tocmo0nlord/axolotl
1
0
Fork 0
Code Issues Pull Requests Actions 4 Packages Projects Releases Wiki Activity

fix: minor patches for multimodal (#2441)

Browse Source 
* fix: update chat_template

* fix: handle gemma3 showing a lot of no content for turn 0

* fix: remove unknown config from examples

* fix: test

* fix: temporary disable gemma2 test

* fix: stop overwriting config.text_config unnecessarily

* fix: handling of set cache to the text_config section

* feat: add liger gemma support and bump liger to 0.5.5

* fix: add double use_cache setting

* fix: add support for final_logit_softcap in CCE for gemma2/3

* fix: set use_cache before model load

* feat: add missing layernorm override

* fix: handle gemma3 rmsnorm

* fix: use wrapper to pass dim as hidden_size

* fix: change dim to positional

* fix: patch with wrong mlp

* chore: refactor use_cache handling

* fix import issues

* fix tests.e2e.utils import

---------

Co-authored-by: Wing Lian <wing@axolotl.ai>
This commit is contained in:
NanoCode012
2025-03-31 13:40:12 +07:00
committed by GitHub
parent 4ba80a0e5a
commit cf0c79d52e
38 changed files with 287 additions and 188 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
tests/e2e/multigpu/test_grpo.py
View File

@@ -8,11 +8,12 @@ from pathlib import Path
import pytest
import yaml
from accelerate.test_utils import execute_subprocess_async
from e2e.utils import require_vllm
from transformers.testing_utils import get_torch_dist_unique_port
from axolotl.utils.dict import DictDefault
from tests.e2e.utils import require_vllm
class TestGRPO:
"""

3
tests/e2e/multigpu/test_llama.py
View File

@@ -9,12 +9,13 @@ from pathlib import Path
import pytest
import yaml
from accelerate.test_utils import execute_subprocess_async
from e2e.utils import check_tensorboard
from huggingface_hub import snapshot_download
from transformers.testing_utils import get_torch_dist_unique_port
from axolotl.utils.dict import DictDefault
from tests.e2e.utils import check_tensorboard
LOG = logging.getLogger("axolotl.tests.e2e.multigpu")
os.environ["WANDB_DISABLED"] = "true"

3
tests/e2e/multigpu/test_ray.py
View File

@@ -9,10 +9,11 @@ from pathlib import Path
import pytest
import yaml
from accelerate.test_utils import execute_subprocess_async
from e2e.utils import check_tensorboard, require_torch_lt_2_6_0
from axolotl.utils.dict import DictDefault
from tests.e2e.utils import check_tensorboard, require_torch_lt_2_6_0
LOG = logging.getLogger(__name__)
os.environ["WANDB_DISABLED"] = "true"