 +228,6 @@ class BlenderbotSmallModelTest(ModelTesterMixin, GenerationTesterMixin, Pipeline
all_generative_model_classes = (BlenderbotSmallForConditionalGeneration,) if is_torch_available() else ()
pipeline_model_mapping = (
{
"conversational": BlenderbotSmallForConditionalGeneration,
"feature-extraction": BlenderbotSmallModel,
"summarization": BlenderbotSmallForConditionalGeneration,
"text-generation": BlenderbotSmallForCausalLM,
@@ -247,7 +246,7 @@ class BlenderbotSmallModelTest(ModelTesterMixin, GenerationTesterMixin, Pipeline
def is_pipeline_test_to_skip(
self, pipeline_test_casse_name, config_class, model_architecture, tokenizer_name, processor_name
):
return pipeline_test_casse_name in ("TextGenerationPipelineTests", "ConversationalPipelineTests")
return pipeline_test_casse_name == "TextGenerationPipelineTests"
def setUp(self):
self.model_tester = BlenderbotSmallModelTester(self)

2
tests/models/blenderbot_small/test_modeling_flax_blenderbot_small.py
View File

@@ -323,7 +323,7 @@ class FlaxBlenderbotSmallModelTest(FlaxModelTesterMixin, unittest.TestCase, Flax
def is_pipeline_test_to_skip(
self, pipeline_test_casse_name, config_class, model_architecture, tokenizer_name, processor_name
):
return pipeline_test_casse_name in ("TextGenerationPipelineTests", "ConversationalPipelineTests")
return pipeline_test_casse_name == "TextGenerationPipelineTests"
def setUp(self):
self.model_tester = FlaxBlenderbotSmallModelTester(self)

3
tests/models/blenderbot_small/test_modeling_tf_blenderbot_small.py
View File

@@ -185,7 +185,6 @@ class TFBlenderbotSmallModelTest(TFModelTesterMixin, PipelineTesterMixin, unitte
all_generative_model_classes = (TFBlenderbotSmallForConditionalGeneration,) if is_tf_available() else ()
pipeline_model_mapping = (
{
"conversational": TFBlenderbotSmallForConditionalGeneration,
"feature-extraction": TFBlenderbotSmallModel,
"summarization": TFBlenderbotSmallForConditionalGeneration,
"text2text-generation": TFBlenderbotSmallForConditionalGeneration,
@@ -201,7 +200,7 @@ class TFBlenderbotSmallModelTest(TFModelTesterMixin, PipelineTesterMixin, unitte
def is_pipeline_test_to_skip(
self, pipeline_test_casse_name, config_class, model_architecture, tokenizer_name, processor_name
):
return pipeline_test_casse_name in ("TextGenerationPipelineTests", "ConversationalPipelineTests")
return pipeline_test_casse_name == "TextGenerationPipelineTests"
def setUp(self):
self.model_tester = TFBlenderbotSmallModelTester(self)

1
tests/models/fsmt/test_modeling_fsmt.py
View File

@@ -166,7 +166,6 @@ class FSMTModelTest(ModelTesterMixin, GenerationTesterMixin, PipelineTesterMixin
all_generative_model_classes = (FSMTForConditionalGeneration,) if is_torch_available() else ()
pipeline_model_mapping = (
{
"conversational": FSMTForConditionalGeneration,
"feature-extraction": FSMTModel,
"summarization": FSMTForConditionalGeneration,
"text2text-generation": FSMTForConditionalGeneration,

1
tests/models/led/test_modeling_led.py
View File

@@ -284,7 +284,6 @@ class LEDModelTest(ModelTesterMixin, GenerationTesterMixin, PipelineTesterMixin,
all_generative_model_classes = (LEDForConditionalGeneration,) if is_torch_available() else ()
pipeline_model_mapping = (
{
"conversational": LEDForConditionalGeneration,
"feature-extraction": LEDModel,
"question-answering": LEDForQuestionAnswering,
"summarization": LEDForConditionalGeneration,

1
tests/models/led/test_modeling_tf_led.py
View File

@@ -197,7 +197,6 @@ class TFLEDModelTest(TFModelTesterMixin, PipelineTesterMixin, unittest.TestCase)
all_generative_model_classes = (TFLEDForConditionalGeneration,) if is_tf_available() else ()
pipeline_model_mapping = (
{
"conversational": TFLEDForConditionalGeneration,
"feature-extraction": TFLEDModel,
"summarization": TFLEDForConditionalGeneration,
"text2text-generation": TFLEDForConditionalGeneration,

1
tests/models/longt5/test_modeling_longt5.py
View File

@@ -504,7 +504,6 @@ class LongT5ModelTest(ModelTesterMixin, GenerationTesterMixin, PipelineTesterMix
all_generative_model_classes = (LongT5ForConditionalGeneration,) if is_torch_available() else ()
pipeline_model_mapping = (
{
"conversational": LongT5ForConditionalGeneration,
"feature-extraction": LongT5Model,
"summarization": LongT5ForConditionalGeneration,
"text2text-generation": LongT5ForConditionalGeneration,

1
tests/models/m2m_100/test_modeling_m2m_100.py
View File

@@ -243,7 +243,6 @@ class M2M100ModelTest(ModelTesterMixin, GenerationTesterMixin, PipelineTesterMix
all_generative_model_classes = (M2M100ForConditionalGeneration,) if is_torch_available() else ()
pipeline_model_mapping = (
{
"conversational": M2M100ForConditionalGeneration,
"feature-extraction": M2M100Model,
"summarization": M2M100ForConditionalGeneration,
"text2text-generation": M2M100ForConditionalGeneration,

4
tests/models/marian/test_modeling_flax_marian.py
View File

@@ -311,10 +311,6 @@ class FlaxMarianModelTest(FlaxModelTesterMixin, unittest.TestCase, FlaxGeneratio
outputs = model(input_ids)
self.assertIsNotNone(outputs)
@unittest.skip("Skipping for now, to fix @ArthurZ or @ydshieh")
def test_pipeline_conversational(self):
pass
@require_flax
@require_sentencepiece

5
tests/models/marian/test_modeling_marian.py
View File

@@ -248,7 +248,6 @@ class MarianModelTest(ModelTesterMixin, GenerationTesterMixin, PipelineTesterMix
all_generative_model_classes = (MarianMTModel,) if is_torch_available() else ()
pipeline_model_mapping = (
{
"conversational": MarianMTModel,
"feature-extraction": MarianModel,
"summarization": MarianMTModel,
"text-generation": MarianForCausalLM,
@@ -350,10 +349,6 @@ class MarianModelTest(ModelTesterMixin, GenerationTesterMixin, PipelineTesterMix
def test_tie_word_embeddings_decoder(self):
pass
@unittest.skip("Skipping for now, to fix @ArthurZ or @ydshieh")
def test_pipeline_conversational(self):
pass
@unittest.skip(
reason="This architecure seem to not compute gradients properly when using GC, check: https://github.com/huggingface/transformers/pull/27124"
)

5
tests/models/marian/test_modeling_tf_marian.py
View File

@@ -184,7 +184,6 @@ class TFMarianModelTest(TFModelTesterMixin, PipelineTesterMixin, unittest.TestCa
all_generative_model_classes = (TFMarianMTModel,) if is_tf_available() else ()
pipeline_model_mapping = (
{
"conversational": TFMarianMTModel,
"feature-extraction": TFMarianModel,
"summarization": TFMarianMTModel,
"text2text-generation": TFMarianMTModel,
@@ -208,10 +207,6 @@ class TFMarianModelTest(TFModelTesterMixin, PipelineTesterMixin, unittest.TestCa
config_and_inputs = self.model_tester.prepare_config_and_inputs_for_common()
self.model_tester.check_decoder_model_past_large_inputs(*config_and_inputs)
@unittest.skip("Skipping for now, to fix @ArthurZ or @ydshieh")
def test_pipeline_conversational(self):
pass
@require_tf
class AbstractMarianIntegrationTest(unittest.TestCase):

1
tests/models/mbart/test_modeling_mbart.py
View File

@@ -240,7 +240,6 @@ class MBartModelTest(ModelTesterMixin, GenerationTesterMixin, PipelineTesterMixi
all_generative_model_classes = (MBartForConditionalGeneration,) if is_torch_available() else ()
pipeline_model_mapping = (
{
"conversational": MBartForConditionalGeneration,
"feature-extraction": MBartModel,
"fill-mask": MBartForConditionalGeneration,
"question-answering": MBartForQuestionAnswering,

1
tests/models/mbart/test_modeling_tf_mbart.py
View File

@@ -161,7 +161,6 @@ class TFMBartModelTest(TFModelTesterMixin, PipelineTesterMixin, unittest.TestCas
all_generative_model_classes = (TFMBartForConditionalGeneration,) if is_tf_available() else ()
pipeline_model_mapping = (
{
"conversational": TFMBartForConditionalGeneration,
"feature-extraction": TFMBartModel,
"summarization": TFMBartForConditionalGeneration,
"text2text-generation": TFMBartForConditionalGeneration,

5
tests/models/mt5/test_modeling_mt5.py
View File

@@ -555,7 +555,6 @@ class MT5ModelTest(ModelTesterMixin, GenerationTesterMixin, PipelineTesterMixin,
all_generative_model_classes = (MT5ForConditionalGeneration,) if is_torch_available() else ()
pipeline_model_mapping = (
{
"conversational": MT5ForConditionalGeneration,
"feature-extraction": MT5Model,
"question-answering": MT5ForQuestionAnswering,
"summarization": MT5ForConditionalGeneration,
@@ -886,10 +885,6 @@ class MT5ModelTest(ModelTesterMixin, GenerationTesterMixin, PipelineTesterMixin,
attn_weights = out[attn_name] if attn_name == attention_names[0] else out[attn_name][-1]
self.assertEqual(sum([w.sum().item() for w in attn_weights]), 0.0)
@unittest.skip("Does not support conversations.")
def test_pipeline_conversational(self):
pass
# Copied from tests.models.t5.test_modeling_t5.T5EncoderOnlyModelTester with T5->MT5
class MT5EncoderOnlyModelTester:

1
tests/models/mvp/test_modeling_mvp.py
View File

@@ -421,7 +421,6 @@ class MvpModelTest(ModelTesterMixin, GenerationTesterMixin, PipelineTesterMixin,
all_generative_model_classes = (MvpForConditionalGeneration,) if is_torch_available() else ()
pipeline_model_mapping = (
{
"conversational": MvpForConditionalGeneration,
"feature-extraction": MvpModel,
"fill-mask": MvpForConditionalGeneration,
"question-answering": MvpForQuestionAnswering,

1
tests/models/nllb_moe/test_modeling_nllb_moe.py
View File

@@ -250,7 +250,6 @@ class NllbMoeModelTest(ModelTesterMixin, GenerationTesterMixin, PipelineTesterMi
all_generative_model_classes = (NllbMoeForConditionalGeneration,) if is_torch_available() else ()
pipeline_model_mapping = (
{
"conversational": NllbMoeForConditionalGeneration,
"feature-extraction": NllbMoeModel,
"summarization": NllbMoeForConditionalGeneration,
"text2text-generation": NllbMoeForConditionalGeneration,

1
tests/models/pegasus/test_modeling_pegasus.py
View File

@@ -246,7 +246,6 @@ class PegasusModelTest(ModelTesterMixin, GenerationTesterMixin, PipelineTesterMi
all_generative_model_classes = (PegasusForConditionalGeneration,) if is_torch_available() else ()
pipeline_model_mapping = (
{
"conversational": PegasusForConditionalGeneration,
"feature-extraction": PegasusModel,
"summarization": PegasusForConditionalGeneration,
"text-generation": PegasusForCausalLM,

1
tests/models/pegasus/test_modeling_tf_pegasus.py
View File

@@ -182,7 +182,6 @@ class TFPegasusModelTest(TFModelTesterMixin, PipelineTesterMixin, unittest.TestC
all_generative_model_classes = (TFPegasusForConditionalGeneration,) if is_tf_available() else ()
pipeline_model_mapping = (
{
"conversational": TFPegasusForConditionalGeneration,
"feature-extraction": TFPegasusModel,
"summarization": TFPegasusForConditionalGeneration,
"text2text-generation": TFPegasusForConditionalGeneration,

1
tests/models/pegasus_x/test_modeling_pegasus_x.py
View File

@@ -206,7 +206,6 @@ class PegasusXModelTest(ModelTesterMixin, GenerationTesterMixin, PipelineTesterM
all_generative_model_classes = (PegasusXForConditionalGeneration,) if is_torch_available() else ()
pipeline_model_mapping = (
{
"conversational": PegasusXForConditionalGeneration,
"feature-extraction": PegasusXModel,
"summarization": PegasusXForConditionalGeneration,
"text2text-generation": PegasusXForConditionalGeneration,

1
tests/models/plbart/test_modeling_plbart.py
View File

@@ -227,7 +227,6 @@ class PLBartModelTest(ModelTesterMixin, GenerationTesterMixin, PipelineTesterMix
all_generative_model_classes = (PLBartForConditionalGeneration,) if is_torch_available() else ()
pipeline_model_mapping = (
{
"conversational": PLBartForConditionalGeneration,
"feature-extraction": PLBartModel,
"summarization": PLBartForConditionalGeneration,
"text-classification": PLBartForSequenceClassification,

1
tests/models/prophetnet/test_modeling_prophetnet.py
View File

@@ -891,7 +891,6 @@ class ProphetNetModelTest(ModelTesterMixin, GenerationTesterMixin, PipelineTeste
all_generative_model_classes = (ProphetNetForConditionalGeneration,) if is_torch_available() else ()
pipeline_model_mapping = (
{
"conversational": ProphetNetForConditionalGeneration,
"feature-extraction": ProphetNetModel,
"summarization": ProphetNetForConditionalGeneration,
"text-generation": ProphetNetForCausalLM,

1
tests/models/seamless_m4t/test_modeling_seamless_m4t.py
View File

@@ -645,7 +645,6 @@ class SeamlessM4TModelWithTextInputTest(
pipeline_model_mapping = (
{
"automatic-speech-recognition": SeamlessM4TForSpeechToText,
"conversational": SeamlessM4TForTextToText,
"feature-extraction": SeamlessM4TModel,
"summarization": SeamlessM4TForTextToText,
"text-to-audio": SeamlessM4TForTextToSpeech,

1
tests/models/switch_transformers/test_modeling_switch_transformers.py
View File

@@ -559,7 +559,6 @@ class SwitchTransformersModelTest(ModelTesterMixin, GenerationTesterMixin, Pipel
all_generative_model_classes = (SwitchTransformersForConditionalGeneration,) if is_torch_available() else ()
pipeline_model_mapping = (
{
"conversational": SwitchTransformersForConditionalGeneration,
"feature-extraction": SwitchTransformersModel,
"summarization": SwitchTransformersForConditionalGeneration,
"text2text-generation": SwitchTransformersForConditionalGeneration,

5
tests/models/t5/test_modeling_t5.py
View File

@@ -558,7 +558,6 @@ class T5ModelTest(ModelTesterMixin, GenerationTesterMixin, PipelineTesterMixin,
all_generative_model_classes = (T5ForConditionalGeneration,) if is_torch_available() else ()
pipeline_model_mapping = (
{
"conversational": T5ForConditionalGeneration,
"feature-extraction": T5Model,
"question-answering": T5ForQuestionAnswering,
"summarization": T5ForConditionalGeneration,
@@ -889,10 +888,6 @@ class T5ModelTest(ModelTesterMixin, GenerationTesterMixin, PipelineTesterMixin,
attn_weights = out[attn_name] if attn_name == attention_names[0] else out[attn_name][-1]
self.assertEqual(sum([w.sum().item() for w in attn_weights]), 0.0)
@unittest.skip("Does not support conversations.")
def test_pipeline_conversational(self):
pass
class T5EncoderOnlyModelTester:
def __init__(

9
tests/models/t5/test_modeling_tf_t5.py
View File

@@ -248,7 +248,6 @@ class TFT5ModelTest(TFModelTesterMixin, PipelineTesterMixin, unittest.TestCase):
all_generative_model_classes = (TFT5ForConditionalGeneration,) if is_tf_available() else ()
pipeline_model_mapping = (
{
"conversational": TFT5ForConditionalGeneration,
"feature-extraction": TFT5Model,
"summarization": TFT5ForConditionalGeneration,
"text2text-generation": TFT5ForConditionalGeneration,
@@ -314,10 +313,6 @@ class TFT5ModelTest(TFModelTesterMixin, PipelineTesterMixin, unittest.TestCase):
def test_keras_save_load(self):
pass
@unittest.skip("Does not support conversations.")
def test_pipeline_conversational(self):
pass
class TFT5EncoderOnlyModelTester:
def __init__(
@@ -611,10 +606,6 @@ class TFT5GenerationIntegrationTests(unittest.TestCase):
expected_output_string = ["Ich liebe es so sehr!", "die Transformatoren sind wirklich erstaunlich"]
self.assertListEqual(expected_output_string, output_strings)
@unittest.skip("Does not support conversations.")
def test_pipeline_conversational(self):
pass
@require_tf
@require_sentencepiece

1
tests/models/umt5/test_modeling_umt5.py
View File

@@ -297,7 +297,6 @@ class UMT5ModelTest(ModelTesterMixin, GenerationTesterMixin, PipelineTesterMixin
all_generative_model_classes = (UMT5ForConditionalGeneration,) if is_torch_available() else ()
pipeline_model_mapping = (
{
"conversational": UMT5ForConditionalGeneration,
"feature-extraction": UMT5Model,
"question-answering": UMT5ForQuestionAnswering,
"summarization": UMT5ForConditionalGeneration,

439
tests/pipelines/test_pipelines_conversational.py
View File

@@ -1,439 +0,0 @@
# Copyright 2020 The HuggingFace Team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import gc
import unittest
from transformers import (
MODEL_FOR_CAUSAL_LM_MAPPING,
MODEL_FOR_SEQ_TO_SEQ_CAUSAL_LM_MAPPING,
TF_MODEL_FOR_CAUSAL_LM_MAPPING,
TF_MODEL_FOR_SEQ_TO_SEQ_CAUSAL_LM_MAPPING,
AutoModelForCausalLM,
AutoModelForSeq2SeqLM,
AutoTokenizer,
BlenderbotSmallForConditionalGeneration,
BlenderbotSmallTokenizer,
Conversation,
ConversationalPipeline,
TFAutoModelForCausalLM,
pipeline,
)
from transformers.testing_utils import (
backend_empty_cache,
is_pipeline_test,
is_torch_available,
require_tf,
require_torch,
slow,
torch_device,
)
from .test_pipelines_common import ANY
@is_pipeline_test
class ConversationalPipelineTests(unittest.TestCase):
def tearDown(self):
super().tearDown()
# clean-up as much as possible GPU memory occupied by PyTorch
gc.collect()
if is_torch_available():
backend_empty_cache(torch_device)
model_mapping = dict(
list(MODEL_FOR_SEQ_TO_SEQ_CAUSAL_LM_MAPPING.items())
if MODEL_FOR_SEQ_TO_SEQ_CAUSAL_LM_MAPPING
else [] + list(MODEL_FOR_CAUSAL_LM_MAPPING.items())
if MODEL_FOR_CAUSAL_LM_MAPPING
else []
)
tf_model_mapping = dict(
list(TF_MODEL_FOR_SEQ_TO_SEQ_CAUSAL_LM_MAPPING.items())
if TF_MODEL_FOR_SEQ_TO_SEQ_CAUSAL_LM_MAPPING
else [] + list(TF_MODEL_FOR_CAUSAL_LM_MAPPING.items())
if TF_MODEL_FOR_CAUSAL_LM_MAPPING
else []
)
def get_test_pipeline(self, model, tokenizer, processor):
conversation_agent = ConversationalPipeline(model=model, tokenizer=tokenizer)
return conversation_agent, [Conversation("Hi there!")]
def run_pipeline_test(self, conversation_agent, _):
# Simple
outputs = conversation_agent(Conversation("Hi there!"), max_new_tokens=5)
self.assertEqual(
outputs,
Conversation([{"role": "user", "content": "Hi there!"}, {"role": "assistant", "content": ANY(str)}]),
)
# Single list
outputs = conversation_agent([Conversation("Hi there!")], max_new_tokens=5)
self.assertEqual(
outputs,
Conversation([{"role": "user", "content": "Hi there!"}, {"role": "assistant", "content": ANY(str)}]),
)
# Batch
conversation_1 = Conversation("Going to the movies tonight - any suggestions?")
conversation_2 = Conversation("What's the last book you have read?")
self.assertEqual(len(conversation_1), 1)
self.assertEqual(len(conversation_2), 1)
outputs = conversation_agent([conversation_1, conversation_2], max_new_tokens=5)
self.assertEqual(outputs, [conversation_1, conversation_2])
self.assertEqual(
outputs,
[
Conversation(
[
{"role": "user", "content": "Going to the movies tonight - any suggestions?"},
{"role": "assistant", "content": ANY(str)},
],
),
Conversation(
[
{"role": "user", "content": "What's the last book you have read?"},
{"role": "assistant", "content": ANY(str)},
]
),
],
)
# One conversation with history
conversation_2.add_message({"role": "user", "content": "Why do you recommend it?"})
outputs = conversation_agent(conversation_2, max_new_tokens=5)
self.assertEqual(outputs, conversation_2)
self.assertEqual(
outputs,
Conversation(
[
{"role": "user", "content": "What's the last book you have read?"},
{"role": "assistant", "content": ANY(str)},
{"role": "user", "content": "Why do you recommend it?"},
{"role": "assistant", "content": ANY(str)},
]
),
)
@require_torch
@slow
def test_integration_torch_conversation(self):
# When
conversation_agent = pipeline(task="conversational", device=torch_device)
conversation_1 = Conversation("Going to the movies tonight - any suggestions?")
conversation_2 = Conversation("What's the last book you have read?")
# Then
self.assertEqual(len(conversation_1.past_user_inputs), 0)
self.assertEqual(len(conversation_2.past_user_inputs), 0)
# When
result = conversation_agent([conversation_1, conversation_2], do_sample=False, max_length=1000)
# Then
self.assertEqual(result, [conversation_1, conversation_2])
self.assertEqual(len(result[0].past_user_inputs), 1)
self.assertEqual(len(result[1].past_user_inputs), 1)
self.assertEqual(len(result[0].generated_responses), 1)
self.assertEqual(len(result[1].generated_responses), 1)
self.assertEqual(result[0].past_user_inputs[0], "Going to the movies tonight - any suggestions?")
self.assertEqual(result[0].generated_responses[0], "The Big Lebowski")
self.assertEqual(result[1].past_user_inputs[0], "What's the last book you have read?")
self.assertEqual(result[1].generated_responses[0], "The Last Question")
# When
conversation_2.add_user_input("Why do you recommend it?")
result = conversation_agent(conversation_2, do_sample=False, max_length=1000)
# Then
self.assertEqual(result, conversation_2)
self.assertEqual(len(result.past_user_inputs), 2)
self.assertEqual(len(result.generated_responses), 2)
self.assertEqual(result.past_user_inputs[1], "Why do you recommend it?")
self.assertEqual(result.generated_responses[1], "It's a good book.")
@require_torch
@slow
def test_integration_torch_conversation_truncated_history(self):
# When
conversation_agent = pipeline(task="conversational", min_length_for_response=24, device=torch_device)
conversation_1 = Conversation("Going to the movies tonight - any suggestions?")
# Then
self.assertEqual(len(conversation_1.past_user_inputs), 0)
# When
result = conversation_agent(conversation_1, do_sample=False, max_length=36)
# Then
self.assertEqual(result, conversation_1)
self.assertEqual(len(result.past_user_inputs), 1)
self.assertEqual(len(result.generated_responses), 1)
self.assertEqual(result.past_user_inputs[0], "Going to the movies tonight - any suggestions?")
self.assertEqual(result.generated_responses[0], "The Big Lebowski")
# When
conversation_1.add_user_input("Is it an action movie?")
result = conversation_agent(conversation_1, do_sample=False, max_length=36)
# Then
self.assertEqual(result, conversation_1)
self.assertEqual(len(result.past_user_inputs), 2)
self.assertEqual(len(result.generated_responses), 2)
self.assertEqual(result.past_user_inputs[1], "Is it an action movie?")
self.assertEqual(result.generated_responses[1], "It's a comedy.")
@require_torch
def test_small_model_pt(self):
tokenizer = AutoTokenizer.from_pretrained("microsoft/DialoGPT-small")
model = AutoModelForCausalLM.from_pretrained("microsoft/DialoGPT-small")
conversation_agent = ConversationalPipeline(model=model, tokenizer=tokenizer)
conversation = Conversation("hello")
output = conversation_agent(conversation)
self.assertEqual(output, Conversation(past_user_inputs=["hello"], generated_responses=["Hi"]))
@require_tf
def test_small_model_tf(self):
tokenizer = AutoTokenizer.from_pretrained("microsoft/DialoGPT-small")
model = TFAutoModelForCausalLM.from_pretrained("microsoft/DialoGPT-small")
conversation_agent = ConversationalPipeline(model=model, tokenizer=tokenizer)
conversation = Conversation("hello")
output = conversation_agent(conversation)
self.assertEqual(output, Conversation(past_user_inputs=["hello"], generated_responses=["Hi"]))
@require_torch
@slow
def test_integration_torch_conversation_dialogpt_input_ids(self):
tokenizer = AutoTokenizer.from_pretrained("microsoft/DialoGPT-small")
model = AutoModelForCausalLM.from_pretrained("microsoft/DialoGPT-small")
conversation_agent = ConversationalPipeline(model=model, tokenizer=tokenizer)
conversation_1 = Conversation("hello")
inputs = conversation_agent.preprocess(conversation_1)
self.assertEqual(inputs["input_ids"].tolist(), [[31373, 50256]])
conversation_2 = Conversation("how are you ?", past_user_inputs=["hello"], generated_responses=["Hi there!"])
inputs = conversation_agent.preprocess(conversation_2)
self.assertEqual(
inputs["input_ids"].tolist(), [[31373, 50256, 17250, 612, 0, 50256, 4919, 389, 345, 5633, 50256]]
)
@unittest.skip("Model is curently gated")
@require_torch
@slow
def test_integration_torch_conversation_llama2_input_ids(self):
tokenizer = AutoTokenizer.from_pretrained("meta-llama/Llama-2-7b-chat-hf", use_default_system_prompt=True)
conversation = Conversation(
"What is so great about #1?",
past_user_inputs=["I am going to Paris, what should I see?"],
generated_responses=[
"""\
Paris, the capital of France, is known for its stunning architecture, art museums, historical landmarks, and romantic atmosphere. Here are some of the top attractions to see in Paris:
1. The Eiffel Tower: The iconic Eiffel Tower is one of the most recognizable landmarks in the world and offers breathtaking views of the city.
2. The Louvre Museum: The Louvre is one of the world's largest and most famous museums, housing an impressive collection of art and artifacts, including the Mona Lisa.
3. Notre-Dame Cathedral: This beautiful cathedral is one of the most famous landmarks in Paris and is known for its Gothic architecture and stunning stained glass windows.
These are just a few of the many attractions that Paris has to offer. With so much to see and do, it's no wonder that Paris is one of the most popular tourist destinations in the world."""
],
)
inputs = tokenizer._build_conversation_input_ids(conversation)
EXPECTED_INPUTS_IDS = [ 1, 518, 25580, 29962, 3532, 14816, 29903, 6778, 13, 3492, 526, 263, 8444, 29892, 3390, 1319, 322, 15993, 20255, 29889, 29849, 1234, 408, 1371, 3730, 408, 1950, 29892, 1550, 1641, 9109, 29889, 29871, 3575, 6089, 881, 451, 3160, 738, 10311, 1319, 29892, 443, 621, 936, 29892, 11021, 391, 29892, 7916, 391, 29892, 304, 27375, 29892, 18215, 29892, 470, 27302, 2793, 29889, 3529, 9801, 393, 596, 20890, 526, 5374, 635, 443, 5365, 1463, 322, 6374, 297, 5469, 29889, 13, 13, 3644, 263, 1139, 947, 451, 1207, 738, 4060, 29892, 470, 338, 451, 2114, 1474, 16165, 261, 296, 29892, 5649, 2020, 2012, 310, 22862, 1554, 451, 1959, 29889, 960, 366, 1016, 29915, 29873, 1073, 278, 1234, 304, 263, 1139, 29892, 3113, 1016, 29915, 29873, 6232, 2089, 2472, 29889, 13, 29966, 829, 14816, 29903, 6778, 13, 13, 29902, 626, 2675, 304, 3681, 29892, 825, 881, 306, 1074, 29973, 518, 29914, 25580, 29962, 3681, 29892, 278, 7483, 310, 3444, 29892, 338, 2998, 363, 967, 380, 27389, 11258, 29892, 1616, 19133, 29879, 29892, 15839, 2982, 22848, 29892, 322, 6017, 7716, 25005, 29889, 2266, 526, 777, 310, 278, 2246, 19650, 1953, 304, 1074, 297, 3681, 29901, 13, 13, 29896, 29889, 450, 382, 2593, 295, 23615, 29901, 450, 9849, 293, 382, 2593, 295, 23615, 338, 697, 310, 278, 1556, 5936, 13902, 2982, 22848, 297, 278, 3186, 322, 16688, 2078, 271, 400, 5086, 8386, 310, 278, 4272, 29889, 13, 29906, 29889, 450, 4562, 12675, 6838, 29901, 450, 4562, 12675, 338, 697, 310, 278, 3186, 29915, 29879, 10150, 322, 1556, 13834, 19133, 29879, 29892, 27261, 385, 21210, 573, 4333, 310, 1616, 322, 24238, 29879, 29892, 3704, 278, 2598, 29874, 29420, 29889, 13, 29941, 29889, 24337, 29899, 29928, 420, 315, 21471, 29901, 910, 9560, 274, 21471, 338, 697, 310, 278, 1556, 13834, 2982, 22848, 297, 3681, 322, 338, 2998, 363, 967, 22883, 293, 11258, 322, 380, 27389, 380, 7114, 12917, 5417, 29889, 13, 13, 1349, 968, 526, 925, 263, 2846, 310, 278, 1784, 19650, 1953, 393, 3681, 756, 304, 5957, 29889, 2973, 577, 1568, 304, 1074, 322, 437, 29892, 372, 29915, 29879, 694, 4997, 393, 3681, 338, 697, 310, 278, 1556, 5972, 6282, 391, 15422, 800, 297, 278, 3186, 29889, 29871, 2, 1, 518, 25580, 29962, 1724, 338, 577, 2107, 1048, 396, 29896, 29973, 518, 29914, 25580, 29962] # fmt: skip
self.assertEqual(inputs, EXPECTED_INPUTS_IDS)
model = AutoModelForCausalLM.from_pretrained("meta-llama/Llama-2-7b-chat-hf")
conversation_agent = ConversationalPipeline(model=model, tokenizer=tokenizer)
EXPECTED_TEXT = "what topic you want to focus on and create content around it. This will help you stand out from other creators and attract a specific audience.\n\nStep 2: Set Up Your Channel\nCreate your YouTube account and customize your channel with your branding and logo. Make sure your channel name and profile picture are consistent with your niche.\n\nStep 3: Plan Your Content\nDevelop a content strategy that includes the type of content you want to create, how often you will post, and when you will post. Consider creating a content calendar to help you stay organized.\n\nStep 4: Invest in Quality Equipment\nInvest in good quality camera and microphone equipment to ensure your videos look and sound professional. You don't need to break the bank, but investing in good equipment will make a big difference in the quality of your videos.\n\nStep 5: Optimize Your Videos for Search\nUse keywords in your video titles, descriptions, and tags to help people find your videos when they search for topics related to your niche"
conversation = Conversation(
"<<SYS>>\n Only answer with emojis, and charades\n<</SYS>>\n\nHow can I build a house in 10 steps?"
)
result = conversation_agent(conversation)
self.assertEqual(result.generated_responses[-1], EXPECTED_TEXT)
@require_torch
@slow
def test_integration_torch_conversation_blenderbot_400M_input_ids(self):
tokenizer = AutoTokenizer.from_pretrained("facebook/blenderbot-400M-distill")
model = AutoModelForSeq2SeqLM.from_pretrained("facebook/blenderbot-400M-distill")
conversation_agent = ConversationalPipeline(model=model, tokenizer=tokenizer)
# test1
conversation_1 = Conversation("hello")
inputs = conversation_agent.preprocess(conversation_1)
self.assertEqual(inputs["input_ids"].tolist(), [[1710, 86, 2]])
# test2
conversation_1 = Conversation(
"I like lasagne.",
past_user_inputs=["hello"],
generated_responses=[
" Do you like lasagne? It is a traditional Italian dish consisting of a shepherd's pie."
],
)
inputs = conversation_agent.preprocess(conversation_1)
self.assertEqual(
inputs["input_ids"].tolist(),
[
# This should be compared with the same conversation on ParlAI `safe_interactive` demo.
[
1710, # hello
86,
228, # Double space
228,
946,
304,
398,
6881,
558,
964,
38,
452,
315,
265,
6252,
452,
322,
968,
6884,
3146,
278,
306,
265,
617,
87,
388,
75,
341,
286,
521,
21,
228, # Double space
228,
281, # I like lasagne.
398,
6881,
558,
964,
21,
2, # EOS
],
],
)
@require_torch
@slow
def test_integration_torch_conversation_blenderbot_400M(self):
tokenizer = AutoTokenizer.from_pretrained("facebook/blenderbot-400M-distill")
model = AutoModelForSeq2SeqLM.from_pretrained("facebook/blenderbot-400M-distill")
conversation_agent = ConversationalPipeline(model=model, tokenizer=tokenizer)
conversation_1 = Conversation("hello")
result = conversation_agent(
conversation_1,
)
self.assertEqual(
result.generated_responses[0],
# ParlAI implementation output, we have a different one, but it's our
# second best, you can check by using num_return_sequences=10
# " Hello! How are you? I'm just getting ready to go to work, how about you?",
" Hello! How are you doing today? I just got back from a walk with my dog.",
)
conversation_1 = Conversation("Lasagne hello")
result = conversation_agent(conversation_1, encoder_no_repeat_ngram_size=3)
self.assertEqual(
result.generated_responses[0],
" Do you like lasagne? It is a traditional Italian dish consisting of a shepherd's pie.",
)
conversation_1 = Conversation(
"Lasagne hello Lasagne is my favorite Italian dish. Do you like lasagne? I like lasagne."
)
result = conversation_agent(
conversation_1,
encoder_no_repeat_ngram_size=3,
)
self.assertEqual(
result.generated_responses[0],
" Me too. I like how it can be topped with vegetables, meats, and condiments.",
)
@require_torch
@slow
def test_integration_torch_conversation_encoder_decoder(self):
# When
tokenizer = AutoTokenizer.from_pretrained("facebook/blenderbot_small-90M")
model = AutoModelForSeq2SeqLM.from_pretrained("facebook/blenderbot_small-90M")
conversation_agent = ConversationalPipeline(model=model, tokenizer=tokenizer, device=torch_device)
conversation_1 = Conversation("My name is Sarah and I live in London")
conversation_2 = Conversation("Going to the movies tonight, What movie would you recommend? ")
# Then
self.assertEqual(len(conversation_1.past_user_inputs), 0)
self.assertEqual(len(conversation_2.past_user_inputs), 0)
# When
result = conversation_agent([conversation_1, conversation_2], do_sample=False, max_length=1000)
# Then
self.assertEqual(result, [conversation_1, conversation_2])
self.assertEqual(len(result[0].past_user_inputs), 1)
self.assertEqual(len(result[1].past_user_inputs), 1)
self.assertEqual(len(result[0].generated_responses), 1)
self.assertEqual(len(result[1].generated_responses), 1)
self.assertEqual(result[0].past_user_inputs[0], "My name is Sarah and I live in London")
self.assertEqual(
result[0].generated_responses[0],
"hi sarah, i live in london as well. do you have any plans for the weekend?",
)
self.assertEqual(
result[1].past_user_inputs[0], "Going to the movies tonight, What movie would you recommend? "
)
self.assertEqual(
result[1].generated_responses[0], "i don't know... i'm not really sure. what movie are you going to see?"
)
# When
conversation_1.add_user_input("Not yet, what about you?")
conversation_2.add_user_input("What's your name?")
result = conversation_agent([conversation_1, conversation_2], do_sample=False, max_length=1000)
# Then
self.assertEqual(result, [conversation_1, conversation_2])
self.assertEqual(len(result[0].past_user_inputs), 2)
self.assertEqual(len(result[1].past_user_inputs), 2)
self.assertEqual(len(result[0].generated_responses), 2)
self.assertEqual(len(result[1].generated_responses), 2)
self.assertEqual(result[0].past_user_inputs[1], "Not yet, what about you?")
self.assertEqual(result[0].generated_responses[1], "i don't have any plans yet. i'm not sure what to do yet.")
self.assertEqual(result[1].past_user_inputs[1], "What's your name?")
self.assertEqual(result[1].generated_responses[1], "i don't have a name, but i'm going to see a horror movie.")
@require_torch
@slow
def test_from_pipeline_conversation(self):
model_id = "facebook/blenderbot_small-90M"
# from model id
conversation_agent_from_model_id = pipeline("conversational", model=model_id, tokenizer=model_id)
# from model object
model = BlenderbotSmallForConditionalGeneration.from_pretrained(model_id)
tokenizer = BlenderbotSmallTokenizer.from_pretrained(model_id)
conversation_agent_from_model = pipeline("conversational", model=model, tokenizer=tokenizer)
conversation = Conversation("My name is Sarah and I live in London")
conversation_copy = Conversation("My name is Sarah and I live in London")
result_model_id = conversation_agent_from_model_id([conversation])
result_model = conversation_agent_from_model([conversation_copy])
# check for equality
self.assertEqual(
result_model_id.generated_responses[0],
"hi sarah, i live in london as well. do you have any plans for the weekend?",
)
self.assertEqual(
result_model_id.generated_responses[0],
result_model.generated_responses[0],
)

14
tests/test_pipeline_mixin.py
View File

@@ -33,7 +33,6 @@ from transformers.utils import direct_transformers_import, logging
from .pipelines.test_pipelines_audio_classification import AudioClassificationPipelineTests
from .pipelines.test_pipelines_automatic_speech_recognition import AutomaticSpeechRecognitionPipelineTests
from .pipelines.test_pipelines_conversational import ConversationalPipelineTests
from .pipelines.test_pipelines_depth_estimation import DepthEstimationPipelineTests
from .pipelines.test_pipelines_document_question_answering import DocumentQuestionAnsweringPipelineTests
from .pipelines.test_pipelines_feature_extraction import FeatureExtractionPipelineTests
@@ -65,7 +64,6 @@ from .pipelines.test_pipelines_zero_shot_object_detection import ZeroShotObjectD
pipeline_test_mapping = {
"audio-classification": {"test": AudioClassificationPipelineTests},
"automatic-speech-recognition": {"test": AutomaticSpeechRecognitionPipelineTests},
"conversational": {"test": ConversationalPipelineTests},
"depth-estimation": {"test": DepthEstimationPipelineTests},
"document-question-answering": {"test": DocumentQuestionAnsweringPipelineTests},
"feature-extraction": {"test": FeatureExtractionPipelineTests},
@@ -314,12 +312,8 @@ class PipelineTesterMixin:
yield copy.deepcopy(random.choice(examples))
out = []
if task == "conversational":
for item in pipeline(data(10), batch_size=4, max_new_tokens=5):
out.append(item)
else:
for item in pipeline(data(10), batch_size=4):
out.append(item)
for item in pipeline(data(10), batch_size=4):
out.append(item)
self.assertEqual(len(out), 10)
run_batch_test(pipeline, examples)
@@ -332,10 +326,6 @@ class PipelineTesterMixin:
def test_pipeline_automatic_speech_recognition(self):
self.run_task_tests(task="automatic-speech-recognition")
@is_pipeline_test
def test_pipeline_conversational(self):
self.run_task_tests(task="conversational")
@is_pipeline_test
@require_vision
@require_timm