fix: minor patches for multimodal (#2441)

* 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>

tests/test_packed_batch_sampler.py

@@ -4,7 +4,6 @@ import pytest
 from datasets import concatenate_datasets, load_dataset
 from torch.utils.data import DataLoader, RandomSampler
 from transformers import AutoTokenizer
-from utils import enable_hf_offline
 
 from axolotl.datasets import TokenizedPromptDataset
 from axolotl.prompt_strategies.completion import load
@@ -13,6 +12,8 @@ from axolotl.utils.data.utils import drop_long_seq_in_dataset
 from axolotl.utils.dict import DictDefault
 from axolotl.utils.samplers import MultipackBatchSampler, get_dataset_lengths
 
+from tests.hf_offline_utils import enable_hf_offline
+
 
 @pytest.fixture(name="tokenizer")
 def fixture_tokenizer():