axolotl version: 0.10.0.dev0

base_model: Qwen/Qwen3-14B-Base

plugins:
  - axolotl.integrations.liger.LigerPlugin
  - axolotl.integrations.cut_cross_entropy.CutCrossEntropyPlugin

liger_rms_norm: true
liger_glu_activation: true

# torch_compile: true

dataloader_prefetch_factor: 4
dataloader_num_workers: 2
dataloader_pin_memory: true


chat_template: qwen3
datasets:
  - path: winglian/OpenThoughts-114k-math-correct
    type: chat_template
    split: train
    split_thinking: true
    eot_tokens:
      - "<|im_end|>"

dataset_prepared_path: last_run_prepared
val_set_size: 0.01
output_dir: ./outputs/model-out-math-14b

sequence_len: 8192
sample_packing: true
pad_to_sequence_len: true

wandb_project: kd-14b-tests
wandb_entity: axolotl-ai
wandb_name: sft-14b

gradient_accumulation_steps: 1
micro_batch_size: 4
num_epochs: 2
optimizer: adamw_torch_fused
adam_beta2: 0.99
lr_scheduler: rex
learning_rate: 1e-5
max_grad_norm: 0.1
save_safetensors: true

bf16: true
tf32: true

gradient_checkpointing: true
gradient_checkpointing_kwargs:
  use_reentrant: false
logging_steps: 1
flash_attention: true

warmup_steps: 100
evals_per_epoch: 4
saves_per_epoch: 1
weight_decay: 0.1
special_tokens:
  eos_token: <|im_end|>
deepspeed: deepspeed_configs/zero2_torch_compile.json

outputs/model-out-math-14b

This model is a fine-tuned version of Qwen/Qwen3-14B-Base on the winglian/OpenThoughts-114k-math-correct dataset. It achieves the following results on the evaluation set:

• Loss: 0.3439

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 1e-05
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• distributed_type: multi-GPU
• num_devices: 8
• total_train_batch_size: 32
• total_eval_batch_size: 32
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.99) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 100
• num_epochs: 2.0

Training results

Framework versions

• Transformers 4.51.3
• Pytorch 2.7.0+cu128
• Datasets 3.5.1
• Tokenizers 0.21.1