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

Clean push to hub API (#12187)

Browse Source 
* Clean push to hub API

* Create working dir if it does not exist

* Different tweak

* New API + all models + test Flax

* Adds the Trainer clean up

* Update src/transformers/file_utils.py

Co-authored-by: Lysandre Debut <lysandre@huggingface.co>

* Address review comments

* (nit) output types

* No need to set clone_from when folder exists

* Update src/transformers/trainer.py

Co-authored-by: Julien Chaumond <julien@huggingface.co>

* Add generated_from_trainer tag

* Update to new version

* Fixes

Co-authored-by: Lysandre Debut <lysandre@huggingface.co>
Co-authored-by: Julien Chaumond <julien@huggingface.co>
Co-authored-by: Lysandre <lysandre.debut@reseau.eseo.fr>
This commit is contained in:
Sylvain Gugger
2021-06-23 10:11:19 -04:00
committed by GitHub
parent 625f512d5e
commit 53c60babe4
17 changed files with 368 additions and 159 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
tests/test_configuration_common.py
View File

@@ -113,7 +113,7 @@ class ConfigPushToHubTester(unittest.TestCase):
vocab_size=99, hidden_size=32, num_hidden_layers=5, num_attention_heads=4, intermediate_size=37
)
with tempfile.TemporaryDirectory() as tmp_dir:
config.save_pretrained(tmp_dir, push_to_hub=True, repo_name="test-config", use_auth_token=self._token)
config.save_pretrained(os.path.join(tmp_dir, "test-config"), push_to_hub=True, use_auth_token=self._token)
new_config = BertConfig.from_pretrained(f"{USER}/test-config")
for k, v in config.__dict__.items():
@@ -127,9 +127,8 @@ class ConfigPushToHubTester(unittest.TestCase):
with tempfile.TemporaryDirectory() as tmp_dir:
config.save_pretrained(
tmp_dir,
os.path.join(tmp_dir, "test-config-org"),
push_to_hub=True,
repo_name="test-config-org",
use_auth_token=self._token,
organization="valid_org",
)

5
tests/test_modeling_common.py
View File

@@ -1613,7 +1613,7 @@ class ModelPushToHubTester(unittest.TestCase):
)
model = BertModel(config)
with tempfile.TemporaryDirectory() as tmp_dir:
model.save_pretrained(tmp_dir, push_to_hub=True, repo_name="test-model", use_auth_token=self._token)
model.save_pretrained(os.path.join(tmp_dir, "test-model"), push_to_hub=True, use_auth_token=self._token)
new_model = BertModel.from_pretrained(f"{USER}/test-model")
for p1, p2 in zip(model.parameters(), new_model.parameters()):
@@ -1626,9 +1626,8 @@ class ModelPushToHubTester(unittest.TestCase):
model = BertModel(config)
with tempfile.TemporaryDirectory() as tmp_dir:
model.save_pretrained(
tmp_dir,
os.path.join(tmp_dir, "test-model-org"),
push_to_hub=True,
repo_name="test-model-org",
use_auth_token=self._token,
organization="valid_org",
)

77
tests/test_modeling_flax_common.py
View File

@@ -16,14 +16,25 @@ import copy
import inspect
import random
import tempfile
import unittest
from typing import List, Tuple
import numpy as np
import transformers
from transformers import is_flax_available, is_torch_available
from huggingface_hub import HfApi
from requests.exceptions import HTTPError
from transformers import BertConfig, FlaxBertModel, is_flax_available, is_torch_available
from transformers.models.auto import get_values
from transformers.testing_utils import is_pt_flax_cross_test, require_flax, slow
from transformers.testing_utils import (
ENDPOINT_STAGING,
PASS,
USER,
is_pt_flax_cross_test,
is_staging_test,
require_flax,
slow,
)
if is_flax_available():
@@ -504,3 +515,65 @@ class FlaxModelTesterMixin:
list(self_attentions[0].shape[-3:]),
[self.model_tester.num_attention_heads, encoder_seq_length, encoder_key_length],
)
@require_flax
@is_staging_test
class FlaxModelPushToHubTester(unittest.TestCase):
@classmethod
def setUpClass(cls):
cls._api = HfApi(endpoint=ENDPOINT_STAGING)
cls._token = cls._api.login(username=USER, password=PASS)
@classmethod
def tearDownClass(cls):
try:
cls._api.delete_repo(token=cls._token, name="test-model-flax")
except HTTPError:
pass
try:
cls._api.delete_repo(token=cls._token, name="test-model-flax-org", organization="valid_org")
except HTTPError:
pass
def test_push_to_hub(self):
config = BertConfig(
vocab_size=99, hidden_size=32, num_hidden_layers=5, num_attention_heads=4, intermediate_size=37
)
model = FlaxBertModel(config)
with tempfile.TemporaryDirectory() as tmp_dir:
model.save_pretrained(
os.path.join(tmp_dir, "test-model-flax"), push_to_hub=True, use_auth_token=self._token
)
new_model = FlaxBertModel.from_pretrained(f"{USER}/test-model-flax")
base_params = flatten_dict(unfreeze(model.params))
new_params = flatten_dict(unfreeze(new_model.params))
for key in base_params.keys():
max_diff = (base_params[key] - new_params[key]).sum().item()
self.assertLessEqual(max_diff, 1e-3, msg=f"{key} not identical")
def test_push_to_hub_in_organization(self):
config = BertConfig(
vocab_size=99, hidden_size=32, num_hidden_layers=5, num_attention_heads=4, intermediate_size=37
)
model = FlaxBertModel(config)
with tempfile.TemporaryDirectory() as tmp_dir:
model.save_pretrained(
os.path.join(tmp_dir, "test-model-flax-org"),
push_to_hub=True,
use_auth_token=self._token,
organization="valid_org",
)
new_model = FlaxBertModel.from_pretrained("valid_org/test-model-flax-org")
base_params = flatten_dict(unfreeze(model.params))
new_params = flatten_dict(unfreeze(new_model.params))
for key in base_params.keys():
max_diff = (base_params[key] - new_params[key]).sum().item()
self.assertLessEqual(max_diff, 1e-3, msg=f"{key} not identical")

5
tests/test_modeling_tf_common.py
View File

@@ -1487,7 +1487,7 @@ class TFModelPushToHubTester(unittest.TestCase):
# Make sure model is properly initialized
_ = model(model.dummy_inputs)
with tempfile.TemporaryDirectory() as tmp_dir:
model.save_pretrained(tmp_dir, push_to_hub=True, repo_name="test-model-tf", use_auth_token=self._token)
model.save_pretrained(os.path.join(tmp_dir, "test-model-tf"), push_to_hub=True, use_auth_token=self._token)
new_model = TFBertModel.from_pretrained(f"{USER}/test-model-tf")
models_equal = True
@@ -1503,9 +1503,8 @@ class TFModelPushToHubTester(unittest.TestCase):
model = TFBertModel(config)
with tempfile.TemporaryDirectory() as tmp_dir:
model.save_pretrained(
tmp_dir,
os.path.join(tmp_dir, "test-model-tf-org"),
push_to_hub=True,
repo_name="test-model-tf-org",
use_auth_token=self._token,
organization="valid_org",
)

5
tests/test_tokenization_common.py
View File

@@ -3173,7 +3173,7 @@ class TokenizerPushToHubTester(unittest.TestCase):
vocab_writer.write("".join([x + "\n" for x in self.vocab_tokens]))
tokenizer = BertTokenizer(vocab_file)
tokenizer.save_pretrained(
tmp_dir, push_to_hub=True, repo_name="test-tokenizer", use_auth_token=self._token
os.path.join(tmp_dir, "test-tokenizer"), push_to_hub=True, use_auth_token=self._token
)
new_tokenizer = BertTokenizer.from_pretrained(f"{USER}/test-tokenizer")
@@ -3186,9 +3186,8 @@ class TokenizerPushToHubTester(unittest.TestCase):
vocab_writer.write("".join([x + "\n" for x in self.vocab_tokens]))
tokenizer = BertTokenizer(vocab_file)
tokenizer.save_pretrained(
tmp_dir,
os.path.join(tmp_dir, "test-tokenizer-org"),
push_to_hub=True,
repo_name="test-tokenizer-org",
use_auth_token=self._token,
organization="valid_org",
)

16
tests/test_trainer.py
View File

@@ -1274,8 +1274,12 @@ class TrainerIntegrationWithHubTester(unittest.TestCase):
def test_push_to_hub(self):
with tempfile.TemporaryDirectory() as tmp_dir:
trainer = get_regression_trainer(output_dir=tmp_dir)
url = trainer.push_to_hub(repo_name="test-trainer", use_auth_token=self._token)
trainer = get_regression_trainer(
output_dir=os.path.join(tmp_dir, "test-trainer"),
push_to_hub=True,
push_to_hub_token=self._token,
)
url = trainer.push_to_hub()
# Extract repo_name from the url
re_search = re.search(ENDPOINT_STAGING + r"/([^/]+/[^/]+)/", url)
@@ -1292,9 +1296,13 @@ class TrainerIntegrationWithHubTester(unittest.TestCase):
with tempfile.TemporaryDirectory() as tmp_dir:
trainer = get_regression_trainer(output_dir=tmp_dir)
trainer.save_model()
url = trainer.push_to_hub(
repo_name="test-trainer-org", organization="valid_org", use_auth_token=self._token
trainer = get_regression_trainer(
output_dir=os.path.join(tmp_dir, "test-trainer-org"),
push_to_hub=True,
push_to_hub_organization="valid_org",
push_to_hub_token=self._token,
)
url = trainer.push_to_hub()
# Extract repo_name from the url
re_search = re.search(ENDPOINT_STAGING + r"/([^/]+/[^/]+)/", url)