additional validation for fsdp2, bump dep versions

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

@@ -24,13 +24,6 @@ jobs:
       fail-fast: false
       matrix:
         include:
-          - cuda: 124
-            cuda_version: 12.4.1
-            python_version: "3.11"
-            pytorch: 2.6.0
-            axolotl_extras: vllm
-            num_gpus: 2
-            nightly_build: "true"
           - cuda: 124
             cuda_version: 12.4.1
             python_version: "3.11"
@@ -45,6 +38,13 @@ jobs:
             axolotl_extras: vllm
             num_gpus: 2
             nightly_build: "true"
+          - cuda: 124
+            cuda_version: 12.4.1
+            python_version: "3.11"
+            pytorch: 2.6.0
+            axolotl_extras: vllm
+            num_gpus: 2
+            nightly_build: "true"
     runs-on: [self-hosted, modal]
     timeout-minutes: 120
     steps:

.github/workflows/tests.yml

@@ -211,7 +211,7 @@ jobs:
           - cuda: 124
             cuda_version: 12.4.1
             python_version: "3.11"
-            pytorch: 2.6.0
+            pytorch: 2.5.1
             num_gpus: 1
             axolotl_extras: vllm
     steps:
@@ -258,7 +258,7 @@ jobs:
           - cuda: 124
             cuda_version: 12.4.1
             python_version: "3.11"
-            pytorch: 2.5.1
+            pytorch: 2.6.0
             num_gpus: 1
             axolotl_extras: vllm
     steps:

CNAME

@@ -1 +0,0 @@
-docs.axolotl.ai

README.md

@@ -63,7 +63,7 @@ axolotl fetch examples
 axolotl fetch deepspeed_configs  # OPTIONAL
 ```
 
-Other installation approaches are described [here](https://docs.axolotl.ai/docs/installation.html).
+Other installation approaches are described [here](https://axolotl-ai-cloud.github.io/axolotl/docs/installation.html).
 
 ### Your First Fine-tune
 
@@ -78,7 +78,7 @@ axolotl fetch examples --dest path/to/folder
 axolotl train examples/llama-3/lora-1b.yml
 ```
 
-That's it! Check out our [Getting Started Guide](https://docs.axolotl.ai/docs/getting-started.html) for a more detailed walkthrough.
+That's it! Check out our [Getting Started Guide](https://axolotl-ai-cloud.github.io/axolotl/docs/getting-started.html) for a more detailed walkthrough.
 
 ## ✨ Key Features
 
@@ -91,20 +91,20 @@ That's it! Check out our [Getting Started Guide](https://docs.axolotl.ai/docs/ge
 
 ## 📚 Documentation
 
-- [Installation Options](https://docs.axolotl.ai/docs/installation.html) - Detailed setup instructions for different environments
+- [Installation Options](https://axolotl-ai-cloud.github.io/axolotl/docs/installation.html) - Detailed setup instructions for different environments
-- [Configuration Guide](https://docs.axolotl.ai/docs/config.html) - Full configuration options and examples
+- [Configuration Guide](https://axolotl-ai-cloud.github.io/axolotl/docs/config.html) - Full configuration options and examples
-- [Dataset Guide](https://docs.axolotl.ai/docs/dataset-formats/) - Supported formats and how to use them
+- [Dataset Guide](https://axolotl-ai-cloud.github.io/axolotl/docs/dataset-formats/) - Supported formats and how to use them
-- [Multi-GPU Training](https://docs.axolotl.ai/docs/multi-gpu.html)
+- [Multi-GPU Training](https://axolotl-ai-cloud.github.io/axolotl/docs/multi-gpu.html)
-- [Multi-Node Training](https://docs.axolotl.ai/docs/multi-node.html)
+- [Multi-Node Training](https://axolotl-ai-cloud.github.io/axolotl/docs/multi-node.html)
-- [Multipacking](https://docs.axolotl.ai/docs/multipack.html)
+- [Multipacking](https://axolotl-ai-cloud.github.io/axolotl/docs/multipack.html)
-- [API Reference](https://docs.axolotl.ai/docs/api/) - Auto-generated code documentation
+- [API Reference](https://axolotl-ai-cloud.github.io/axolotl/docs/api/) - Auto-generated code documentation
-- [FAQ](https://docs.axolotl.ai/docs/faq.html) - Frequently asked questions
+- [FAQ](https://axolotl-ai-cloud.github.io/axolotl/docs/faq.html) - Frequently asked questions
 
 ## 🤝 Getting Help
 
 - Join our [Discord community](https://discord.gg/HhrNrHJPRb) for support
 - Check out our [Examples](https://github.com/axolotl-ai-cloud/axolotl/tree/main/examples/) directory
-- Read our [Debugging Guide](https://docs.axolotl.ai/docs/debugging.html)
+- Read our [Debugging Guide](https://axolotl-ai-cloud.github.io/axolotl/docs/debugging.html)
 - Need dedicated support? Please contact [✉️wing@axolotl.ai](mailto:wing@axolotl.ai) for options
 
 ## 🌟 Contributing

_quarto.yml

@@ -231,7 +231,6 @@ website:
             - docs/reward_modelling.qmd
             - docs/lr_groups.qmd
             - docs/lora_optims.qmd
-            - docs/dataset_loading.qmd
 
         - section: "Core Concepts"
           contents:

cicd/multigpu.py

@@ -68,7 +68,7 @@ def run_cmd(cmd: str, run_folder: str):
 @app.function(
     image=cicd_image,
     gpu=GPU_CONFIG,
-    timeout=90 * 60,
+    timeout=60 * 60,
     cpu=8.0,
     memory=131072 * N_GPUS,
     volumes=VOLUME_CONFIG,

docker/Dockerfile-base

@@ -29,7 +29,7 @@ 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 && \
-    python3 -m pip install --no-cache-dir -U torch==${PYTORCH_VERSION}+cu${CUDA} torchvision --extra-index-url https://download.pytorch.org/whl/cu$CUDA && \
+    python3 -m pip install --no-cache-dir -U torch==${PYTORCH_VERSION}+cu${CUDA} --extra-index-url https://download.pytorch.org/whl/cu$CUDA && \
     python3 -m pip install --no-cache-dir "causal_conv1d @ git+https://github.com/Dao-AILab/causal-conv1d.git@main" && \
     python3 -m pip install --no-cache-dir "mamba_ssm @ git+https://github.com/state-spaces/mamba.git@main"
 

docs/config.qmd

@@ -90,7 +90,7 @@ lora_on_cpu: true
 
 # List[str]. Add plugins to extend the pipeline.
 # See `src/axolotl/integrations` for the available plugins or doc below for more details.
-# https://docs.axolotl.ai/docs/custom_integrations.html
+# https://axolotl-ai-cloud.github.io/axolotl/docs/custom_integrations.html
 plugins:
   # - axolotl.integrations.cut_cross_entropy.CutCrossEntropyPlugin
 
@@ -109,7 +109,7 @@ datasets:
     preprocess_shards: # Optional[int] process dataset in N sequential chunks for memory efficiency (exclusive with `shards`)
 
     name: # Optional[str] name of dataset configuration to load
-    split: train # Optional[str] name of dataset split to load from
+    train_on_split: train # Optional[str] name of dataset split to load from
     revision: # Optional[str] The specific revision of the dataset to use when loading from the Hugging Face Hub. This can be a commit hash, tag, or branch name. If not specified, the latest version will be used. This parameter is ignored for local datasets.
     trust_remote_code: # Optional[bool] Trust remote code for untrusted source
 
@@ -165,9 +165,7 @@ datasets:
       content: value
       # ...
 
-    # Optional[Dict[str, List]]. Roles mapping in the messages.
+    # Optional[Dict[str, List]]. Roles mapping in the messages. The default is:
-    # The format is {target_role: [source_roles]}. All source roles will be mapped to the target role.
-    # The default is:
     roles:
       user: ["human", "user"]
       assistant: ["gpt", "assistant"]
@@ -394,7 +392,7 @@ lora_fan_in_fan_out: false
 
 # Apply custom LoRA autograd functions and activation function Triton kernels for
 # speed and memory savings
-# See: https://docs.axolotl.ai/docs/lora_optims.html
+# See: https://axolotl-ai-cloud.github.io/axolotl/docs/lora_optims.html
 lora_mlp_kernel: true
 lora_qkv_kernel: true
 lora_o_kernel: true
@@ -688,7 +686,7 @@ ddp_broadcast_buffers:
 # Use in long context training to prevent OOM when sequences cannot fit into a single GPU's VRAM.
 # E.g., if 4 GPUs are available, set this value to 2 to split each sequence into two equal-sized
 # subsequences, or set to 4 to split into four equal-sized subsequences.
-# See https://docs.axolotl.ai/docs/sequence_parallelism.html for more details.
+# See https://axolotl-ai-cloud.github.io/axolotl/docs/sequence_parallelism.html for more details.
 sequence_parallel_degree:
 # Optional; strides across the key dimension. Larger values use more memory but should make training faster.
 # Must evenly divide the number of KV heads in your model.

docs/dataset-formats/index.qmd

@@ -13,13 +13,6 @@ As there are a lot of available options in Axolotl, this guide aims to provide a
 
 Axolotl supports 3 kinds of training methods: pre-training, supervised fine-tuning, and preference-based post-training (e.g. DPO, ORPO, PRMs). Each method has their own dataset format which are described below.
 
-::: {.callout-tip}
-
-This guide will mainly use JSONL as an introduction. Please refer to the [dataset loading docs](../dataset_loading.qmd) to understand how to load datasets from other sources.
-
-For `pretraining_dataset:` specifically, please refer to the [Pre-training section](#pre-training).
-:::
-
 ## Pre-training
 
 When aiming to train on large corpora of text datasets, pre-training is your go-to choice. Due to the size of these datasets, downloading the entire-datasets before beginning training would be prohibitively time-consuming. Axolotl supports [streaming](https://huggingface.co/docs/datasets/en/stream) to only load batches into memory at a time.
@@ -457,7 +450,10 @@ datasets:
     type: alpaca
 ```
 
-Axolotl supports many kinds of instruction dataset. All of them can be found in the [Instruction Dataset Documentation](inst_tune.qmd) with their respective type and sample row format.
+Axolotl supports many kinds of instruction dataset. All of them can be found here (https://axolotl-ai-cloud.github.io/axolotl/docs/dataset-formats/inst_tune.html) with their respective type and sample row format.
+
+
+Reference: [Instruction Dataset Documentation](inst_tune.qmd).
 
 #### Custom Instruct Prompt Format
 

docs/dataset_loading.qmd

@@ -1,276 +0,0 @@
----
-title: Dataset Loading
-description: Understanding how to load datasets from different sources
-back-to-top-navigation: true
-toc: true
-toc-depth: 5
----
-
-## Overview
-
-Datasets can be loaded in a number of different ways depending on the how it is saved (the extension of the file) and where it is stored.
-
-## Loading Datasets
-
-We use the `datasets` library to load datasets and a mix of `load_dataset` and `load_from_disk` to load them.
-
-You may recognize the similar named configs between `load_dataset` and the `datasets` section of the config file.
-
-```yaml
-datasets:
-  - path:
-    name:
-    data_files:
-    split:
-    revision:
-    trust_remote_code:
-```
-
-::: {.callout-tip}
-
-Do not feel overwhelmed by the number of options here. A lot of them are optional. In fact, the most common config to use would be `path` and sometimes `data_files`.
-
-:::
-
-This matches the API of [`datasets.load_dataset`](https://github.com/huggingface/datasets/blob/0b5998ac62f08e358f8dcc17ec6e2f2a5e9450b6/src/datasets/load.py#L1838-L1858), so if you're familiar with that, you will feel right at home.
-
-For HuggingFace's guide to load different dataset types, see [here](https://huggingface.co/docs/datasets/loading).
-
-For full details on the config, see [config.qmd](config.qmd).
-
-::: {.callout-note}
-
-You can set multiple datasets in the config file by more than one entry under `datasets`.
-
-```yaml
-datasets:
-  - path: /path/to/your/dataset
-  - path: /path/to/your/other/dataset
-```
-
-:::
-
-### Local dataset
-
-#### Files
-
-Usually, to load a JSON file, you would do something like this:
-
-```python
-from datasets import load_dataset
-
-dataset = load_dataset("json", data_files="data.json")
-```
-
-Which translates to the following config:
-
-```yaml
-datasets:
-  - path: json
-    data_files: /path/to/your/file.jsonl
-```
-
-However, to make things easier, we have added a few shortcuts for loading local dataset files.
-
-You can just point the `path` to the file or directory along with the `ds_type` to load the dataset. The below example shows for a JSON file:
-
-```yaml
-datasets:
-  - path: /path/to/your/file.jsonl
-    ds_type: json
-```
-
-This works for CSV, JSON, Parquet, and Arrow files.
-
-::: {.callout-tip}
-
-If `path` points to a file and `ds_type` is not specified, we will automatically infer the dataset type from the file extension, so you could omit `ds_type` if you'd like.
-
-:::
-
-#### Directory
-
-If you're loading a directory, you can point the `path` to the directory.
-
-Then, you have two options:
-
-##### Loading entire directory
-
-You do not need any additional configs.
-
-We will attempt to load in the following order:
-- datasets saved with `datasets.save_to_disk`
-- loading entire directory of files (such as with parquet/arrow files)
-
-```yaml
-datasets:
-  - path: /path/to/your/directory
-```
-
-##### Loading specific files in directory
-
-Provide `data_files` with a list of files to load.
-
-```yaml
-datasets:
-    # single file
-  - path: /path/to/your/directory
-    ds_type: csv
-    data_files: file1.csv
-
-    # multiple files
-  - path: /path/to/your/directory
-    ds_type: json
-    data_files:
-      - file1.jsonl
-      - file2.jsonl
-
-    # multiple files for parquet
-  - path: /path/to/your/directory
-    ds_type: parquet
-    data_files:
-      - file1.parquet
-      - file2.parquet
-
-```
-
-### HuggingFace Hub
-
-The method you use to load the dataset depends on how the dataset was created, whether a folder was uploaded directly or a HuggingFace Dataset was pushed.
-
-::: {.callout-note}
-
-If you're using a private dataset, you will need to enable the `hf_use_auth_token` flag in the root-level of the config file.
-
-:::
-
-#### Folder uploaded
-
-This would mean that the dataset is a single file or file(s) uploaded to the Hub.
-
-```yaml
-datasets:
-  - path: org/dataset-name
-    data_files:
-      - file1.jsonl
-      - file2.jsonl
-```
-
-#### HuggingFace Dataset
-
-This means that the dataset is created as a HuggingFace Dataset and pushed to the Hub via `datasets.push_to_hub`.
-
-```yaml
-datasets:
-  - path: org/dataset-name
-```
-
-::: {.callout-note}
-
-There are some other configs which may be required like `name`, `split`, `revision`, `trust_remote_code`, etc depending on the dataset.
-
-:::
-
-### Remote Filesystems
-
-Via the `storage_options` config under `load_dataset`, you can load datasets from remote filesystems like S3, GCS, Azure, and OCI.
-
-::: {.callout-warning}
-
-This is currently experimental. Please let us know if you run into any issues!
-
-:::
-
-The only difference between the providers is that you need to prepend the path with the respective protocols.
-
-```yaml
-datasets:
-    # Single file
-  - path: s3://bucket-name/path/to/your/file.jsonl
-
-    # Directory
-  - path: s3://bucket-name/path/to/your/directory
-```
-
-For directory, we load via `load_from_disk`.
-
-#### S3
-
-Prepend the path with `s3://`.
-
-The credentials are pulled in the following order:
-
-- `AWS_ACCESS_KEY_ID`, `AWS_SECRET_ACCESS_KEY`, and `AWS_SESSION_TOKEN` environment variables
-- from the `~/.aws/credentials` file
-- for nodes on EC2, the IAM metadata provider
-
-::: {.callout-note}
-
-We assume you have credentials setup and not using anonymous access. If you want to use anonymous access, let us know! We may have to open a config option for this.
-
-:::
-
-Other environment variables that can be set can be found in [boto3 docs](https://boto3.amazonaws.com/v1/documentation/api/latest/guide/configuration.html#using-environment-variables)
-
-#### GCS
-
-Prepend the path with `gs://` or `gcs://`.
-
-The credentials are loaded in the following order:
-
-- gcloud credentials
-- for nodes on GCP, the google metadata service
-- anonymous access
-
-#### Azure
-
-##### Gen 1
-
-Prepend the path with `adl://`.
-
-Ensure you have the following environment variables set:
-
-- `AZURE_STORAGE_TENANT_ID`
-- `AZURE_STORAGE_CLIENT_ID`
-- `AZURE_STORAGE_CLIENT_SECRET`
-
-##### Gen 2
-
-Prepend the path with `abfs://` or `az://`.
-
-Ensure you have the following environment variables set:
-
-- `AZURE_STORAGE_ACCOUNT_NAME`
-- `AZURE_STORAGE_ACCOUNT_KEY`
-
-Other environment variables that can be set can be found in [adlfs docs](https://github.com/fsspec/adlfs?tab=readme-ov-file#setting-credentials)
-
-#### OCI
-
-Prepend the path with `oci://`.
-
-It would attempt to read in the following order:
-
-- `OCIFS_IAM_TYPE`, `OCIFS_CONFIG_LOCATION`, and `OCIFS_CONFIG_PROFILE` environment variables
-- when on OCI resource, resource principal
-
-Other environment variables:
-
-- `OCI_REGION_METADATA`
-
-Please see the [ocifs docs](https://ocifs.readthedocs.io/en/latest/getting-connected.html#Using-Environment-Variables).
-
-### HTTPS
-
-The path should start with `https://`.
-
-```yaml
-datasets:
-  - path: https://path/to/your/dataset/file.jsonl
-```
-
-This must be publically accessible.
-
-## Next steps
-
-Now that you know how to load datasets, you can learn more on how to load your specific dataset format into your target output format [dataset formats docs](dataset-formats).

docs/multi-gpu.qmd

@@ -36,9 +36,6 @@ deepspeed: deepspeed_configs/zero1.json
 ### Usage {#sec-deepspeed-usage}
 
 ```{.bash}
-# Fetch deepspeed configs (if not already present)
-axolotl fetch deepspeed_configs
-
 # Passing arg via config
 axolotl train config.yml
 
@@ -51,20 +48,10 @@ axolotl train config.yml --deepspeed deepspeed_configs/zero1.json
 We provide default configurations for:
 
 - ZeRO Stage 1 (`zero1.json`)
-- ZeRO Stage 1 with torch compile (`zero1_torch_compile.json`)
 - ZeRO Stage 2 (`zero2.json`)
 - ZeRO Stage 3 (`zero3.json`)
-- ZeRO Stage 3 with bf16 (`zero3_bf16.json`)
-- ZeRO Stage 3 with bf16 and CPU offload params(`zero3_bf16_cpuoffload_params.json`)
-- ZeRO Stage 3 with bf16 and CPU offload params and optimizer (`zero3_bf16_cpuoffload_all.json`)
 
-::: {.callout-tip}
+Choose based on your memory requirements and performance needs.
-
-Choose the configuration that offloads the least amount to memory while still being able to fit on VRAM for best performance.
-
-Start from Stage 1 -> Stage 2 -> Stage 3.
-
-:::
 
 ## FSDP {#sec-fsdp}
 

docs/multimodal.qmd

@@ -9,7 +9,6 @@ format:
 ## Supported Models
 
 - [Mllama](#sec-mllama)
-- [Llama4](#sec-llama4)
 - [Pixtral](#sec-pixtral)
 - [Llava-1.5](#sec-llava-15)
 - [Mistral-Small-3.1](#sec-mistral-small-31)
@@ -64,14 +63,6 @@ base_model: meta-llama/Llama-3.2-11B-Vision-Instruct
 chat_template: llama3_2_vision
 ```
 
-### Llama4 {#sec-llama4}
-
-```yaml
-base_model: meta-llama/Llama-4-Scout-17B-16E-Instruct
-
-chat_template: llama4
-```
-
 ### Pixtral {#sec-pixtral}
 
 ```yaml

docs/rlhf.qmd

@@ -530,7 +530,7 @@ trl:
 ```
 
 ```bash
-CUDA_VISIBLE_DEVICES=2,3 axolotl vllm-serve grpo.yaml
+CUDA_VISIBLE_DEVICES=2,3 axolotl vllm_serve grpo.yaml
 ```
 
 Your `vLLM` instance will now attempt to spin up, and it's time to kick off training utilizing our remaining two GPUs. In another terminal, execute:

examples/cerebras/btlm-ft.yml

@@ -8,6 +8,7 @@ tokenizer_type: GPT2Tokenizer
 trust_remote_code: true
 tokenizer_use_fast: true
 tokenizer_legacy: true
+strict: false
 push_dataset_to_hub:
 hf_use_auth_token: true
 datasets:

examples/cerebras/qlora.yml

@@ -4,6 +4,7 @@ base_model: cerebras/Cerebras-GPT-1.3B
 
 load_in_8bit: false
 load_in_4bit: true
+strict: false
 push_dataset_to_hub:
 datasets:
   - path: teknium/GPT4-LLM-Cleaned

examples/code-llama/13b/lora.yml

@@ -7,6 +7,7 @@ tokenizer_type: CodeLlamaTokenizer
 
 load_in_8bit: true
 load_in_4bit: false
+strict: false
 
 datasets:
   - path: mhenrichsen/alpaca_2k_test

examples/code-llama/13b/qlora.yml

@@ -7,6 +7,7 @@ tokenizer_type: CodeLlamaTokenizer
 
 load_in_8bit: false
 load_in_4bit: true
+strict: false
 
 datasets:
   - path: mhenrichsen/alpaca_2k_test

examples/code-llama/34b/lora.yml

@@ -7,6 +7,7 @@ tokenizer_type: CodeLlamaTokenizer
 
 load_in_8bit: true
 load_in_4bit: false
+strict: false
 
 datasets:
   - path: mhenrichsen/alpaca_2k_test

examples/code-llama/34b/qlora.yml

@@ -7,6 +7,7 @@ tokenizer_type: CodeLlamaTokenizer
 
 load_in_8bit: false
 load_in_4bit: true
+strict: false
 
 datasets:
   - path: mhenrichsen/alpaca_2k_test

examples/code-llama/7b/lora.yml

@@ -7,6 +7,7 @@ tokenizer_type: CodeLlamaTokenizer
 
 load_in_8bit: true
 load_in_4bit: false
+strict: false
 
 datasets:
   - path: mhenrichsen/alpaca_2k_test

examples/code-llama/7b/qlora.yml

@@ -7,6 +7,7 @@ tokenizer_type: CodeLlamaTokenizer
 
 load_in_8bit: false
 load_in_4bit: true
+strict: false
 
 datasets:
   - path: mhenrichsen/alpaca_2k_test

examples/cohere/command-r-7b-qlora.yml

@@ -4,6 +4,7 @@ tokenizer_type: AutoTokenizer
 
 load_in_8bit: false
 load_in_4bit: true
+strict: false
 
 # huggingface repo
 chat_template: cohere

examples/dbrx/16bit-lora.yaml

@@ -3,6 +3,7 @@ base_model: LnL-AI/dbrx-base-converted-v2
 # hub_model_id: username/custom_model_name
 
 trust_remote_code: true
+strict: false
 
 datasets:
   - path: tatsu-lab/alpaca

examples/dbrx/8bit-lora.yaml

@@ -6,6 +6,7 @@ trust_remote_code: true
 
 load_in_8bit: true
 load_in_4bit: false
+strict: false
 
 datasets:
   - path: tatsu-lab/alpaca

examples/dbrx/fft-ds-zero3.yaml

@@ -3,6 +3,7 @@ base_model: LnL-AI/dbrx-base-converted-v2
 # hub_model_id: username/custom_model_name
 
 trust_remote_code: true
+strict: false
 
 datasets:
   - path: tatsu-lab/alpaca

examples/deepcoder/deepcoder-14B-preview-lora.yml

@@ -1,58 +0,0 @@
-base_model: agentica-org/DeepCoder-14B-Preview
-# Automatically upload checkpoint and final model to HF
-# hub_model_id: username/custom_model_name
-
-load_in_8bit: true
-load_in_4bit: false
-strict: false
-
-datasets:
-  - path: fozziethebeat/alpaca_messages_2k_test
-    type: chat_template
-    field_messages: messages
-    message_property_mappings:
-      role: role
-      content: content
-
-dataset_prepared_path:
-val_set_size: 0.05
-output_dir: ./outputs/lora-out
-
-sequence_len: 4096
-sample_packing: true
-eval_sample_packing: false
-pad_to_sequence_len: true
-
-adapter: lora
-lora_model_dir:
-lora_r: 32
-lora_alpha: 16
-lora_dropout: 0.05
-lora_target_linear: true
-
-wandb_project:
-wandb_entity:
-wandb_watch:
-wandb_name:
-wandb_log_model:
-
-gradient_accumulation_steps: 2
-micro_batch_size: 2
-num_epochs: 4
-optimizer: adamw_bnb_8bit
-lr_scheduler: cosine
-learning_rate: 0.0002
-
-bf16: auto
-tf32: true
-
-gradient_checkpointing: true
-resume_from_checkpoint:
-logging_steps: 1
-flash_attention: true
-
-warmup_steps: 10
-evals_per_epoch: 1
-saves_per_epoch: 1
-weight_decay: 0.0
-special_tokens:

examples/deepcogito/cogito-v1-preview-llama-3B-lora.yml

@@ -1,58 +0,0 @@
-base_model: deepcogito/cogito-v1-preview-llama-3B
-# Automatically upload checkpoint and final model to HF
-# hub_model_id: username/custom_model_name
-
-load_in_8bit: true
-load_in_4bit: false
-strict: false
-
-datasets:
-  - path: fozziethebeat/alpaca_messages_2k_test
-    type: chat_template
-    field_messages: messages
-    message_property_mappings:
-      role: role
-      content: content
-
-dataset_prepared_path:
-val_set_size: 0.05
-output_dir: ./outputs/lora-out
-
-sequence_len: 4096
-sample_packing: true
-eval_sample_packing: false
-pad_to_sequence_len: true
-
-adapter: lora
-lora_model_dir:
-lora_r: 32
-lora_alpha: 16
-lora_dropout: 0.05
-lora_target_linear: true
-
-wandb_project:
-wandb_entity:
-wandb_watch:
-wandb_name:
-wandb_log_model:
-
-gradient_accumulation_steps: 2
-micro_batch_size: 2
-num_epochs: 1
-optimizer: adamw_bnb_8bit
-lr_scheduler: cosine
-learning_rate: 0.0002
-
-bf16: auto
-tf32: true
-
-gradient_checkpointing: true
-resume_from_checkpoint:
-logging_steps: 1
-flash_attention: true
-
-warmup_steps: 10
-evals_per_epoch: 1
-saves_per_epoch: 1
-weight_decay: 0.0
-special_tokens:

examples/deepcogito/cogito-v1-preview-qwen-14B-lora.yml

@@ -1,58 +0,0 @@
-base_model: deepcogito/cogito-v1-preview-qwen-14B
-# Automatically upload checkpoint and final model to HF
-# hub_model_id: username/custom_model_name
-
-load_in_8bit: true
-load_in_4bit: false
-strict: false
-
-datasets:
-  - path: fozziethebeat/alpaca_messages_2k_test
-    type: chat_template
-    field_messages: messages
-    message_property_mappings:
-      role: role
-      content: content
-
-dataset_prepared_path:
-val_set_size: 0.05
-output_dir: ./outputs/lora-out
-
-sequence_len: 4096
-sample_packing: true
-eval_sample_packing: false
-pad_to_sequence_len: true
-
-adapter: lora
-lora_model_dir:
-lora_r: 32
-lora_alpha: 16
-lora_dropout: 0.05
-lora_target_linear: true
-
-wandb_project:
-wandb_entity:
-wandb_watch:
-wandb_name:
-wandb_log_model:
-
-gradient_accumulation_steps: 2
-micro_batch_size: 2
-num_epochs: 1
-optimizer: adamw_bnb_8bit
-lr_scheduler: cosine
-learning_rate: 0.0002
-
-bf16: auto
-tf32: true
-
-gradient_checkpointing: true
-resume_from_checkpoint:
-logging_steps: 1
-flash_attention: true
-
-warmup_steps: 10
-evals_per_epoch: 1
-saves_per_epoch: 1
-weight_decay: 0.0
-special_tokens:

examples/deepseek-v2/fft-fsdp-16b.yaml

@@ -2,6 +2,7 @@ base_model: deepseek-ai/DeepSeek-V2-Lite
 # Automatically upload checkpoint and final model to HF
 # hub_model_id: username/custom_model_name
 trust_remote_code: true
+strict: false
 
 datasets:
   - path: tatsu-lab/alpaca

examples/deepseek-v2/qlora-fsdp-2_5.yaml

@@ -6,6 +6,7 @@ trust_remote_code: true
 
 load_in_8bit: false
 load_in_4bit: true
+strict: false
 
 
 plugins:

examples/falcon/config-7b-lora.yml

@@ -11,6 +11,7 @@ trust_remote_code: true
 load_in_8bit: true
 load_in_4bit: false
 gptq: false
+strict: false
 push_dataset_to_hub:
 datasets:
   - path: teknium/GPT4-LLM-Cleaned

examples/falcon/config-7b-qlora.yml

@@ -15,6 +15,7 @@ load_in_8bit: false
 # enable 4bit for QLoRA
 load_in_4bit: true
 gptq: false
+strict: false
 push_dataset_to_hub:
 datasets:
   - path: QingyiSi/Alpaca-CoT

examples/falcon/config-7b.yml

@@ -8,6 +8,7 @@ tokenizer_type: AutoTokenizer
 # required by falcon custom model code: https://huggingface.co/tiiuae/falcon-7b/tree/main
 trust_remote_code: true
 gptq: false
+strict: false
 push_dataset_to_hub:
 datasets:
   - path: teknium/GPT4-LLM-Cleaned

examples/gemma/qlora.yml

@@ -8,6 +8,7 @@ tokenizer_type: AutoTokenizer
 
 load_in_8bit: false
 load_in_4bit: true
+strict: false
 
 # huggingface repo
 datasets:

examples/gemma2/qlora.yml

@@ -7,6 +7,7 @@ tokenizer_type: AutoTokenizer
 
 load_in_8bit: false
 load_in_4bit: true
+strict: false
 
 # huggingface repo
 chat_template: gemma

examples/gemma2/reward-model.yaml

@@ -5,6 +5,7 @@ num_labels: 1
 tokenizer_type: AutoTokenizer
 # Automatically upload checkpoint and final model to HF
 # hub_model_id: username/custom_model_name
+strict: false
 
 reward_model: true
 chat_template: gemma

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

@@ -10,6 +10,7 @@ ddp_find_unused_parameters: true
 
 load_in_8bit: false
 load_in_4bit: true
+strict: false
 
 # huggingface repo
 chat_template: gemma3

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

@@ -1,4 +1,5 @@
 base_model: google/gemma-3-4b-it
+strict: false
 
 load_in_4bit: true
 

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

@@ -1,5 +1,6 @@
 base_model: google/gemma-3-4b-it
 processor_type: AutoProcessor
+strict: false
 
 load_in_4bit: true
 

examples/gptj/qlora.yml

@@ -4,6 +4,7 @@ base_model: EleutherAI/gpt-j-6b
 
 load_in_8bit: false
 load_in_4bit: true
+strict: false
 push_dataset_to_hub:
 datasets:
   - path: teknium/GPT4-LLM-Cleaned

examples/jamba/qlora.yaml

@@ -6,6 +6,7 @@ trust_remote_code: true
 
 load_in_8bit: false
 load_in_4bit: true
+strict: false
 
 datasets:
   - path: mhenrichsen/alpaca_2k_test

examples/jamba/qlora_deepspeed.yaml

@@ -5,6 +5,7 @@ trust_remote_code: true
 
 load_in_8bit: false
 load_in_4bit: true
+strict: false
 
 datasets:
   - path: mhenrichsen/alpaca_2k_test

examples/jamba/qlora_fsdp_large.yaml

@@ -5,6 +5,7 @@ tokenizer_type: AutoTokenizer
 # hub_model_id: username/custom_model_name
 
 load_in_4bit: true
+strict: false
 use_tensorboard: true
 chat_template: jamba
 datasets:

examples/llama-2/fft_optimized.yml

@@ -4,6 +4,7 @@ model_type: LlamaForCausalLM
 tokenizer_type: LlamaTokenizer
 # Automatically upload checkpoint and final model to HF
 # hub_model_id: username/custom_model_name
+strict: false
 
 datasets:
   - path: mhenrichsen/alpaca_2k_test

examples/llama-2/gptq-lora.yml

@@ -10,6 +10,7 @@ gptq_disable_exllama: true
 
 tokenizer_use_fast: true
 tokenizer_legacy: true
+strict: false
 push_dataset_to_hub:
 hf_use_auth_token: true
 datasets:

examples/llama-2/lisa.yml

@@ -4,6 +4,7 @@ model_type: LlamaForCausalLM
 tokenizer_type: LlamaTokenizer
 # Automatically upload checkpoint and final model to HF
 # hub_model_id: username/custom_model_name
+strict: false
 
 datasets:
   - path: teknium/GPT4-LLM-Cleaned

examples/llama-2/loftq.yml

@@ -4,6 +4,7 @@ model_type: LlamaForCausalLM
 tokenizer_type: LlamaTokenizer
 # Automatically upload checkpoint and final model to HF
 # hub_model_id: username/custom_model_name
+strict: false
 
 datasets:
   - path: mhenrichsen/alpaca_2k_test

examples/llama-2/lora.yml

@@ -7,6 +7,7 @@ tokenizer_type: LlamaTokenizer
 
 load_in_8bit: true
 load_in_4bit: false
+strict: false
 
 datasets:
   - path: mhenrichsen/alpaca_2k_test

examples/llama-2/qlora-fsdp.yml

@@ -7,6 +7,7 @@ tokenizer_type: LlamaTokenizer
 
 load_in_8bit: false
 load_in_4bit: true
+strict: false
 
 datasets:
   - path: yahma/alpaca-cleaned

examples/llama-2/qlora.yml

@@ -7,6 +7,7 @@ tokenizer_type: LlamaTokenizer
 
 load_in_8bit: false
 load_in_4bit: true
+strict: false
 
 datasets:
   - path: mhenrichsen/alpaca_2k_test

examples/llama-2/relora.yml

@@ -5,6 +5,7 @@ tokenizer_type: LlamaTokenizer
 
 load_in_8bit: false
 load_in_4bit: true
+strict: false
 
 datasets:
   - path: teknium/GPT4-LLM-Cleaned

examples/llama-3-vision/lora-11b.yaml

@@ -4,6 +4,7 @@ processor_type: AutoProcessor
 # Automatically upload checkpoint and final model to HF
 # hub_model_id: username/custom_model_name
 
+strict: false
 
 # these 3 lines are needed for now to handle vision chat templates w images
 skip_prepare_dataset: true

examples/llama-3/fft-8b-liger-fsdp.yaml

@@ -9,6 +9,7 @@ liger_rms_norm: true
 liger_glu_activation: true
 liger_fused_linear_cross_entropy: true
 
+strict: false
 
 chat_template: llama3
 datasets:

examples/llama-3/fft-8b.yaml

@@ -1,6 +1,7 @@
 base_model: NousResearch/Meta-Llama-3.1-8B
 # Automatically upload checkpoint and final model to HF
 # hub_model_id: username/custom_model_name
+strict: false
 
 datasets:
   - path: tatsu-lab/alpaca

examples/llama-3/instruct-dpo-lora-8b.yml

@@ -7,6 +7,7 @@ tokenizer_type: AutoTokenizer
 
 load_in_8bit: true
 load_in_4bit: false
+strict: false
 
 chat_template: llama3
 rl: dpo

examples/llama-3/instruct-lora-8b.yml

@@ -7,6 +7,7 @@ tokenizer_type: AutoTokenizer
 
 load_in_8bit: true
 load_in_4bit: false
+strict: false
 
 chat_template: llama3
 datasets:

examples/llama-3/lora-1b-deduplicate-dpo.yml

@@ -7,6 +7,7 @@ tokenizer_type: AutoTokenizer
 
 load_in_8bit: true
 load_in_4bit: false
+strict: false
 
 chat_template: llama3
 rl: dpo

examples/llama-3/lora-1b-deduplicate-sft.yml

@@ -7,6 +7,7 @@ tokenizer_type: AutoTokenizer
 
 load_in_8bit: true
 load_in_4bit: false
+strict: false
 
 datasets:
   - path: mhenrichsen/alpaca_2k_test

examples/llama-3/lora-1b-kernels.yml

@@ -1,6 +1,7 @@
 base_model: NousResearch/Llama-3.2-1B
 # Automatically upload checkpoint and final model to HF
 # hub_model_id: username/custom_model_name
+strict: false
 
 datasets:
   - path: teknium/GPT4-LLM-Cleaned

examples/llama-3/lora-1b-ray.yml

@@ -1,6 +1,7 @@
 base_model: NousResearch/Llama-3.2-1B
 # Automatically upload checkpoint and final model to HF
 # hub_model_id: username/custom_model_name
+strict: false
 
 datasets:
   - path: teknium/GPT4-LLM-Cleaned

examples/llama-3/lora-1b-sample-packing-sequentially.yml

@@ -7,6 +7,7 @@ tokenizer_type: AutoTokenizer
 
 load_in_8bit: true
 load_in_4bit: false
+strict: false
 
 datasets:
   - path: mhenrichsen/alpaca_2k_test

examples/llama-3/lora-1b.yml

@@ -1,6 +1,7 @@
 base_model: NousResearch/Llama-3.2-1B
 # Automatically upload checkpoint and final model to HF
 # hub_model_id: username/custom_model_name
+strict: false
 
 datasets:
   - path: teknium/GPT4-LLM-Cleaned

examples/llama-3/lora-8b.yml

@@ -7,6 +7,7 @@ tokenizer_type: AutoTokenizer
 
 load_in_8bit: true
 load_in_4bit: false
+strict: false
 
 datasets:
   - path: mhenrichsen/alpaca_2k_test

examples/llama-3/qlora-1b-kto.yaml

@@ -4,6 +4,7 @@ base_model: meta-llama/Llama-3.2-1B
 
 load_in_8bit: false
 load_in_4bit: true
+strict: false
 
 rl: kto
 rl_beta: 0.5

examples/llama-3/qlora-1b.yml

@@ -4,6 +4,7 @@ base_model: NousResearch/Llama-3.2-1B
 
 load_in_8bit: false
 load_in_4bit: true
+strict: false
 
 datasets:
   - path: teknium/GPT4-LLM-Cleaned

examples/llama-3/qlora-fsdp-405b.yaml

@@ -5,6 +5,7 @@ tokenizer_type: AutoTokenizer
 # hub_model_id: username/custom_model_name
 
 load_in_4bit: true
+strict: false
 
 datasets:
   - path: tatsu-lab/alpaca

examples/llama-3/qlora-fsdp-70b.yaml

@@ -7,6 +7,7 @@ tokenizer_type: AutoTokenizer  # PreTrainedTokenizerFast
 
 load_in_8bit: false
 load_in_4bit: true
+strict: false
 
 datasets:
   - path: tatsu-lab/alpaca

examples/llama-3/qlora.yml

@@ -7,6 +7,7 @@ tokenizer_type: AutoTokenizer
 
 load_in_8bit: false
 load_in_4bit: true
+strict: false
 
 datasets:
   - path: aaditya/alpaca_subset_1

examples/llama-4/README.md

@@ -1,16 +0,0 @@
-# Llama 4 by Meta AI
-
-## Available Examples
-
-### Llama 4 Scout 17Bx16Experts (109B)
-- [Multi-Modal/Vision QLoRA w/ FSDP1](./scout-vision-qlora-fsdp.yaml)
-- [Text Single GPU (H100) QLoRA](./scout-qlora-single-h100.yaml)
-- [Text Multi GPU QLoRA w/ FSDP1](./scout-qlora-fsdp1.yaml)
-
-Our Single H100 implementation for Llama 4 Scout uses only 68.5GB VRAM for post-training with 4k context length @ 546 tokens/second. [WandB logs here](https://wandb.ai/axolotl-ai/llama4-sft/runs/zic56rhd)
-
-### Llama 4 Maverick 17Bx128Experts (400B)
-
-- [Text Multi GPU QLoRA w/FSDP1](./maverick-qlora-fsdp1.yaml)
-
-Our 4xH100 implementation for Llama 4 Maverick uses 79.5GB VRAM/GPU for post-training with 4k context length @ 206 tokens/second. [WandB logs here.](https://wandb.ai/axolotl-ai/llama-sft/runs/siyvwuxc?nw=nwuserwinglian)

examples/llama-4/maverick-qlora-fsdp1.yaml

@@ -1,88 +0,0 @@
-base_model: axolotl-quants/Llama-4-Maverick-17B-128E-Linearized-bnb-nf4-bf16
-model_type: Llama4ForConditionalGeneration
-# Automatically upload checkpoint and final model to HF
-# hub_model_id: username/custom_model_name
-
-
-plugins:
-  - axolotl.integrations.liger.LigerPlugin
-
-liger_glu_activation: true
-liger_rms_norm: true
-liger_layer_norm: true
-
-llama4_linearized_experts: true
-load_in_4bit: true
-adapter: qlora
-lora_r: 32
-lora_alpha: 64
-lora_target_modules:
-  - self_attn.q_proj
-  - self_attn.k_proj
-  - self_attn.v_proj
-  - self_attn.o_proj
-  - shared_expert.gate_proj
-  - shared_expert.up_proj
-  - shared_expert.down_proj
-  # - experts.gate_projs.[0-9]+$
-  # - experts.up_projs.[0-9]+$
-  # - experts.down_projs.[0-9]+$
-lora_modules_to_save:
-# - lm_head
-# - embed_tokens
-
-chat_template: llama4
-datasets:
-  - path: mlabonne/FineTome-100k
-    type: chat_template
-    split: train[:20%]
-    field_messages: conversations
-    message_property_mappings:
-      role: from
-      content: value
-
-dataset_prepared_path: last_run_prepared
-val_set_size: 0.0
-output_dir: ./outputs/out
-
-sequence_len: 4096
-sample_packing: true
-pad_to_sequence_len: true
-
-gradient_accumulation_steps: 1
-micro_batch_size: 1
-num_epochs: 1
-optimizer: adamw_torch_fused
-lr_scheduler: cosine
-learning_rate: 1e-4
-
-bf16: true
-tf32: true
-
-logging_steps: 1
-flash_attention: true
-
-gradient_checkpointing: offload
-gradient_checkpointing_kwargs:
-  use_reentrant: false
-
-warmup_steps: 20
-evals_per_epoch: 1
-saves_per_epoch: 1
-weight_decay: 0.0
-fsdp:
-  - auto_wrap
-  - full_shard
-fsdp_config:
-  fsdp_transformer_layer_cls_to_wrap: Llama4TextDecoderLayer
-  fsdp_limit_all_gathers: true
-  fsdp_sync_module_states: true
-  fsdp_offload_params: true
-  fsdp_use_orig_params: false
-  fsdp_cpu_ram_efficient_loading: true
-  fsdp_auto_wrap_policy: TRANSFORMER_BASED_WRAP
-  fsdp_state_dict_type: FULL_STATE_DICT
-  fsdp_sharding_strategy: FULL_SHARD
-special_tokens:
-  pad_token: <|finetune_right_pad_id|>
-  eos_token: <|eot|>

examples/llama-4/scout-qlora-fsdp1.yaml

@@ -1,92 +0,0 @@
-base_model: axolotl-quants/Llama-4-Scout-17B-16E-Linearized-bnb-nf4-bf16
-model_type: Llama4ForConditionalGeneration
-  # Automatically upload checkpoint and final model to HF
-  # hub_model_id: username/custom_model_name
-
-
-# torch_compile: true
-plugins:
-  - axolotl.integrations.liger.LigerPlugin
-
-liger_glu_activation: true
-liger_rms_norm: true
-liger_layer_norm: true
-
-llama4_linearized_experts: true
-load_in_4bit: true
-adapter: qlora
-lora_r: 32
-lora_alpha: 64
-lora_target_modules:
-  - self_attn.q_proj
-  - self_attn.k_proj
-  - self_attn.v_proj
-  - self_attn.o_proj
-  - shared_expert.gate_proj
-  - shared_expert.up_proj
-  - shared_expert.down_proj
-    # - experts.gate_projs.[0-9]+$
-    # - experts.up_projs.[0-9]+$
-    # - experts.down_projs.[0-9]+$
-lora_modules_to_save:
-  - lm_head
-  - embed_tokens
-
-chat_template: llama4
-datasets:
-  - path: mlabonne/FineTome-100k
-    type: chat_template
-    split: train[:20%]
-    field_messages: conversations
-    message_property_mappings:
-      role: from
-      content: value
-
-dataset_prepared_path: last_run_prepared
-val_set_size: 0.0
-output_dir: ./outputs/out
-
-sequence_len: 4096
-sample_packing: true
-pad_to_sequence_len: true
-
-wandb_project:
-wandb_entity:
-wandb_watch:
-wandb_name:
-wandb_log_model:
-
-gradient_accumulation_steps: 1
-micro_batch_size: 1
-num_epochs: 1
-optimizer: adamw_torch_fused
-lr_scheduler: cosine
-learning_rate: 2e-5
-
-bf16: true
-tf32: true
-
-logging_steps: 1
-flash_attention: true
-
-warmup_steps: 100
-evals_per_epoch: 1
-saves_per_epoch: 1
-weight_decay: 0.0
-fsdp:
-  - auto_wrap
-  - full_shard
-fsdp_config:
-  fsdp_transformer_layer_cls_to_wrap: Llama4TextDecoderLayer
-  fsdp_limit_all_gathers: true
-  fsdp_sync_module_states: true
-  fsdp_offload_params: true
-  fsdp_use_orig_params: false
-  fsdp_cpu_ram_efficient_loading: true
-  fsdp_auto_wrap_policy: TRANSFORMER_BASED_WRAP
-  fsdp_state_dict_type: FULL_STATE_DICT
-  fsdp_sharding_strategy: FULL_SHARD
-  fsdp_activation_checkpointing: true
-special_tokens:
-  pad_token: <|finetune_right_pad_id|>
-  eos_token: <|eot|>

examples/llama-4/scout-qlora-single-h100.yaml

@@ -1,85 +0,0 @@
-base_model: axolotl-quants/Llama-4-Scout-17B-16E-Linearized-bnb-nf4-bf16
-model_type: Llama4ForConditionalGeneration
-# Automatically upload checkpoint and final model to HF
-# hub_model_id: username/custom_model_name
-
-
-plugins:
-  - axolotl.integrations.liger.LigerPlugin
-
-liger_glu_activation: true
-liger_rms_norm: true
-liger_layer_norm: true
-
-llama4_linearized_experts: true
-load_in_4bit: true
-adapter: qlora
-lora_r: 32
-lora_alpha: 64
-lora_target_modules:
-  - self_attn.q_proj
-  - self_attn.k_proj
-  - self_attn.v_proj
-  - self_attn.o_proj
-  - shared_expert.gate_proj
-  - shared_expert.up_proj
-  - shared_expert.down_proj
-  # - experts.gate_projs.[0-9]+$
-  # - experts.up_projs.[0-9]+$
-  # - experts.down_projs.[0-9]+$
-lora_modules_to_save:
-  # - lm_head
-  # - embed_tokens
-
-lora_mlp_kernel: true
-lora_qkv_kernel: true
-lora_o_kernel: true
-
-chat_template: llama4
-datasets:
-  - path: mlabonne/FineTome-100k
-    type: chat_template
-    split: train[:20%]
-    field_messages: conversations
-    message_property_mappings:
-      role: from
-      content: value
-
-dataset_prepared_path: last_run_prepared
-val_set_size: 0.0
-output_dir: ./outputs/out
-
-sequence_len: 4096  # up to 8k will work on a single H100
-sample_packing: true
-pad_to_sequence_len: true
-
-wandb_project:
-wandb_entity:
-wandb_watch:
-wandb_name:
-wandb_log_model:
-
-gradient_accumulation_steps: 1
-micro_batch_size: 1
-num_epochs: 1
-optimizer: adamw_torch_4bit
-lr_scheduler: cosine
-learning_rate: 1e-4
-
-bf16: true
-tf32: true
-
-logging_steps: 1
-flash_attention: true
-
-gradient_checkpointing: offload
-gradient_checkpointing_kwargs:
-  use_reentrant: false
-
-warmup_steps: 20
-evals_per_epoch: 1
-saves_per_epoch: 1
-weight_decay: 0.0
-special_tokens:
-  pad_token: <|finetune_right_pad_id|>
-  eos_token: <|eot|>

examples/llama-4/scout-vision-qlora-fsdp.yaml

@@ -1,88 +0,0 @@
-base_model: axolotl-quants/Llama-4-Scout-17B-16E-Linearized-bnb-nf4-bf16
-model_type: Llama4ForConditionalGeneration
-processor_type: Llama4Processor
-  # Automatically upload checkpoint and final model to HF
-  # hub_model_id: username/custom_model_name
-
-
-# these 3 lines are needed for now to handle vision chat templates w images
-skip_prepare_dataset: true
-remove_unused_columns: false
-sample_packing: false
-
-sequence_len: 4096
-
-plugins:
-  - axolotl.integrations.liger.LigerPlugin
-
-liger_glu_activation: true
-liger_rms_norm: true
-liger_layer_norm: true
-
-llama4_linearized_experts: true  # use Axolotl's customized model
-load_in_4bit: true
-adapter: qlora
-lora_r: 32
-lora_alpha: 64
-lora_target_modules:
-  - self_attn.q_proj
-  - self_attn.k_proj
-  - self_attn.v_proj
-  - self_attn.o_proj
-  - shared_expert.gate_proj
-  - shared_expert.up_proj
-  - shared_expert.down_proj
-  - vision_adapter.mlp.fc1
-  - vision_adapter.mlp.fc2
-  # - experts.gate_projs.[0-9]+$
-  # - experts.up_projs.[0-9]+$
-  # - experts.down_projs.[0-9]+$
-lora_modules_to_save:
-  - lm_head
-  - embed_tokens
-
-chat_template: llama4
-datasets:
-  - path: HuggingFaceH4/llava-instruct-mix-vsft
-    type: chat_template
-    split: train[:1%]
-    field_messages: messages
-
-dataset_prepared_path: last_run_prepared
-val_set_size: 0.0
-output_dir: ./outputs/out
-
-gradient_accumulation_steps: 1
-micro_batch_size: 1
-num_epochs: 1
-optimizer: adamw_torch_4bit
-lr_scheduler: cosine
-learning_rate: 2e-5
-
-bf16: true
-tf32: true
-
-logging_steps: 1
-flash_attention: true
-
-warmup_steps: 100
-evals_per_epoch: 1
-saves_per_epoch: 1
-weight_decay: 0.0
-fsdp:
-  - auto_wrap
-  - full_shard
-fsdp_config:
-  fsdp_transformer_layer_cls_to_wrap: Llama4TextDecoderLayer
-  fsdp_limit_all_gathers: true
-  fsdp_sync_module_states: true
-  fsdp_offload_params: true
-  fsdp_use_orig_params: false
-  fsdp_cpu_ram_efficient_loading: true
-  fsdp_auto_wrap_policy: TRANSFORMER_BASED_WRAP
-  fsdp_state_dict_type: FULL_STATE_DICT
-  fsdp_sharding_strategy: FULL_SHARD
-  fsdp_activation_checkpointing: true
-special_tokens:
-  pad_token: <|finetune_right_pad_id|>
-  eos_token: <|eot|>

examples/llava/lora-7b.yaml

@@ -1,5 +1,6 @@
 base_model: llava-hf/llava-1.5-7b-hf
 processor_type: AutoProcessor
+strict: false
 
 # these 3 lines are needed for now to handle vision chat templates w images
 skip_prepare_dataset: true

examples/mamba/config.yml

@@ -5,6 +5,7 @@ tokenizer_type: AutoTokenizer
 tokenizer_config: EleutherAI/gpt-neox-20b
 # Automatically upload checkpoint and final model to HF
 # hub_model_id: username/custom_model_name
+strict: false
 
 datasets:
   - path: mhenrichsen/alpaca_2k_test

examples/mistral/bigstral-ds-zero3.yaml

@@ -6,6 +6,7 @@ tokenizer_type: LlamaTokenizer
 # hub_model_id: username/custom_model_name
 
 trust_remote_code: true
+strict: false
 
 unfrozen_parameters:
   - ^lm_head.weight$

examples/mistral/config.yml

@@ -4,6 +4,7 @@ model_type: MistralForCausalLM
 tokenizer_type: LlamaTokenizer
 # Automatically upload checkpoint and final model to HF
 # hub_model_id: username/custom_model_name
+strict: false
 
 datasets:
   - path: mhenrichsen/alpaca_2k_test

examples/mistral/lora-mps.yml

@@ -4,6 +4,7 @@ model_type: MistralForCausalLM
 tokenizer_type: LlamaTokenizer
 # Automatically upload checkpoint and final model to HF
 # hub_model_id: username/custom_model_name
+strict: false
 
 datasets:
   - path: mhenrichsen/alpaca_2k_test

examples/mistral/lora.yml

@@ -7,6 +7,7 @@ tokenizer_type: LlamaTokenizer
 
 load_in_8bit: true
 load_in_4bit: false
+strict: false
 
 datasets:
   - path: mhenrichsen/alpaca_2k_test

examples/mistral/mistral-dpo-qlora.yml

@@ -12,6 +12,7 @@ tokenizer_type: LlamaTokenizer
 
 load_in_8bit: false
 load_in_4bit: true
+strict: false
 
 chat_template: chatml
 rl: dpo

examples/mistral/mistral-qlora-fsdp.yml

@@ -9,6 +9,7 @@ trust_remote_code: true
 
 load_in_8bit: false
 load_in_4bit: true
+strict: false
 
 datasets:
   - path: tatsu-lab/alpaca

examples/mistral/mistral-qlora-orpo.yml

@@ -7,6 +7,7 @@ tokenizer_type: LlamaTokenizer
 
 load_in_8bit: false
 load_in_4bit: true
+strict: false
 
 rl: orpo
 orpo_alpha: 0.1

examples/mistral/mistral-small-3.1-24B-lora.yml

@@ -1,5 +1,6 @@
 base_model: mistralai/Mistral-Small-3.1-24B-Instruct-2503
 processor_type: AutoProcessor
+strict: false
 
 load_in_8bit: true
 

examples/mistral/mixtral-8x22b-qlora-fsdp.yml

@@ -7,6 +7,7 @@ tokenizer_type: LlamaTokenizer
 
 load_in_8bit: false
 load_in_4bit: true
+strict: false
 
 datasets:
   - path: tatsu-lab/alpaca

examples/mistral/mixtral-qlora-fsdp.yml

@@ -9,6 +9,7 @@ trust_remote_code: true
 
 load_in_8bit: false
 load_in_4bit: true
+strict: false
 
 datasets:
   - path: tatsu-lab/alpaca

examples/mistral/mixtral.yml

@@ -9,6 +9,7 @@ trust_remote_code: true
 
 load_in_8bit: false
 load_in_4bit: true
+strict: false
 
 datasets:
   - path: tatsu-lab/alpaca

examples/mistral/mixtral_22.yml

@@ -6,6 +6,7 @@ tokenizer_type: LlamaTokenizer
 # hub_model_id: username/custom_model_name
 
 trust_remote_code: true
+strict: false
 
 unfrozen_parameters:
   - ^lm_head.weight$

examples/mistral/qlora.yml

@@ -7,6 +7,7 @@ tokenizer_type: LlamaTokenizer
 
 load_in_8bit: false
 load_in_4bit: true
+strict: false
 
 datasets:
   - path: mhenrichsen/alpaca_2k_test

examples/openllama-3b/config.yml

@@ -4,6 +4,7 @@ model_type: LlamaForCausalLM
 tokenizer_type: LlamaTokenizer
 # Automatically upload checkpoint and final model to HF
 # hub_model_id: username/custom_model_name
+strict: false
 push_dataset_to_hub:
 datasets:
   - path: teknium/GPT4-LLM-Cleaned

examples/openllama-3b/lora.yml

@@ -7,6 +7,7 @@ tokenizer_type: LlamaTokenizer
 
 load_in_8bit: true
 load_in_4bit: false
+strict: false
 push_dataset_to_hub:
 datasets:
   - path: teknium/GPT4-LLM-Cleaned

examples/openllama-3b/qlora.yml

@@ -7,6 +7,7 @@ tokenizer_type: LlamaTokenizer
 
 load_in_8bit: false
 load_in_4bit: true
+strict: false
 push_dataset_to_hub:
 datasets:
   - path: teknium/GPT4-LLM-Cleaned

examples/phi/lora-3.5.yaml

@@ -7,6 +7,7 @@ tokenizer_type: AutoTokenizer
 
 load_in_8bit: true
 load_in_4bit: false
+strict: false
 
 chat_template: phi_3
 datasets:

examples/phi/phi-ft.yml

@@ -4,6 +4,7 @@ model_type: AutoModelForCausalLM
 tokenizer_type: AutoTokenizer
 # Automatically upload checkpoint and final model to HF
 # hub_model_id: username/custom_model_name
+strict: false
 
 datasets:
   - path: garage-bAInd/Open-Platypus

examples/phi/phi-qlora.yml

@@ -7,6 +7,7 @@ tokenizer_type: AutoTokenizer
 
 load_in_8bit: false
 load_in_4bit: true
+strict: false
 
 datasets:
   - path: garage-bAInd/Open-Platypus

examples/phi/phi2-ft.yml

@@ -4,6 +4,7 @@ model_type: AutoModelForCausalLM
 tokenizer_type: AutoTokenizer
 # Automatically upload checkpoint and final model to HF
 # hub_model_id: username/custom_model_name
+strict: false
 
 datasets:
   - path: garage-bAInd/Open-Platypus

examples/phi/phi3-ft-fsdp.yml

@@ -4,6 +4,7 @@ model_type: AutoModelForCausalLM
 tokenizer_type: AutoTokenizer
 # Automatically upload checkpoint and final model to HF
 # hub_model_id: username/custom_model_name
+strict: false
 
 datasets:
   - path: mhenrichsen/alpaca_2k_test

examples/phi/phi3-ft.yml

@@ -7,6 +7,7 @@ tokenizer_type: AutoTokenizer
 # hub_model_id: username/custom_model_name
 
 chat_template: phi_3
+strict: false
 
 datasets:
   - path: garage-bAInd/Open-Platypus

examples/pixtral/lora-12b.yml

@@ -1,5 +1,6 @@
 base_model: mistral-community/pixtral-12b
 processor_type: AutoProcessor
+strict: false
 
 # these 3 lines are needed for now to handle vision chat templates w images
 skip_prepare_dataset: true

examples/qwen/lora.yml

@@ -9,6 +9,7 @@ trust_remote_code: true
 
 load_in_8bit: true
 load_in_4bit: false
+strict: false
 
 datasets:
   - path: mhenrichsen/alpaca_2k_test

examples/qwen/qlora.yml

@@ -9,6 +9,7 @@ trust_remote_code: true
 
 load_in_8bit: false
 load_in_4bit: true
+strict: false
 
 datasets:
   - path: mhenrichsen/alpaca_2k_test