diff --git a/.github/workflows/pr_tests.yml b/.github/workflows/pr_tests.yml
index 1208a0d2..f63d4ffd 100644
--- a/.github/workflows/pr_tests.yml
+++ b/.github/workflows/pr_tests.yml
@@ -41,4 +41,15 @@ jobs:
 
     - name: Run all non-slow selected tests on CPU
       run: |
-        python -m pytest -n 2 --max-worker-restart=0 --dist=loadfile -s tests/
+        python -m pytest -n 2 --max-worker-restart=0 --dist=loadfile -s -v --make-reports=tests_torch_cpu tests/
+
+    - name: Failure short reports
+      if: ${{ failure() }}
+      run: cat reports/tests_torch_cpu_failures_short.txt
+
+    - name: Test suite reports artifacts
+      if: ${{ always() }}
+      uses: actions/upload-artifact@v2
+      with:
+        name: pr_torch_test_reports
+        path: reports
diff --git a/.github/workflows/push_tests.yml b/.github/workflows/push_tests.yml
index 58ecc0d3..d1317ac9 100644
--- a/.github/workflows/push_tests.yml
+++ b/.github/workflows/push_tests.yml
@@ -49,4 +49,15 @@ jobs:
       env:
         HUGGING_FACE_HUB_TOKEN: ${{ secrets.HUGGING_FACE_HUB_TOKEN }}
       run: |
-        python -m pytest -n 1 --max-worker-restart=0 --dist=loadfile -s tests/
+        python -m pytest -n 1 --max-worker-restart=0 --dist=loadfile -s -v --make-reports=tests_torch_gpu tests/
+
+    - name: Failure short reports
+      if: ${{ failure() }}
+      run: cat reports/tests_torch_gpu_failures_short.txt
+
+    - name: Test suite reports artifacts
+      if: ${{ always() }}
+      uses: actions/upload-artifact@v2
+      with:
+        name: push_torch_test_reports
+        path: reports
diff --git a/src/diffusers/testing_utils.py b/src/diffusers/testing_utils.py
index 7daf2bc6..d3f6fa62 100644
--- a/src/diffusers/testing_utils.py
+++ b/src/diffusers/testing_utils.py
@@ -1,7 +1,9 @@
 import os
 import random
+import re
 import unittest
 from distutils.util import strtobool
+from pathlib import Path
 from typing import Union
 
 import torch
@@ -92,3 +94,157 @@ def load_image(image: Union[str, PIL.Image.Image]) -> PIL.Image.Image:
     image = PIL.ImageOps.exif_transpose(image)
     image = image.convert("RGB")
     return image
+
+
+# --- pytest conf functions --- #
+
+# to avoid multiple invocation from tests/conftest.py and examples/conftest.py - make sure it's called only once
+pytest_opt_registered = {}
+
+
+def pytest_addoption_shared(parser):
+    """
+    This function is to be called from `conftest.py` via `pytest_addoption` wrapper that has to be defined there.
+
+    It allows loading both `conftest.py` files at once without causing a failure due to adding the same `pytest`
+    option.
+
+    """
+    option = "--make-reports"
+    if option not in pytest_opt_registered:
+        parser.addoption(
+            option,
+            action="store",
+            default=False,
+            help="generate report files. The value of this option is used as a prefix to report names",
+        )
+        pytest_opt_registered[option] = 1
+
+
+def pytest_terminal_summary_main(tr, id):
+    """
+    Generate multiple reports at the end of test suite run - each report goes into a dedicated file in the current
+    directory. The report files are prefixed with the test suite name.
+
+    This function emulates --duration and -rA pytest arguments.
+
+    This function is to be called from `conftest.py` via `pytest_terminal_summary` wrapper that has to be defined
+    there.
+
+    Args:
+    - tr: `terminalreporter` passed from `conftest.py`
+    - id: unique id like `tests` or `examples` that will be incorporated into the final reports filenames - this is
+      needed as some jobs have multiple runs of pytest, so we can't have them overwrite each other.
+
+    NB: this functions taps into a private _pytest API and while unlikely, it could break should
+    pytest do internal changes - also it calls default internal methods of terminalreporter which
+    can be hijacked by various `pytest-` plugins and interfere.
+
+    """
+    from _pytest.config import create_terminal_writer
+
+    if not len(id):
+        id = "tests"
+
+    config = tr.config
+    orig_writer = config.get_terminal_writer()
+    orig_tbstyle = config.option.tbstyle
+    orig_reportchars = tr.reportchars
+
+    dir = "reports"
+    Path(dir).mkdir(parents=True, exist_ok=True)
+    report_files = {
+        k: f"{dir}/{id}_{k}.txt"
+        for k in [
+            "durations",
+            "errors",
+            "failures_long",
+            "failures_short",
+            "failures_line",
+            "passes",
+            "stats",
+            "summary_short",
+            "warnings",
+        ]
+    }
+
+    # custom durations report
+    # note: there is no need to call pytest --durations=XX to get this separate report
+    # adapted from https://github.com/pytest-dev/pytest/blob/897f151e/src/_pytest/runner.py#L66
+    dlist = []
+    for replist in tr.stats.values():
+        for rep in replist:
+            if hasattr(rep, "duration"):
+                dlist.append(rep)
+    if dlist:
+        dlist.sort(key=lambda x: x.duration, reverse=True)
+        with open(report_files["durations"], "w") as f:
+            durations_min = 0.05  # sec
+            f.write("slowest durations\n")
+            for i, rep in enumerate(dlist):
+                if rep.duration < durations_min:
+                    f.write(f"{len(dlist)-i} durations < {durations_min} secs were omitted")
+                    break
+                f.write(f"{rep.duration:02.2f}s {rep.when:<8} {rep.nodeid}\n")
+
+    def summary_failures_short(tr):
+        # expecting that the reports were --tb=long (default) so we chop them off here to the last frame
+        reports = tr.getreports("failed")
+        if not reports:
+            return
+        tr.write_sep("=", "FAILURES SHORT STACK")
+        for rep in reports:
+            msg = tr._getfailureheadline(rep)
+            tr.write_sep("_", msg, red=True, bold=True)
+            # chop off the optional leading extra frames, leaving only the last one
+            longrepr = re.sub(r".*_ _ _ (_ ){10,}_ _ ", "", rep.longreprtext, 0, re.M | re.S)
+            tr._tw.line(longrepr)
+            # note: not printing out any rep.sections to keep the report short
+
+    # use ready-made report funcs, we are just hijacking the filehandle to log to a dedicated file each
+    # adapted from https://github.com/pytest-dev/pytest/blob/897f151e/src/_pytest/terminal.py#L814
+    # note: some pytest plugins may interfere by hijacking the default `terminalreporter` (e.g.
+    # pytest-instafail does that)
+
+    # report failures with line/short/long styles
+    config.option.tbstyle = "auto"  # full tb
+    with open(report_files["failures_long"], "w") as f:
+        tr._tw = create_terminal_writer(config, f)
+        tr.summary_failures()
+
+    # config.option.tbstyle = "short" # short tb
+    with open(report_files["failures_short"], "w") as f:
+        tr._tw = create_terminal_writer(config, f)
+        summary_failures_short(tr)
+
+    config.option.tbstyle = "line"  # one line per error
+    with open(report_files["failures_line"], "w") as f:
+        tr._tw = create_terminal_writer(config, f)
+        tr.summary_failures()
+
+    with open(report_files["errors"], "w") as f:
+        tr._tw = create_terminal_writer(config, f)
+        tr.summary_errors()
+
+    with open(report_files["warnings"], "w") as f:
+        tr._tw = create_terminal_writer(config, f)
+        tr.summary_warnings()  # normal warnings
+        tr.summary_warnings()  # final warnings
+
+    tr.reportchars = "wPpsxXEf"  # emulate -rA (used in summary_passes() and short_test_summary())
+    with open(report_files["passes"], "w") as f:
+        tr._tw = create_terminal_writer(config, f)
+        tr.summary_passes()
+
+    with open(report_files["summary_short"], "w") as f:
+        tr._tw = create_terminal_writer(config, f)
+        tr.short_test_summary()
+
+    with open(report_files["stats"], "w") as f:
+        tr._tw = create_terminal_writer(config, f)
+        tr.summary_stats()
+
+    # restore:
+    tr._tw = orig_writer
+    tr.reportchars = orig_reportchars
+    config.option.tbstyle = orig_tbstyle
diff --git a/tests/conftest.py b/tests/conftest.py
new file mode 100644
index 00000000..e116f40e
--- /dev/null
+++ b/tests/conftest.py
@@ -0,0 +1,44 @@
+# Copyright 2022 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.
+
+# tests directory-specific settings - this file is run automatically
+# by pytest before any tests are run
+
+import sys
+import warnings
+from os.path import abspath, dirname, join
+
+
+# allow having multiple repository checkouts and not needing to remember to rerun
+# 'pip install -e .[dev]' when switching between checkouts and running tests.
+git_repo_path = abspath(join(dirname(dirname(__file__)), "src"))
+sys.path.insert(1, git_repo_path)
+
+# silence FutureWarning warnings in tests since often we can't act on them until
+# they become normal warnings - i.e. the tests still need to test the current functionality
+warnings.simplefilter(action="ignore", category=FutureWarning)
+
+
+def pytest_addoption(parser):
+    from diffusers.testing_utils import pytest_addoption_shared
+
+    pytest_addoption_shared(parser)
+
+
+def pytest_terminal_summary(terminalreporter):
+    from diffusers.testing_utils import pytest_terminal_summary_main
+
+    make_reports = terminalreporter.config.getoption("--make-reports")
+    if make_reports:
+        pytest_terminal_summary_main(terminalreporter, id=make_reports)