Merge branch 'main' into map-dataset-fetcher-fix

.github/workflows/base.yml

@@ -5,13 +5,11 @@ on:
     branches:
       - "main"
     paths:
-      - 'docker/Dockerfile-base'
+      - 'Dockerfile-base'
-      - 'docker/Dockerfile-uv-base'
       - '.github/workflows/base.yml'
   pull_request:
     paths:
-      - 'docker/Dockerfile-base'
+      - 'Dockerfile-base'
-      - 'docker/Dockerfile-uv-base'
       - '.github/workflows/base.yml'
   workflow_dispatch:
 

.github/workflows/main.yml

@@ -20,11 +20,12 @@ jobs:
             python_version: "3.11"
             pytorch: 2.5.1
             axolotl_extras:
-          - cuda: 126
+          - cuda: 124
-            cuda_version: 12.6.3
+            cuda_version: 12.4.1
             python_version: "3.11"
             pytorch: 2.6.0
             axolotl_extras: vllm
+            is_latest: true
           - cuda: 126
             cuda_version: 12.6.3
             python_version: "3.11"
@@ -87,8 +88,8 @@ jobs:
             python_version: "3.11"
             pytorch: 2.5.1
             axolotl_extras:
-          - cuda: 126
+          - cuda: 124
-            cuda_version: 12.6.3
+            cuda_version: 12.4.1
             python_version: "3.11"
             pytorch: 2.6.0
             axolotl_extras:
@@ -145,8 +146,8 @@ jobs:
     strategy:
       matrix:
         include:
-          - cuda: 126
+          - cuda: 124
-            cuda_version: 12.6.3
+            cuda_version: 12.4.1
             python_version: "3.11"
             pytorch: 2.6.0
             axolotl_extras:

.github/workflows/multi-gpu-e2e.yml

@@ -26,11 +26,11 @@ jobs:
       fail-fast: false
       matrix:
         include:
-          - cuda: 126
+          - cuda: 124
-            cuda_version: 12.6.3
+            cuda_version: 12.4.1
             python_version: "3.11"
             pytorch: 2.6.0
-            axolotl_extras:
+            axolotl_extras: vllm
             num_gpus: 2
             nightly_build: "true"
           - cuda: 124

.github/workflows/tests-nightly.yml

@@ -18,9 +18,96 @@ jobs:
         env:
           SKIP: no-commit-to-branch
 
+  preload-cache:
+    name: Preload HF cache
+    runs-on: ubuntu-latest
+    strategy:
+      fail-fast: false
+      matrix:
+        python_version: ["3.11"]
+        pytorch_version: ["2.6.0"]
+    timeout-minutes: 20
+
+    env:
+      AXOLOTL_IS_CI_CACHE_PRELOAD: "1"
+
+    steps:
+      - name: Check out repository code
+        uses: actions/checkout@v4
+
+      - name: Restore HF cache
+        id: hf-cache-restore
+        uses: actions/cache/restore@v4
+        with:
+          path: |
+            /home/runner/.cache/huggingface/hub/datasets--*
+            /home/runner/.cache/huggingface/hub/models--*
+          key: ${{ runner.os }}-hf-hub-cache-v2
+
+      - name: Setup Python
+        uses: actions/setup-python@v5
+        with:
+          python-version: ${{ matrix.python_version }}
+          cache: 'pip' # caching pip dependencies
+
+      - name: upgrade pip
+        run: |
+          pip3 install --upgrade pip
+          pip3 install --upgrade packaging==23.2 setuptools==75.8.0 wheel
+
+      - name: Install PyTorch
+        run: |
+          pip3 install torch==${{ matrix.pytorch_version }}
+
+      - name: Install dependencies
+        run: |
+          pip3 show torch
+          pip3 install --no-build-isolation -U -e .
+          python scripts/unsloth_install.py | sh
+          python scripts/cutcrossentropy_install.py | sh
+          pip3 install -r requirements-dev.txt -r requirements-tests.txt
+
+      - name: Make sure PyTorch version wasn't clobbered
+        run: |
+          python -c "import torch; assert '${{ matrix.pytorch_version }}' in torch.__version__"
+
+      - name: Ensure axolotl CLI was installed
+        run: |
+          axolotl --help
+
+      - name: Pre-Download dataset fixture
+        run: |
+          huggingface-cli download --repo-type=dataset axolotl-ai-internal/axolotl-oss-dataset-fixtures
+
+      - name: Run tests
+        run: |
+          pytest -v tests/conftest.py
+
+      - name: Upload coverage to Codecov
+        uses: codecov/codecov-action@v5
+        with:
+          token: ${{ secrets.CODECOV_TOKEN }}
+          files: ./coverage.xml
+          flags: unittests,pytorch-${{ matrix.pytorch_version }}
+          fail_ci_if_error: false
+
+      - name: cleanup pip cache
+        run: |
+          find "$(pip cache dir)/http-v2" -type f -mtime +14 -exec rm {} \;
+
+      - name: Save HF cache
+        id: hf-cache
+        uses: actions/cache/save@v4
+        with:
+          path: |
+            /home/runner/.cache/huggingface/hub/datasets--*
+            /home/runner/.cache/huggingface/hub/models--*
+          key: ${{ steps.hf-cache-restore.outputs.cache-primary-key }}
+
   pytest:
     name: PyTest
     runs-on: ubuntu-latest
+    needs: [preload-cache]
     strategy:
       fail-fast: false
       max-parallel: 2
@@ -33,11 +120,14 @@ jobs:
       - name: Check out repository code
         uses: actions/checkout@v4
 
-      - name: Restore Cache from S3
+      - name: Restore HF cache
-        id: hf-cache-restore-s3
+        id: hf-cache-restore
-        run: |
+        uses: actions/cache/restore@v4
-          mkdir -p /home/runner/.cache/huggingface/hub
+        with:
-          curl -L https://d1dttdx32dkk5p.cloudfront.net/hf-cache.tar.zst | tar -xf - -C /home/runner/.cache/huggingface/hub/  --use-compress-program unzstd
+          path: |
+            /home/runner/.cache/huggingface/hub/datasets--*
+            /home/runner/.cache/huggingface/hub/models--*
+          key: ${{ runner.os }}-hf-hub-cache-v2
 
       - name: Setup Python
         uses: actions/setup-python@v5
@@ -78,6 +168,10 @@ jobs:
         run: |
           axolotl --help
 
+      - name: Pre-Download dataset fixture
+        run: |
+          huggingface-cli download --repo-type=dataset axolotl-ai-internal/axolotl-oss-dataset-fixtures
+
       - name: Run tests
         run: |
           pytest -v -n8 --dist loadfile --ignore=tests/e2e/ --ignore=tests/patched/ --ignore=tests/cli/ tests/
@@ -99,8 +193,15 @@ jobs:
       fail-fast: false
       matrix:
         include:
-          - cuda: 126
+          - cuda: 124
-            cuda_version: 12.6.3
+            cuda_version: 12.4.1
+            python_version: "3.11"
+            pytorch: 2.5.1
+            num_gpus: 1
+            axolotl_extras:
+            nightly_build: "true"
+          - cuda: 124
+            cuda_version: 12.4.1
             python_version: "3.11"
             pytorch: 2.6.0
             num_gpus: 1

.github/workflows/tests.yml

@@ -195,12 +195,12 @@ jobs:
       fail-fast: false
       matrix:
         include:
-          - cuda: 126
+          - cuda: 124
-            cuda_version: 12.6.3
+            cuda_version: 12.4.1
             python_version: "3.11"
             pytorch: 2.6.0
             num_gpus: 1
-            axolotl_extras:
+            axolotl_extras: vllm
           - cuda: 126
             cuda_version: 12.6.3
             python_version: "3.11"
@@ -247,8 +247,8 @@ jobs:
       fail-fast: false
       matrix:
         include:
-          - cuda: 126
+          - cuda: 124
-            cuda_version: 12.6.3
+            cuda_version: 12.4.1
             python_version: "3.11"
             pytorch: 2.6.0
             num_gpus: 1
@@ -311,7 +311,7 @@ jobs:
             python_version: "3.11"
             pytorch: 2.6.0
             num_gpus: 1
-            axolotl_extras:
+            axolotl_extras: vllm
     steps:
       - name: Checkout
         uses: actions/checkout@v4

README.md

@@ -59,8 +59,6 @@ Features:
 
 ### Installation
 
-#### Using pip
-
 ```bash
 pip3 install -U packaging==23.2 setuptools==75.8.0 wheel ninja
 pip3 install --no-build-isolation axolotl[flash-attn,deepspeed]
@@ -70,13 +68,6 @@ axolotl fetch examples
 axolotl fetch deepspeed_configs  # OPTIONAL
 ```
 
-#### Using Docker
-
-Installing with Docker can be less error prone than installing in your own environment.
-```bash
-docker run --gpus '"all"' --rm -it axolotlai/axolotl:main-latest
-```
-
 Other installation approaches are described [here](https://docs.axolotl.ai/docs/installation.html).
 
 ### Your First Fine-tune

cicd/single_gpu.py

@@ -32,8 +32,6 @@ df_args = {
     "NIGHTLY_BUILD": os.environ.get("NIGHTLY_BUILD", ""),
     "CODECOV_TOKEN": os.environ.get("CODECOV_TOKEN", ""),
     "HF_HOME": "/workspace/data/huggingface-cache/hub",
-    "PYTHONUNBUFFERED": os.environ.get("PYTHONUNBUFFERED", "1"),
-    "DEEPSPEED_LOG_LEVEL": os.environ.get("DEEPSPEED_LOG_LEVEL", "WARNING"),
 }
 
 dockerfile_contents = df_template.render(**df_args)

docker/Dockerfile-base

@@ -22,11 +22,9 @@ RUN apt-get update \
     && mkdir /root/.conda \
     && bash Miniconda3-latest-Linux-x86_64.sh -b \
     && rm -f Miniconda3-latest-Linux-x86_64.sh \
-    && conda create -n "axolotl-py${PYTHON_VERSION}" python="${PYTHON_VERSION}" \
+    && conda create -n "py${PYTHON_VERSION}" python="${PYTHON_VERSION}"
-    && conda init bash \
-    && echo "conda activate axolotl-py${PYTHON_VERSION}" >> ~/.bashrc
 
-ENV PATH="/root/miniconda3/envs/axolotl-py${PYTHON_VERSION}/bin:${PATH}"
+ENV PATH="/root/miniconda3/envs/py${PYTHON_VERSION}/bin:${PATH}"
 
 WORKDIR /workspace
 
@@ -40,6 +38,6 @@ RUN git lfs install --skip-repo && \
     # The base image ships with `pydantic==1.8.2` which is not working
     pip3 install -U --no-cache-dir pydantic==1.10.10
 
-RUN if [ "$PYTORCH_VERSION" = "2.6.0" ] && [ "$CUDA" = "124" ] ; then \
+RUN if [ "$PYTORCH_VERSION" = "2.7.1" ] ; then \
-        FLASH_ATTENTION_FORCE_BUILD="TRUE" pip3 install --no-build-isolation flash-attn==2.8.0.post2; \
+        pip3 install flash-attn==2.7.4.post1; \
     fi

docker/Dockerfile-base-next

@@ -22,11 +22,9 @@ RUN apt-get update \
     && mkdir /root/.conda \
     && bash Miniconda3-latest-Linux-x86_64.sh -b \
     && rm -f Miniconda3-latest-Linux-x86_64.sh \
-    && conda create -n "axolotl-py${PYTHON_VERSION}" python="${PYTHON_VERSION}" \
+    && conda create -n "py${PYTHON_VERSION}" python="${PYTHON_VERSION}"
-    && conda init bash \
-    && echo "conda activate axolotl-py${PYTHON_VERSION}" >> ~/.bashrc
 
-ENV PATH="/root/miniconda3/envs/axolotl-py${PYTHON_VERSION}/bin:${PATH}"
+ENV PATH="/root/miniconda3/envs/py${PYTHON_VERSION}/bin:${PATH}"
 
 WORKDIR /workspace
 

docker/Dockerfile-base-nightly

@@ -22,11 +22,9 @@ RUN apt-get update \
     && mkdir /root/.conda \
     && bash Miniconda3-latest-Linux-x86_64.sh -b \
     && rm -f Miniconda3-latest-Linux-x86_64.sh \
-    && conda create -n "axolotl-py${PYTHON_VERSION}" python="${PYTHON_VERSION}" \
+    && conda create -n "py${PYTHON_VERSION}" python="${PYTHON_VERSION}"
-    && conda init bash \
-    && echo "conda activate axolotl-py${PYTHON_VERSION}" >> ~/.bashrc
 
-ENV PATH="/root/miniconda3/envs/axolotl-py${PYTHON_VERSION}/bin:${PATH}"
+ENV PATH="/root/miniconda3/envs/py${PYTHON_VERSION}/bin:${PATH}"
 
 WORKDIR /workspace
 

docker/Dockerfile-uv-base

@@ -34,3 +34,7 @@ RUN uv pip install packaging setuptools wheel psutil \
     && uv pip install --no-build-isolation "causal_conv1d @ git+https://github.com/Dao-AILab/causal-conv1d.git@main" \
     && uv pip install "mamba_ssm @ git+https://github.com/state-spaces/mamba.git@main" \
     && uv pip install awscli pydantic
+
+RUN if [ "$PYTORCH_VERSION" = "2.7.1" ] ; then \
+        uv pip install --no-build-isolation flash-attn==2.7.4.post1; \
+    fi

docs/docker.qmd

@@ -9,7 +9,7 @@ format:
 This section describes the different Docker images that are released by AxolotlAI at [Docker Hub](https://hub.docker.com/u/axolotlai).
 
 ::: {.callout-important}
-For Blackwell GPUs, please use the tags with PyTorch 2.7.1 and CUDA 12.8.
+For Blackwell GPUs, please use the tags with Pytorch 2.7.1 and CUDA 12.8.
 :::
 
 ## Base
@@ -34,7 +34,6 @@ Tags examples:
 
 - `main-base-py3.11-cu128-2.7.1`
 - `main-base-py3.11-cu126-2.7.1`
-- `main-base-py3.11-cu126-2.6.0`
 - `main-base-py3.11-cu124-2.6.0`
 - `main-base-py3.11-cu124-2.5.1`
 
@@ -74,15 +73,13 @@ There may be some extra tags appended to the image, like `-vllm` which installs
 
 Tags examples:
 
-- `main-py3.11-cu128-2.7.1`
+- `main-py3.11-cu126-2.7.0`
-- `main-py3.11-cu126-2.7.1`
-- `main-py3.11-cu126-2.6.0`
 - `main-py3.11-cu124-2.6.0`
 - `main-py3.11-cu124-2.5.1`
 - `main-latest`
 - `main-20250303-py3.11-cu124-2.6.0`
 - `main-20250303-py3.11-cu124-2.5.1`
-- `0.10.1`
+- `0.9.2`
 
 ## Cloud
 

docs/faq.qmd

@@ -51,10 +51,6 @@ description: Frequently asked questions
 >   pad_token: "..."
 > ```
 
-**Q: `IterableDataset error` or `KeyError: 'input_ids'` when using `preprocess` CLI**
-
-> A: This is because you may be using `preprocess` CLI with `pretraining_dataset:` or `skip_prepare_dataset: true` respectively. Please use `axolotl train` CLI directly instead as these datasets are prepared on demand.
-
 ### Chat templates
 
 **Q: `jinja2.exceptions.UndefinedError: 'dict object' has no attribute 'content' / 'role' / ____`**

requirements.txt

@@ -1,7 +1,7 @@
 --extra-index-url https://huggingface.github.io/autogptq-index/whl/cu118/
 
 # START section of dependencies that don't install on Darwin/MacOS
-bitsandbytes==0.46.0
+bitsandbytes==0.45.4
 triton>=3.0.0
 mamba-ssm==1.2.0.post1
 xformers>=0.0.23.post1
@@ -15,7 +15,7 @@ huggingface_hub==0.32.2
 peft==0.15.2
 transformers==4.52.4
 tokenizers>=0.21.1
-accelerate==1.8.1
+accelerate==1.7.0
 datasets==3.6.0
 deepspeed>=0.17.0
 trl==0.18.2
@@ -68,4 +68,4 @@ schedulefree==1.4.1
 axolotl-contribs-lgpl==0.0.6
 axolotl-contribs-mit==0.0.3
 
-mistral-common==1.6.3
+mistral-common==1.6.0

setup.py

@@ -111,9 +111,9 @@ def get_package_version():
 
 
 extras_require = {
-    "flash-attn": ["flash-attn==2.8.0.post2"],
+    "flash-attn": ["flash-attn==2.7.4.post1"],
     "ring-flash-attn": [
-        "flash-attn==2.8.0.post2",
+        "flash-attn==2.7.4.post1",
         "ring-flash-attn>=0.1.4",
         "yunchang==0.6.0",
     ],

src/axolotl/cli/preprocess.py

@@ -35,12 +35,6 @@ def do_preprocess(cfg: DictDefault, cli_args: PreprocessCliArgs) -> None:
     check_accelerate_default_config()
     check_user_token()
 
-    for key in ["skip_prepare_dataset", "pretraining_dataset"]:
-        if cfg.get("key"):
-            raise ValueError(
-                f"You have set `{key}:`. `preprocess` is not needed. Run the `axolotl train` CLI directly instead."
-            )
-
     if not cfg.dataset_prepared_path:
         msg = (
             Fore.RED

src/axolotl/common/datasets.py

@@ -75,17 +75,13 @@ def load_datasets(
 
         num_examples = cli_args.debug_num_examples if cli_args else 1
         text_only = cli_args.debug_text_only if cli_args else False
-        try:
+        train_samples = sample_dataset(train_dataset, num_examples)
-            train_samples = sample_dataset(train_dataset, num_examples)
+        check_dataset_labels(
-            check_dataset_labels(
+            train_samples,
-                train_samples,
+            tokenizer,
-                tokenizer,
+            num_examples=num_examples,
-                num_examples=num_examples,
+            text_only=text_only,
-                text_only=text_only,
+        )
-            )
-        except AttributeError:
-            # can't sample iterable datasets
-            pass
 
         LOG.info("printing prompters...")
         for prompter in prompters:

src/axolotl/core/builders/base.py

@@ -219,9 +219,7 @@ class TrainerBuilderBase(abc.ABC):
         if self.cfg.bf16 == "full":
             training_args_kwargs["bf16_full_eval"] = True
         else:
-            bf16 = self.cfg.bf16 or self.cfg.bfloat16
+            training_args_kwargs["bf16"] = self.cfg.bf16 or self.cfg.bfloat16
-            bf16 = bf16 if bf16 is not None else False
-            training_args_kwargs["bf16"] = bf16
 
     def _configure_scheduler(self, training_args_kwargs: dict):
         if self.cfg.lr_scheduler in ["one_cycle", "rex"]:

src/axolotl/core/builders/causal.py

@@ -253,10 +253,6 @@ class HFCausalTrainerBuilder(TrainerBuilderBase):
         training_arguments_kwargs["eval_sample_packing"] = bool(
             self.cfg.eval_sample_packing
         )
-        if self.cfg.sample_packing_sequentially is not None:
-            training_arguments_kwargs["sample_packing_sequentially"] = (
-                self.cfg.sample_packing_sequentially
-            )
         if self.cfg.sample_packing_bin_size is not None:
             training_arguments_kwargs["sample_packing_bin_size"] = (
                 self.cfg.sample_packing_bin_size
@@ -417,8 +413,7 @@ class HFCausalTrainerBuilder(TrainerBuilderBase):
                 or self.cfg.micro_batch_size > 1
             ):
                 return DataCollatorForSeq2Seq(self.tokenizer, **kwargs)
-            if not (self.cfg.sample_packing and self.cfg.pretrain_multipack_attn):
+            return None
-                return None
 
         if self.cfg.model_config_type == "mamba":
             return MambaDataCollator(tokenizer=self.tokenizer)

src/axolotl/core/trainers/dpo/__init__.py

@@ -28,7 +28,7 @@ class DPOStrategy:
         training_args_kwargs["max_completion_length"] = None
         training_args_kwargs["max_length"] = cfg.sequence_len
         training_args_kwargs["max_prompt_length"] = cfg.sequence_len
-        training_args_kwargs["generate_during_eval"] = cfg.dpo_generate_during_eval
+        training_args_kwargs["generate_during_eval"] = cfg.use_wandb
         if cfg.dpo_use_weighting is not None:
             training_args_kwargs["use_weighting"] = cfg.dpo_use_weighting
         if cfg.dpo_padding_free is not None:

src/axolotl/loaders/model.py

@@ -776,9 +776,6 @@ class ModelLoader:
         dist_dtype: torch.dtype,
         before_kbit_train_or_finetune: bool,
     ):
-        dest = {"dtype": dist_dtype}
-        if self.cfg.lora_on_cpu:
-            dest["device"] = "cpu"
         for name, module in self.model.named_modules():
             if "norm" in name:
                 module.to(dist_dtype)
@@ -789,4 +786,4 @@ class ModelLoader:
                     # don't upcast lm_head for btlm
                     continue
             if any(m in name for m in embedding_modules) and hasattr(module, "weight"):
-                module.to(**dest)
+                module.to(dist_dtype)

src/axolotl/loaders/patch_manager.py

@@ -65,7 +65,6 @@ class PatchManager:
         self._apply_mistral_cross_entropy_patch()
         self._apply_self_attention_lora_patch()
         self._apply_gemma3_conditional_generation_forward_patch()
-        self._apply_sequence_parallel_patches()
 
     def apply_post_model_load_patches(self, model: PreTrainedModel):
         """Apply patches that require the model instance."""
@@ -232,17 +231,6 @@ class PatchManager:
 
             patch_gemma3_conditional_generation_forward()
 
-    def _apply_sequence_parallel_patches(self):
-        """Apply sequence parallelism patches."""
-        if self.cfg.sequence_parallel_degree and self.cfg.sequence_parallel_degree > 1:
-            from axolotl.monkeypatch.ring_attn.patch import (
-                patch_prepare_data_loader,
-                patch_prepare_device_mesh,
-            )
-
-            patch_prepare_data_loader()
-            patch_prepare_device_mesh(self.cfg.sequence_parallel_degree, self.cfg.fsdp)
-
     def _patch_attention(self):
         """Apply attention-specific patches based on model type."""
         if not (self.cfg.flash_attention and hasattr(self.model_config, "model_type")):

src/axolotl/monkeypatch/data/batch_dataset_fetcher.py

@@ -9,6 +9,9 @@ from torch.utils.data._utils.worker import _worker_loop
 
 class _MapDatasetFetcher(_BaseDatasetFetcher):
     def fetch(self, possibly_batched_index):
+        if not possibly_batched_index:
+            return self.collate_fn([])
+
         if isinstance(possibly_batched_index[0], list):
             data = [None for i in possibly_batched_index]
             for i, possibly_batched_index_ in enumerate(possibly_batched_index):

src/axolotl/monkeypatch/multipack.py

@@ -42,10 +42,6 @@ def patch_for_multipack(model_type, model_name=None, has_remote_code=False):
     if has_remote_code:
         patch_remote(model_name)
     elif hasattr(transformers, "modeling_flash_attention_utils"):
-        # sanity check in case upstream api changes on this
-        assert hasattr(
-            transformers.modeling_flash_attention_utils, "_get_unpad_data"
-        ), "transformers api changed for _get_unpad_data for flash attention"
         transformers.modeling_flash_attention_utils._get_unpad_data = (  # pylint: disable=protected-access
             get_unpad_data
         )

src/axolotl/monkeypatch/ring_attn/patch.py

@@ -152,7 +152,7 @@ def update_ring_attn_params(position_ids: torch.Tensor | None):
 def patch_prepare_data_loader():
     """Patch `accelerate.data_loader.prepare_data_loader` to respect the SP degree.
 
-    Raises:
+    Raies:
         RuntimeError: If source code to patch does not exist.
     """
     original_fn = accelerate.data_loader.prepare_data_loader
@@ -168,34 +168,23 @@ def patch_prepare_data_loader():
         ORIGINAL_PREPARE_DATALOADER_CODE, NEW_PREPARE_DATALOADER_CODE
     )
 
-    items_to_import = []
-    for item in dir(accelerate.data_loader):
-        if item in patched_source:
-            items_to_import.append(item)
-
     # Create a new function from the patched source
     namespace = {}
     exec(  # pylint: disable=exec-used  # nosec B102
-        f"from accelerate.data_loader import ({', '.join(items_to_import)})",
+        patched_source, accelerate.data_loader.__dict__, namespace
-        globals(),
     )
-    exec(  # pylint: disable=exec-used  # nosec B102
-        patched_source, globals(), namespace
-    )
-
     patched_function = namespace["prepare_data_loader"]
-    original_fn.__code__ = patched_function.__code__
 
+    accelerate.data_loader.prepare_data_loader = patched_function
     LOG.info("Patched accelerate.data_loader.prepare_data_loader for SP support")
 
 
-def patch_prepare_device_mesh(sequence_parallel_degree: int, fsdp: bool = False):
+def patch_prepare_device_mesh(sequence_parallel_degree: int):
     """Patches the `Accelerator._prepare_device_mesh` method to create a device mesh
     that includes sequence parallelism with the specified degree.
 
     Args:
-        sequence_parallel_degree: The degree of sequence parallelism to use.
+        sequence_parallel_degree (int): The degree of sequence parallelism to use.
-        fsdp: Whether to use FSDP.
     """
 
     def _prepare_device_mesh(self):
@@ -218,14 +207,12 @@ def patch_prepare_device_mesh(sequence_parallel_degree: int, fsdp: bool = False)
         )
         device_ids = list(range(world_size))
 
-        # NOTE: We use "cp" instead of "sp" to match the PyTorch native "context
+        # Note that we use "cp" instead of "sp" to match the PyTorch native "context
-        # parallelism" implementation naming.
+        # parallelism" implementation naming
-        # NOTE: We have a simplified FSDP handling here; i.e., if FSDP is enabled, we
-        # only use "fsdp" and "cp" for the device mesh.
         return dist.DeviceMesh(
             "cuda",
             torch.tensor(device_ids).reshape(mesh_shape),
-            mesh_dim_names=("dp", "cp") if not fsdp else ("fsdp", "cp"),
+            mesh_dim_names=("dp", "cp"),
         )
 
     # Replace the original method with our new method

src/axolotl/prompt_strategies/chat_template.py

@@ -103,7 +103,6 @@ class ChatTemplatePrompter(Prompter):
         chat_template_kwargs = {
             "chat_template": self.chat_template,
             "add_generation_prompt": add_generation_prompt,
-            **self.chat_template_kwargs,
         }
 
         if tools:

src/axolotl/train.py

@@ -223,9 +223,6 @@ def execute_training(
             )
 
         LOG.info("Starting trainer...")
-        # TODO: disabling for now as not compatible with FSDP2 + torchao low bit optimizers
-        # if cfg.bf16:
-        #     torch.set_default_dtype(torch.bfloat16)
         trainer.train(resume_from_checkpoint=resume_from_checkpoint)
 
 

src/axolotl/utils/ctx_managers/sequence_parallel.py

@@ -12,6 +12,8 @@ from transformers.utils import ModelOutput
 
 from axolotl.monkeypatch.ring_attn import (
     get_ring_attn_group,
+    patch_prepare_data_loader,
+    patch_prepare_device_mesh,
     register_ring_attn,
     update_ring_attn_params,
 )
@@ -236,6 +238,12 @@ class SequenceParallelContextManager:
             ring_attn_func=self.ring_attn_func,
         )
 
+        # Patches for accelerate functionality
+        patch_prepare_data_loader()
+        patch_prepare_device_mesh(
+            sequence_parallel_degree=self.sequence_parallel_degree
+        )
+
     def _register_model_hooks(self):
         # Forward pre-hook to apply sequence parallelism
         def sequence_parallel_pre_hook(_, args, kwargs):

src/axolotl/utils/data/pretraining.py

@@ -224,10 +224,10 @@ def wrap_pretraining_dataset(
     remove_columns = []
     if dataset.features is None:
         for first_row in dataset:
-            remove_columns = list(first_row.keys())
+            remove_columns = first_row.keys()
             break
     else:
-        remove_columns = list(dataset.features.keys())
+        remove_columns = dataset.features.keys()
 
     dataset = dataset.map(
         encode,
@@ -267,7 +267,6 @@ def encode_packed_pretraining(
         batch_size=1,
         batch_max_len=batch_size * max_seq_length,
         drop_last=True,
-        num_processes=1,
     )
 
     chunked_data = defaultdict(list)

src/axolotl/utils/data/sft.py

@@ -334,10 +334,7 @@ def _load_raw_datasets(
     dataset = merge_datasets(datasets, cfg)
 
     if not cfg.skip_prepare_dataset:
-        if split == "test" and cfg.eval_sequence_len:
+        dataset = drop_long_seq_in_dataset(dataset, cfg)
-            dataset = drop_long_seq_in_dataset(dataset, cfg.eval_sequence_len, cfg)
-        else:
-            dataset = drop_long_seq_in_dataset(dataset, cfg.sequence_len, cfg)
         if cfg.sample_packing:
             dataset, _ = process_datasets_for_packing(cfg, dataset, None)
 

src/axolotl/utils/data/shared.py

@@ -524,25 +524,13 @@ def merge_datasets(datasets: list[Dataset], cfg: DictDefault) -> Dataset:
         Merged dataset.
     """
     if len(datasets) == 1:
-        ds = datasets[0]
+        return datasets[0]
-
-        # Do not shuffle if curriculum sampling is enabled or
-        # shuffle_merged_datasets is disabled
-        if cfg.curriculum_sampling or not cfg.shuffle_merged_datasets:
-            return ds
-
-        return ds.shuffle(seed=cfg.seed)
 
     LOG.info("Merging datasets...")
     merged_dataset = concatenate_datasets(datasets)
 
     if cfg.shuffle_merged_datasets:
         LOG.debug("Shuffling merged datasets...")
-        if cfg.curriculum_sampling:
-            LOG.warning(
-                "Shuffling merged datasets with curriculum sampling is not recommended. "
-                "This will randomize the order of samples."
-            )
         merged_dataset = merged_dataset.shuffle(seed=cfg.seed)
     else:
         LOG.debug("Not shuffling merged datasets.")

src/axolotl/utils/data/utils.py

@@ -148,14 +148,11 @@ def deduplicate_and_log_datasets(
     return dataset, other_dataset
 
 
-def drop_long_seq_in_dataset(
+def drop_long_seq_in_dataset(dataset: Dataset, cfg: DictDefault) -> Dataset:
-    dataset: Dataset, sequence_len: int, cfg: DictDefault
-) -> Dataset:
     """Remove sequences longer than configured maximum from dataset.
 
     Args:
         dataset: Dataset to filter.
-        sequence_len: Maximum length for sequences to keep
         cfg: Dictionary mapping `axolotl` config keys to values.
 
     Returns:
@@ -170,7 +167,7 @@ def drop_long_seq_in_dataset(
 
     drop_long = functools.partial(
         drop_long_seq,
-        sequence_len=sequence_len,
+        sequence_len=cfg.sequence_len,
         min_sequence_len=cfg.min_sample_len,
     )
 
@@ -190,7 +187,7 @@ def drop_long_seq_in_dataset(
 
     drop_long_kwargs = {}
     if filter_map_kwargs:
-        drop_long_kwargs["desc"] = f"Dropping Long Sequences (>{sequence_len})"
+        drop_long_kwargs["desc"] = "Dropping Long Sequences"
 
     dataset = dataset.filter(
         drop_long,

src/axolotl/utils/mistral_tokenizer.py

@@ -8,7 +8,7 @@ from typing import TYPE_CHECKING, Optional
 import numpy as np
 from huggingface_hub import hf_hub_download
 from mistral_common.tokens.tokenizers.mistral import MistralTokenizer
-from mistral_common.tokens.tokenizers.tekken import SpecialTokenPolicy, Tekkenizer
+from mistral_common.tokens.tokenizers.tekken import Tekkenizer
 from torch import Tensor
 from transformers.utils import PaddingStrategy
 
@@ -251,13 +251,10 @@ class HFMistralTokenizer:
             token_ids = [token_ids]
 
         if skip_special_tokens:
-            return self._mistral.instruct_tokenizer.tokenizer.decode(
+            return self._mistral.instruct_tokenizer.tokenizer.decode(token_ids)
-                token_ids, special_token_policy=SpecialTokenPolicy.IGNORE
-            )
 
-        return self._mistral.instruct_tokenizer.tokenizer.decode(
+        # to_string returns a string with special tokens
-            token_ids, special_token_policy=SpecialTokenPolicy.KEEP
+        return self._mistral.instruct_tokenizer.tokenizer.to_string(token_ids)
-        )
 
     def _create_mistral_chat_completion_request(
         self, conversation: list[dict], tools: list[dict] | None = None

src/axolotl/utils/samplers/multipack.py

@@ -260,7 +260,7 @@ class MultipackBatchSampler(BatchSampler):
         lengths: np.ndarray,  # Sequence lengths
         packing_efficiency_estimate: float = 1.0,  # Initial efficiency estimate
         drop_last: bool = True,  # Whether to drop final batches (might be incomplete)
-        num_count_samples: int = 4,  # Number of times to estimate batch count
+        num_count_samples: int = 8,  # Number of times to estimate batch count
         sequential: bool = False,  # Whether to use sequential packing
         group_size: int = 100_000,  # Size of groups for parallel packing
         bin_size: int = 200,  # The max number of samples that can be packed in a single bin
@@ -335,13 +335,12 @@ class MultipackBatchSampler(BatchSampler):
             bins = [[indices[b_idx] for b_idx in bin_indices] for bin_indices in bins]
         else:
             # Use parallel packing
-            num_processes = self.num_processes or 1
             all_bins = pack_parallel(
                 lengths,
                 bin_capacity=self.batch_max_len,
                 group_size=self.group_size,
                 bin_size=self.bin_size,
-                num_processes=min(4, num_processes) if num_processes else 4,
+                num_processes=max(4, self.num_processes) if self.num_processes else 4,
                 safe_mode=self.safe_mode,
                 mp_start_method=self.mp_start_method,
             )

src/axolotl/utils/schemas/config.py

@@ -146,7 +146,6 @@ class AxolotlInputConfig(
     dpo_label_smoothing: float | None = None
     dpo_norm_loss: bool | None = None
     dpo_padding_free: bool | None = None
-    dpo_generate_during_eval: bool | None = None
 
     datasets: (
         Annotated[
@@ -367,12 +366,6 @@ class AxolotlInputConfig(
             "description": "The maximum length of an input to train with, this should typically be less than 2048 as most models have a token/context limit of 2048"
         },
     )
-    eval_sequence_len: int | None = Field(
-        default=None,
-        json_schema_extra={
-            "description": "The maximum length of an input for evaluation. If not specified, defaults to sequence_len"
-        },
-    )
     min_sample_len: int | None = None
     max_prompt_len: int = Field(
         default=512,
@@ -785,12 +778,6 @@ class AxolotlInputConfig(
             "description": "Custom jinja template for chat template. This will be only used if chat_template is set to `jinja` or `null` (in which case chat_template is automatically set to `jinja`). Default is null."
         },
     )
-    chat_template_kwargs: dict[str, Any] | None = Field(
-        default=None,
-        json_schema_extra={
-            "description": "Additional kwargs to pass to the chat template. This is useful for customizing the chat template. For example, you can pass `thinking=False` to add a generation prompt to the chat template."
-        },
-    )
     eot_tokens: list[str] | None = Field(
         default=None,
         json_schema_extra={

src/axolotl/utils/schemas/validation.py

@@ -462,20 +462,6 @@ class TrainingValidationMixin:
 
         return data
 
-    @model_validator(mode="before")
-    @classmethod
-    def pretrain_with_tps(cls, data):
-        if data.get("pretraining_dataset") and data.get(
-            "include_tokens_per_second", False
-        ):
-            # combining these would raise `TypeError: cannot pickle 'dict_keys' object`
-            # due to trying to count the number of tokens total in the dataset
-            raise ValueError(
-                "pretraining_dataset and include_tokens_per_second cannot be used together."
-            )
-
-        return data
-
 
 class LoRAValidationMixin:
     """Validation methods related to LoRA/QLoRA configuration."""

src/axolotl/utils/trainer.py

@@ -381,7 +381,6 @@ def process_pretraining_datasets_for_packing(
     if not skip_position_ids:
         train_dataset = train_dataset.map(
             add_position_ids,
-            batched=True,
             desc="Add position_id column (Pretraining Sample Packing)",
         )
     if drop_attention_mask:
@@ -609,9 +608,6 @@ def prepare_opinionated_env(cfg):
     if cfg.qlora_sharded_model_loading:
         # model loading is forked after the tokenizer
         os.environ["TOKENIZERS_PARALLELISM"] = "false"
-    if cfg.sample_packing:
-        # multipack parallel packing sampler defaults to using fork
-        os.environ["TOKENIZERS_PARALLELISM"] = "false"
 
 
 def setup_trainer(

tests/conftest.py

@@ -4,14 +4,12 @@ shared pytest fixtures
 
 import functools
 import importlib
-import logging
 import os
 import shutil
 import sys
 import tempfile
 import time
-from pathlib import Path, PosixPath
+from pathlib import Path
-from typing import Generator
 
 import datasets
 import pytest
@@ -26,8 +24,6 @@ from tests.hf_offline_utils import (
     hf_offline_context,
 )
 
-logging.getLogger("filelock").setLevel(logging.CRITICAL)
-
 
 def retry_on_request_exceptions(max_retries=3, delay=1):
     # pylint: disable=duplicate-code
@@ -415,7 +411,7 @@ def tokenizer_mistral_7b_instruct_chatml(tokenizer_mistral_7b_instruct):
 
 
 @pytest.fixture
-def temp_dir() -> Generator[str, None, None]:
+def temp_dir():
     # Create a temporary directory
     _temp_dir = tempfile.mkdtemp()
     yield _temp_dir
@@ -423,11 +419,6 @@ def temp_dir() -> Generator[str, None, None]:
     shutil.rmtree(_temp_dir)
 
 
-@pytest.fixture(scope="function", autouse=True)
-def unique_triton_cache_dir(temp_dir: str | PosixPath) -> None:
-    os.environ["TRITON_CACHE_DIR"] = str(temp_dir) + "/.triton/cache"
-
-
 @pytest.fixture(scope="function", autouse=True)
 def cleanup_monkeypatches():
     from transformers import Trainer

tests/e2e/multigpu/patched/test_sp.py

@@ -54,7 +54,6 @@ class TestSequenceParallelism:
                 "micro_batch_size": micro_batch_size,
                 "gradient_accumulation_steps": 2,
                 "output_dir": temp_dir,
-                "dataset_prepared_path": temp_dir + "/last_run_prepared",
                 "learning_rate": 0.00001,
                 "optimizer": "adamw_8bit",
                 "lr_scheduler": "cosine",

tests/e2e/multigpu/solo/test_flex.py

@@ -54,7 +54,6 @@ class TestPackedFlex:
                 "gradient_accumulation_steps": 2,
                 "gradient_checkpointing": True,
                 "output_dir": temp_dir,
-                "dataset_prepared_path": temp_dir + "/last_run_prepared",
                 "learning_rate": 0.00001,
                 "optimizer": "adamw_torch_fused",
                 "lr_scheduler": "cosine",

tests/e2e/multigpu/solo/test_grpo.py

@@ -309,7 +309,6 @@ def oai_gsm8k_transform(cfg, *args, **kwargs):
                 "warmup_steps": 10,
                 "val_set_size": 0.0,
                 "output_dir": temp_dir,
-                "dataset_prepared_path": temp_dir + "/last_run_prepared",
                 "learning_rate": 0.0001,
                 "optimizer": "adamw_torch_fused",
                 "lr_scheduler": "cosine",
@@ -401,7 +400,6 @@ def oai_gsm8k_transform(cfg, *args, **kwargs):
                 "warmup_steps": 10,
                 "val_set_size": 0.0,
                 "output_dir": temp_dir,
-                "dataset_prepared_path": temp_dir + "/last_run_prepared",
                 "learning_rate": 0.0001,
                 "optimizer": "adamw_torch_fused",
                 "lr_scheduler": "cosine",

tests/e2e/multigpu/test_eval.py

@@ -38,13 +38,12 @@ class TestMultiGPUEval:
                 "lora_dropout": 0.05,
                 "lora_target_linear": True,
                 "lora_modules_to_save": ["embed_tokens", "lm_head"],
-                "val_set_size": 0.05,
+                "val_set_size": 0.004,
                 "special_tokens": {"pad_token": "<|endoftext|>"},
                 "datasets": [
                     {
                         "path": "teknium/GPT4-LLM-Cleaned",
                         "type": "alpaca",
-                        "split": "train[:5%]",
                     },
                 ],
                 "num_epochs": 1,
@@ -52,7 +51,6 @@ class TestMultiGPUEval:
                 "micro_batch_size": 2,
                 "gradient_accumulation_steps": 2,
                 "output_dir": temp_dir,
-                "dataset_prepared_path": temp_dir + "/last_run_prepared",
                 "learning_rate": 0.00001,
                 "optimizer": "adamw_8bit",
                 "lr_scheduler": "cosine",
@@ -109,13 +107,12 @@ class TestMultiGPUEval:
                 "lora_dropout": 0.05,
                 "lora_target_linear": True,
                 "lora_modules_to_save": ["embed_tokens", "lm_head"],
-                "val_set_size": 0.01,
+                "val_set_size": 0.0004,
                 "special_tokens": {"pad_token": "<|endoftext|>"},
                 "datasets": [
                     {
                         "path": "teknium/GPT4-LLM-Cleaned",
                         "type": "alpaca",
-                        "split": "train[:5%]",
                     },
                 ],
                 "num_epochs": 1,
@@ -123,7 +120,6 @@ class TestMultiGPUEval:
                 "micro_batch_size": 2,
                 "gradient_accumulation_steps": 2,
                 "output_dir": temp_dir,
-                "dataset_prepared_path": temp_dir + "/last_run_prepared",
                 "learning_rate": 0.00001,
                 "optimizer": "adamw_8bit",
                 "lr_scheduler": "cosine",

tests/e2e/multigpu/test_gemma3.py

@@ -64,7 +64,6 @@ class TestMultiGPUGemma3:
                 },
                 "gradient_accumulation_steps": 2,
                 "output_dir": temp_dir,
-                "dataset_prepared_path": temp_dir + "/last_run_prepared",
                 "learning_rate": 0.0001,
                 "optimizer": "adamw_8bit",
                 "lr_scheduler": "cosine",

tests/e2e/multigpu/test_llama.py

@@ -62,7 +62,6 @@ class TestMultiGPULlama:
                 "gradient_accumulation_steps": 2,
                 # "gradient_checkpointing": True,
                 "output_dir": temp_dir,
-                "dataset_prepared_path": temp_dir + "/last_run_prepared",
                 "learning_rate": 0.00001,
                 "optimizer": "adamw_8bit",
                 "lr_scheduler": "cosine",
@@ -128,7 +127,6 @@ class TestMultiGPULlama:
                 "gradient_accumulation_steps": gradient_accumulation_steps,
                 # "gradient_checkpointing": True,
                 "output_dir": temp_dir,
-                "dataset_prepared_path": temp_dir + "/last_run_prepared",
                 "learning_rate": 0.00001,
                 "optimizer": "adamw_8bit",
                 "lr_scheduler": "cosine",
@@ -202,7 +200,6 @@ class TestMultiGPULlama:
                 "gradient_accumulation_steps": 2,
                 # "gradient_checkpointing": True,
                 "output_dir": temp_dir,
-                "dataset_prepared_path": temp_dir + "/last_run_prepared",
                 "warmup_steps": 0,
                 "learning_rate": 0.00001,
                 "optimizer": "adamw_8bit",
@@ -281,7 +278,6 @@ class TestMultiGPULlama:
                 "gradient_accumulation_steps": 2,
                 # "gradient_checkpointing": True,
                 "output_dir": temp_dir,
-                "dataset_prepared_path": temp_dir + "/last_run_prepared",
                 "warmup_steps": 0,
                 "learning_rate": 0.00001,
                 "optimizer": "adamw_8bit",
@@ -344,7 +340,6 @@ class TestMultiGPULlama:
                 "gradient_accumulation_steps": gradient_accumulation_steps,
                 # "gradient_checkpointing": True,
                 "output_dir": temp_dir,
-                "dataset_prepared_path": temp_dir + "/last_run_prepared",
                 "learning_rate": 0.00001,
                 "optimizer": "adamw_torch_fused",
                 "lr_scheduler": "cosine",
@@ -417,7 +412,6 @@ class TestMultiGPULlama:
                 "gradient_accumulation_steps": 2,
                 # "gradient_checkpointing": True,
                 "output_dir": temp_dir,
-                "dataset_prepared_path": temp_dir + "/last_run_prepared",
                 "learning_rate": 0.00001,
                 "optimizer": "adamw_torch_fused",
                 "lr_scheduler": "cosine",
@@ -497,7 +491,6 @@ class TestMultiGPULlama:
                 "gradient_accumulation_steps": 2,
                 "gradient_checkpointing": True,
                 "output_dir": temp_dir,
-                "dataset_prepared_path": temp_dir + "/last_run_prepared",
                 "learning_rate": 0.00001,
                 "optimizer": "adamw_torch_8bit",
                 "lr_scheduler": "cosine",
@@ -580,7 +573,6 @@ class TestMultiGPULlama:
                 "gradient_accumulation_steps": 2,
                 # "gradient_checkpointing": True,
                 "output_dir": temp_dir,
-                "dataset_prepared_path": temp_dir + "/last_run_prepared",
                 "learning_rate": 0.00001,
                 "optimizer": "adamw_torch_fused",
                 "lr_scheduler": "cosine",
@@ -677,7 +669,6 @@ class TestMultiGPULlama:
                 "micro_batch_size": 1,
                 "gradient_accumulation_steps": gradient_accumulation_steps,
                 "output_dir": temp_dir,
-                "dataset_prepared_path": temp_dir + "/last_run_prepared",
                 "learning_rate": 0.00001,
                 "optimizer": "adamw_torch_fused",
                 "lr_scheduler": "cosine",
@@ -752,7 +743,6 @@ class TestMultiGPULlama:
                 "micro_batch_size": 1,
                 "gradient_accumulation_steps": gradient_accumulation_steps,
                 "output_dir": temp_dir,
-                "dataset_prepared_path": temp_dir + "/last_run_prepared",
                 "learning_rate": 0.00001,
                 "optimizer": "adamw_torch_fused",
                 "lr_scheduler": "cosine",
@@ -827,7 +817,6 @@ class TestMultiGPULlama:
                 "micro_batch_size": 1,
                 "gradient_accumulation_steps": gradient_accumulation_steps,
                 "output_dir": temp_dir,
-                "dataset_prepared_path": temp_dir + "/last_run_prepared",
                 "learning_rate": 0.00001,
                 "optimizer": "adamw_torch_fused",
                 "lr_scheduler": "cosine",

tests/e2e/multigpu/test_qwen2.py

@@ -46,7 +46,6 @@ class TestMultiGPUQwen2:
                 "micro_batch_size": 2,
                 "gradient_accumulation_steps": 2,
                 "output_dir": temp_dir,
-                "dataset_prepared_path": temp_dir + "/last_run_prepared",
                 "learning_rate": 0.00001,
                 "optimizer": "adamw_torch_fused",
                 "lr_scheduler": "cosine",

tests/e2e/multigpu/test_ray.py

@@ -48,7 +48,6 @@ class TestMultiGPURay:
                 "micro_batch_size": 4,
                 "gradient_accumulation_steps": 2,
                 "output_dir": temp_dir,
-                "dataset_prepared_path": temp_dir + "/last_run_prepared",
                 "learning_rate": 0.00001,
                 "optimizer": "adamw_8bit",
                 "lr_scheduler": "cosine",
@@ -108,7 +107,6 @@ class TestMultiGPURay:
                 "micro_batch_size": 1,
                 "gradient_accumulation_steps": gradient_accumulation_steps,
                 "output_dir": temp_dir,
-                "dataset_prepared_path": temp_dir + "/last_run_prepared",
                 "learning_rate": 0.00001,
                 "optimizer": "adamw_torch",
                 "lr_scheduler": "cosine",

tests/e2e/patched/lora_kernels/test_lora_kernel_patching.py

@@ -396,7 +396,7 @@ def test_model_architecture(model_config):
 
 
 # pylint: disable=duplicate-code
-def test_kernel_training_integration(temp_dir):
+def test_kernel_training_integration():
     """Test model loading with kernel patches enabled."""
     from axolotl.cli.utils import load_model_and_tokenizer
 
@@ -426,14 +426,6 @@ def test_kernel_training_integration(temp_dir):
         }
     )
 
-    # Write cfg to yaml file
-    path = Path(temp_dir) / "config.yaml"
-    with open(path, "w", encoding="utf-8") as fout:
-        fout.write(yaml.dump(cfg.to_dict(), Dumper=yaml.Dumper))
-
-    # Load config
-    cfg = load_cfg(str(path))
-
     # Load model
     model, _, _ = load_model_and_tokenizer(cfg=cfg)
 
@@ -513,7 +505,7 @@ def test_kernel_training_integration_auto_enable(temp_dir):
     assert found_patched_attn
 
 
-def test_kernel_training_integration_dropout_non_zero(temp_dir):
+def test_kernel_training_integration_dropout_non_zero():
     """Test model loading with dropout non-zero should not patch."""
 
     from axolotl.cli.utils import load_model_and_tokenizer
@@ -541,14 +533,6 @@ def test_kernel_training_integration_dropout_non_zero(temp_dir):
         }
     )
 
-    # Write cfg to yaml file
-    path = Path(temp_dir) / "config.yaml"
-    with open(path, "w", encoding="utf-8") as fout:
-        fout.write(yaml.dump(cfg.to_dict(), Dumper=yaml.Dumper))
-
-    # Load config
-    cfg = load_cfg(str(path))
-
     # Get original attention class
     attention_cls = get_attention_cls_from_config(cfg)
 

tests/test_packed_batch_sampler.py

@@ -70,7 +70,7 @@ class TestBatchedSamplerPacking:
         )
         train_dataset = concatenate_datasets([dataset_wrapper])
 
-        train_dataset = drop_long_seq_in_dataset(train_dataset, cfg.sequence_len, cfg)
+        train_dataset = drop_long_seq_in_dataset(train_dataset, cfg)
 
         lengths = get_dataset_lengths(train_dataset)
         batch_sampler = MultipackBatchSampler(