set TORCH_CUDA_ARCH_LIST correctly

.github/workflows/base.yml

@@ -60,13 +60,6 @@ jobs:
             pytorch: 2.9.0
             torch_cuda_arch_list: "7.0 7.5 8.0 8.6 8.7 8.9 9.0+PTX"
             dockerfile: "Dockerfile-base"
-          - cuda: "130"
-            cuda_version: 13.0.0
-            cudnn_version: ""
-            python_version: "3.11"
-            pytorch: 2.9.0
-            torch_cuda_arch_list: "9.0+PTX"
-            dockerfile: "Dockerfile-base"
 #          - cuda: "128"
 #            cuda_version: 12.8.1
 #            cudnn_version: ""
@@ -150,13 +143,6 @@ jobs:
             pytorch: 2.9.0
             torch_cuda_arch_list: "7.0 7.5 8.0 8.6 8.7 8.9 9.0+PTX"
             dockerfile: "Dockerfile-uv-base"
-          - cuda: "130"
-            cuda_version: 13.0.0
-            cudnn_version: ""
-            python_version: "3.11"
-            pytorch: 2.9.0
-            torch_cuda_arch_list: "9.0+PTX"
-            dockerfile: "Dockerfile-uv-base"
     steps:
       - name: Checkout
         uses: actions/checkout@v4

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

@@ -40,13 +40,6 @@ jobs:
             axolotl_extras: fbgemm-gpu
             num_gpus: 2
             nightly_build: "true"
-          - cuda: 128
-            cuda_version: 12.8.1
-            python_version: "3.11"
-            pytorch: 2.9.0
-            axolotl_extras: fbgemm-gpu
-            num_gpus: 2
-            nightly_build: "true"
     runs-on: [self-hosted, modal]
     timeout-minutes: 120
     steps:

.github/workflows/precommit-autoupdate.yml

@@ -2,7 +2,7 @@ name: Pre-commit auto-update
 
 on:
   schedule:
-    - cron: '0 0 1 * *'  # Run monthly
+    - cron: '0 0 * * 0'  # Run weekly
   workflow_dispatch:  # Manual kickoff
 
 jobs:

.github/workflows/tests.yml

@@ -55,7 +55,7 @@ jobs:
       fail-fast: false
       matrix:
         python_version: ["3.11"]
-        pytorch_version: ["2.7.1", "2.8.0", "2.9.0"]
+        pytorch_version: ["2.7.1", "2.8.0"]
     timeout-minutes: 20
 
     steps:
@@ -130,7 +130,7 @@ jobs:
       fail-fast: false
       matrix:
         python_version: ["3.11"]
-        pytorch_version: ["2.7.1", "2.8.0", "2.9.0"]
+        pytorch_version: ["2.7.1", "2.8.0"]
     timeout-minutes: 20
 
     steps:
@@ -152,7 +152,7 @@ jobs:
       - name: upgrade pip
         run: |
           pip3 install --upgrade pip
-          pip3 install --upgrade packaging==23.2 setuptools==75.8.0 setuptools_scm build wheel psutil
+          pip3 install --upgrade packaging==23.2 setuptools==75.8.0 setuptools_scm build wheel
 
       - name: Install PyTorch
         run: |
@@ -231,10 +231,16 @@ jobs:
       fail-fast: false
       matrix:
         include:
-          - cuda: 128
-            cuda_version: 12.8.1
+          - cuda: 126
+            cuda_version: 12.6.3
             python_version: "3.11"
-            pytorch: 2.8.0
+            pytorch: 2.7.1
+            num_gpus: 1
+            axolotl_extras:
+          - cuda: 126
+            cuda_version: 12.6.3
+            python_version: "3.11"
+            pytorch: 2.7.1
             num_gpus: 1
             axolotl_extras:
             dockerfile: "Dockerfile-uv.jinja"
@@ -280,18 +286,12 @@ jobs:
       fail-fast: false
       matrix:
         include:
-          - cuda: 126
-            cuda_version: 12.6.3
+          - cuda: 128
+            cuda_version: 12.8.1
             python_version: "3.11"
             pytorch: 2.7.1
             num_gpus: 1
             axolotl_extras:
-#          - cuda: 128
-#            cuda_version: 12.8.1
-#            python_version: "3.11"
-#            pytorch: 2.7.1
-#            num_gpus: 1
-#            axolotl_extras:
           - cuda: 128
             cuda_version: 12.8.1
             python_version: "3.11"
@@ -299,12 +299,6 @@ jobs:
             num_gpus: 1
             gpu_type: "B200"
             axolotl_extras: fbgemm-gpu
-          - cuda: 128
-            cuda_version: 12.8.1
-            python_version: "3.11"
-            pytorch: 2.9.0
-            num_gpus: 1
-            axolotl_extras:
     steps:
       - name: Checkout
         uses: actions/checkout@v4

.pre-commit-config.yaml

@@ -11,7 +11,7 @@ repos:
     -   id: no-commit-to-branch
         args: ['--branch', 'main']
 -   repo: https://github.com/astral-sh/ruff-pre-commit
-    rev: v0.14.3
+    rev: v0.14.0
     hooks:
     -   id: ruff
         args: [--fix]

cicd/Dockerfile.jinja

@@ -32,7 +32,7 @@ RUN if [ "$NIGHTLY_BUILD" = "true" ] ; then \
         sed -i 's#^datasets.*#datasets @ git+https://github.com/huggingface/datasets.git@main#' requirements.txt; \
     fi
 
-RUN pip install packaging==23.2 setuptools==75.8.0 psutil
+RUN pip install packaging==23.2 setuptools==75.8.0
 RUN if [ "$AXOLOTL_EXTRAS" != "" ] ; then \
         pip install --no-build-isolation -e .[deepspeed,flash-attn,ring-flash-attn,optimizers,ray,$AXOLOTL_EXTRAS] $AXOLOTL_ARGS; \
     else \

docker/Dockerfile-base

@@ -5,7 +5,7 @@ ARG MAX_JOBS=4
 
 FROM nvidia/cuda:$CUDA_VERSION-cudnn$CUDNN_VERSION-devel-ubuntu$UBUNTU_VERSION AS base-builder
 
-ENV PATH="/workspace/miniconda3/bin:${PATH}"
+ENV PATH="/root/miniconda3/bin:${PATH}"
 
 ARG PYTHON_VERSION="3.10"
 ARG PYTORCH_VERSION="2.1.2"
@@ -24,27 +24,23 @@ RUN apt-get update \
     && rm -rf /var/lib/apt/lists/* \
     && wget \
     https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh \
-    && mkdir -p /workspace/.conda \
-    && bash Miniconda3-latest-Linux-x86_64.sh -b -p /workspace/miniconda3 \
+    && mkdir /root/.conda \
+    && bash Miniconda3-latest-Linux-x86_64.sh -b \
     && rm -f Miniconda3-latest-Linux-x86_64.sh \
     && conda tos accept --override-channels --channel https://repo.anaconda.com/pkgs/main \
     && conda tos accept --override-channels --channel https://repo.anaconda.com/pkgs/r \
     && conda create -n "py${PYTHON_VERSION}" python="${PYTHON_VERSION}"
 
-ENV PATH="/workspace/miniconda3/envs/py${PYTHON_VERSION}/bin:${PATH}"
+ENV PATH="/root/miniconda3/envs/py${PYTHON_VERSION}/bin:${PATH}"
 
 WORKDIR /workspace
 
-RUN python3 -m pip install --upgrade pip && pip3 install -U packaging==23.2 setuptools==75.8.0 wheel psutil && \
+RUN python3 -m pip install --upgrade pip && pip3 install -U packaging==23.2 setuptools==75.8.0 wheel && \
     python3 -m pip install --no-cache-dir -U torch==${PYTORCH_VERSION}+cu${CUDA} torchvision --extra-index-url https://download.pytorch.org/whl/cu$CUDA && \
+    CAUSAL_CONV1D_FORCE_CXX11_ABI=TRUE CAUSAL_CONV1D_FORCE_BUILD=TRUE python3 -m pip install --no-cache-dir causal_conv1d==1.5.2 && \
+    python3 -m pip install --no-cache-dir "mamba_ssm @ git+https://github.com/state-spaces/mamba.git@main" && \
     python3 -m pip cache purge
 
-RUN if [ "$CUDA" != "130" ] ; then \
-        CAUSAL_CONV1D_FORCE_CXX11_ABI=TRUE CAUSAL_CONV1D_FORCE_BUILD=TRUE python3 -m pip install --no-cache-dir "causal_conv1d @ git+https://github.com/Dao-AILab/causal-conv1d.git@v1.5.4"; \
-        python3 -m pip install --no-cache-dir "mamba_ssm @ git+https://github.com/state-spaces/mamba.git@main"; \
-        python3 -m pip cache purge; \
-    fi
-
 RUN git lfs install --skip-repo && \
     pip3 install awscli && \
     # The base image ships with `pydantic==1.8.2` which is not working

docker/Dockerfile-base-next

@@ -5,7 +5,7 @@ ARG MAX_JOBS=4
 
 FROM nvidia/cuda:$CUDA_VERSION-cudnn$CUDNN_VERSION-devel-ubuntu$UBUNTU_VERSION AS base-builder
 
-ENV PATH="/workspace/miniconda3/bin:${PATH}"
+ENV PATH="/root/miniconda3/bin:${PATH}"
 
 ARG PYTHON_VERSION="3.11"
 ARG PYTORCH_VERSION="next"
@@ -19,12 +19,12 @@ RUN apt-get update \
     && apt-get install -y wget git build-essential ninja-build git-lfs libaio-dev pkg-config && rm -rf /var/lib/apt/lists/* \
     && wget \
     https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh \
-    && mkdir -p /workspace/.conda \
-    && bash Miniconda3-latest-Linux-x86_64.sh -b -p /workspace/miniconda3 \
+    && mkdir /root/.conda \
+    && bash Miniconda3-latest-Linux-x86_64.sh -b \
     && rm -f Miniconda3-latest-Linux-x86_64.sh \
     && conda create -n "py${PYTHON_VERSION}" python="${PYTHON_VERSION}"
 
-ENV PATH="/workspace/miniconda3/envs/py${PYTHON_VERSION}/bin:${PATH}"
+ENV PATH="/root/miniconda3/envs/py${PYTHON_VERSION}/bin:${PATH}"
 
 WORKDIR /workspace
 

docker/Dockerfile-base-nightly

@@ -5,7 +5,7 @@ ARG MAX_JOBS=4
 
 FROM nvidia/cuda:$CUDA_VERSION-cudnn$CUDNN_VERSION-devel-ubuntu$UBUNTU_VERSION AS base-builder
 
-ENV PATH="/workspace/miniconda3/bin:${PATH}"
+ENV PATH="/root/miniconda3/bin:${PATH}"
 
 ARG PYTHON_VERSION="3.11"
 ARG PYTORCH_VERSION="nightly"
@@ -19,14 +19,14 @@ RUN apt-get update \
     && apt-get install -y wget git build-essential ninja-build git-lfs libaio-dev pkg-config && rm -rf /var/lib/apt/lists/* \
     && wget \
     https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh \
-    && mkdir -p /workspace/.conda \
-    && bash Miniconda3-latest-Linux-x86_64.sh -b -p /workspace/miniconda3 \
+    && mkdir /root/.conda \
+    && bash Miniconda3-latest-Linux-x86_64.sh -b \
     && rm -f Miniconda3-latest-Linux-x86_64.sh \
     && conda tos accept --override-channels --channel https://repo.anaconda.com/pkgs/main \
     && conda tos accept --override-channels --channel https://repo.anaconda.com/pkgs/r \
     && conda create -n "py${PYTHON_VERSION}" python="${PYTHON_VERSION}"
 
-ENV PATH="/workspace/miniconda3/envs/py${PYTHON_VERSION}/bin:${PATH}"
+ENV PATH="/root/miniconda3/envs/py${PYTHON_VERSION}/bin:${PATH}"
 
 WORKDIR /workspace
 

examples/gemma3/gemma-3-1b-qlora.yml

@@ -1,7 +1,7 @@
 base_model: google/gemma-3-1b-it
-
-model_type: Gemma3ForCausalLM
-
+# optionally might have model_type or tokenizer_type
+model_type: AutoModelForCausalLM
+tokenizer_type: AutoTokenizer
 # Automatically upload checkpoint and final model to HF
 # hub_model_id: username/custom_model_name
 

examples/gemma3/gemma-3-270m-qlora.yml

@@ -1,7 +1,7 @@
 base_model: google/gemma-3-270m-it
-
-model_type: Gemma3ForCausalLM
-
+# optionally might have model_type or tokenizer_type
+model_type: AutoModelForCausalLM
+tokenizer_type: AutoTokenizer
 # Automatically upload checkpoint and final model to HF
 # hub_model_id: username/custom_model_name
 

examples/gemma3/gemma-3-4b-qlora.yml

@@ -1,8 +1,5 @@
 base_model: google/gemma-3-4b-it
 
-# Need to set else transformers tries to load vision too
-model_type: Gemma3ForCausalLM
-
 load_in_4bit: true
 
 # gemma3 doesn't seem to play nice with ddp

examples/gpt-oss/README.md

@@ -2,8 +2,6 @@
 
 [GPT-OSS](https://huggingface.co/collections/openai/gpt-oss-68911959590a1634ba11c7a4) are a family of open-weight MoE models trained by OpenAI, released in August 2025. There are two variants: 20B and 120B.
 
-In October 2025, OpenAI released safeguard models built upon GPT-OSS called [GPT-OSS-Safeguard](https://huggingface.co/collections/openai/gpt-oss-safeguard). They use the same architecture, so the same examples below can be re-used.
-
 This guide shows how to fine-tune it with Axolotl with multi-turn conversations and proper masking.
 
 ## Getting started
@@ -66,16 +64,6 @@ axolotl merge-sharded-fsdp-weights examples/gpt-oss/gpt-oss-120b-fft-fsdp2-offlo
 mv ./outputs/gpt-oss-out/merged/* ./outputs/gpt-oss-out/
 ```
 
-### How to set reasoning_effort in template?
-
-The harmony template has a feature to set the `reasoning_effort` during prompt building. The default is `medium`. If you would like to adjust this, you can add the following to your config:
-
-```yaml
-chat_template_kwargs:
-  reasoning_effort: "high"  # low | medium | high
-```
-
-Currently, this applies globally. There is no method to apply per sample yet. If you are interested in adding this, please feel free to create an Issue to discuss.
 
 ### Inferencing your fine-tuned model
 

examples/gpt-oss/gpt-oss-safeguard-20b-sft-lora-singlegpu.yaml

@@ -1,67 +0,0 @@
-base_model: openai/gpt-oss-safeguard-20b
-use_kernels: true
-model_quantization_config: Mxfp4Config
-model_quantization_config_kwargs:
-  dequantize: true
-
-plugins:
-  - axolotl.integrations.cut_cross_entropy.CutCrossEntropyPlugin
-
-experimental_skip_move_to_device: true  # prevent OOM by not putting model to GPU before sharding
-
-datasets:
-  - path: HuggingFaceH4/Multilingual-Thinking
-    type: chat_template
-    field_thinking: thinking
-    template_thinking_key: thinking
-
-dataset_prepared_path: last_run_prepared
-val_set_size: 0
-output_dir: ./outputs/gpt-oss-safeguard-out/
-
-sequence_len: 4096
-sample_packing: true
-
-adapter: lora
-lora_r: 8
-lora_alpha: 16
-lora_dropout: 0.0  # dropout not supported when using LoRA over expert parameters
-lora_target_linear: true
-
-# TODO: not supported for now, see peft#2710
-#lora_target_parameters:  # target the experts in the last two layers
-#  - "22._checkpoint_wrapped_module.mlp.experts.gate_up_proj"
-#  - "22._checkpoint_wrapped_module.mlp.experts.down_proj"
-#  - "23._checkpoint_wrapped_module.mlp.experts.gate_up_proj"
-#  - "23._checkpoint_wrapped_module.mlp.experts.down_proj"
-
-wandb_project:
-wandb_entity:
-wandb_watch:
-wandb_name:
-wandb_log_model:
-
-gradient_accumulation_steps: 8
-micro_batch_size: 1
-num_epochs: 1
-
-optimizer: adamw_torch_8bit
-lr_scheduler: constant_with_warmup
-learning_rate: 2e-4
-
-bf16: true
-tf32: true
-
-flash_attention: true
-attn_implementation: kernels-community/vllm-flash-attn3  # this is not needed if using flash_attn >= 2.8.3
-
-gradient_checkpointing: true
-activation_offloading: true
-
-logging_steps: 1
-saves_per_epoch: 1
-warmup_ratio: 0.1
-
-special_tokens:
-eot_tokens:
-  - "<|end|>"

requirements.txt

@@ -10,25 +10,25 @@ liger-kernel==0.6.3
 
 packaging==23.2
 
-huggingface_hub>=0.36.0
+huggingface_hub>=0.33.0
 peft>=0.17.1
 tokenizers>=0.21.1
 transformers==4.57.1
 accelerate==1.10.1
-datasets==4.3.0
+datasets==4.0.0
 deepspeed>=0.17.0
-trl==0.24.0
-hf_xet==1.2.0
-kernels>=0.9.0
+trl==0.23.1
+hf_xet==1.1.5
+kernels==0.9.0
 trackio
 
 optimum==1.16.2
 hf_transfer
 sentencepiece
-gradio==5.49.1
+gradio==5.41.1
 
 modal==1.0.2
-pydantic>=2.10.6
+pydantic==2.10.6
 addict
 fire
 PyYAML>=6.0
@@ -36,8 +36,8 @@ requests
 wandb
 einops
 colorama
-numba>=0.61.2
-numpy>=2.2.6
+numba
+numpy>=1.24.4,<=2.0.1
 
 # qlora things
 evaluate==0.4.1
@@ -50,7 +50,7 @@ python-dotenv==1.0.1
 
 # remote filesystems
 s3fs>=2024.5.0
-gcsfs>=2025.3.0
+gcsfs>=2024.5.0
 adlfs>=2024.5.0
 ocifs==1.3.2
 
@@ -66,7 +66,7 @@ antlr4-python3-runtime==4.13.2
 torchao==0.13.0
 schedulefree==1.4.1
 
-axolotl-contribs-lgpl==0.0.7
+axolotl-contribs-lgpl==0.0.6
 axolotl-contribs-mit==0.0.5
 
 mistral-common==1.8.5

setup.py

@@ -62,15 +62,8 @@ def parse_requirements(extras_require_map):
             else:
                 raise ValueError("Invalid version format")
 
-            if (major, minor) >= (2, 9):
-                extras_require_map.pop("fbgemm-gpu")
-                extras_require_map["fbgemm-gpu"] = ["fbgemm-gpu-genai==1.4.1"]
-                extras_require_map["vllm"] = ["vllm==0.11.1"]
-                _install_requires.pop(_install_requires.index(xformers_version))
-            elif (major, minor) >= (2, 8):
-                extras_require_map.pop("fbgemm-gpu")
-                extras_require_map["fbgemm-gpu"] = ["fbgemm-gpu-genai==1.3.0"]
-                extras_require_map["vllm"] = ["vllm==0.11.0"]
+            if (major, minor) >= (2, 8):
+                pass
             elif (major, minor) >= (2, 7):
                 _install_requires.pop(_install_requires.index(xformers_version))
                 if patch == 0:
@@ -79,7 +72,7 @@ def parse_requirements(extras_require_map):
                     extras_require_map.pop("vllm")
                 else:
                     _install_requires.append("xformers==0.0.31")
-                    extras_require_map["vllm"] = ["vllm==0.10.1"]
+                    extras_require_map["vllm"] = ["vllm>=0.10.0"]
             elif (major, minor) >= (2, 6):
                 _install_requires.pop(_install_requires.index(xformers_version))
                 _install_requires.append("xformers==0.0.29.post3")
@@ -165,7 +158,7 @@ extras_require = {
     "llmcompressor": [
         "llmcompressor==0.5.1",
     ],
-    "fbgemm-gpu": ["fbgemm-gpu-genai==1.3.0"],
+    "fbgemm-gpu": ["fbgemm-gpu-genai>=1.2.0"],
     "opentelemetry": [
         "opentelemetry-api",
         "opentelemetry-sdk",

src/axolotl/core/builders/causal.py

@@ -12,7 +12,7 @@ from transformers import (
     EarlyStoppingCallback,
     Trainer,
 )
-from trl.trainer.reward_trainer import DataCollatorForPreference
+from trl.trainer.utils import RewardDataCollatorWithPadding
 
 from axolotl.core.builders.base import TrainerBuilderBase
 from axolotl.core.trainers import (
@@ -453,7 +453,7 @@ class HFCausalTrainerBuilder(TrainerBuilderBase):
                 BatchSamplerDataCollatorForSeq2Seq,
                 DataCollatorForSeq2Seq,
                 DataCollatorWithFlattening,
-                DataCollatorForPreference,
+                RewardDataCollatorWithPadding,
             ]
         ]
         collator_args = [self.tokenizer]
@@ -470,10 +470,7 @@ class HFCausalTrainerBuilder(TrainerBuilderBase):
             if kwargs and isinstance(kwargs, dict):
                 kwargs.update(collator_cls_and_kwargs[1])
         elif self.cfg.reward_model:
-            collator = DataCollatorForPreference
-            tokenizer = collator_args.pop(0)
-            kwargs["pad_token_id"] = tokenizer.pad_token_id
-            kwargs.pop("padding")
+            collator = RewardDataCollatorWithPadding
         elif use_batch_sampler_collator:
             # Use V2BatchSamplerDataCollatorForSeq2Seq for flex attention,
             # supported multipack models, or non-flash-attention llama

src/axolotl/prompt_strategies/bradley_terry/chat_template.py

@@ -71,10 +71,10 @@ class BTChatTemplateStrategy(ChatTemplateStrategy):
             ]
 
         return {
-            "chosen_input_ids": chosen_tokenized["input_ids"],
+            "input_ids_chosen": chosen_tokenized["input_ids"],
             "attention_mask_chosen": chosen_tokenized["attention_mask"],
             "labels_chosen": 1.0,
-            "rejected_input_ids": rejected_tokenized["input_ids"],
+            "input_ids_rejected": rejected_tokenized["input_ids"],
             "attention_mask_rejected": rejected_tokenized["attention_mask"],
             "labels_rejected": 0.0,
         }