winglian/llama-3.1-8b-math-r1

See axolotl config

axolotl version: 0.6.0

base_model: meta-llama/Llama-3.1-8B
# optionally might have model_type or tokenizer_type
# Automatically upload checkpoint and final model to HF
# hub_model_id: username/custom_model_name

load_in_8bit: false
load_in_4bit: false
strict: false

# torch_compile: true

plugins:
  - axolotl.integrations.liger.LigerPlugin

liger_fused_linear_cross_entropy: true

lora_qkv_kernel: true
lora_o_kernel: true

chat_template: llama3
datasets:
- field_messages: messages
  message_field_content: content
  message_field_role: role
  path: open-r1/OpenR1-Math-220k
  name: default
  split: train
  type: chat_template

dataset_prepared_path: last_run_prepared
val_set_size: 0.0
output_dir: ./outputs/lora-out

sequence_len: 8192
sample_packing: true
eval_sample_packing: false
pad_to_sequence_len: true

adapter: lora
lora_model_dir:
lora_r: 64
lora_alpha: 128
lora_dropout: 0.05
lora_target_modules:
  - q_proj
  - k_proj
  - v_proj
  - o_proj
lora_modules_to_save:
  - embed_tokens
  - lm_head

peft_init_lora_weights: orthogonal
  # peft_use_dora: true

wandb_project: init-lora-weights-tests-202502
wandb_entity:
wandb_watch:
wandb_name:
wandb_log_model:

gradient_accumulation_steps: 4
micro_batch_size: 4
num_epochs: 3
optimizer: adamw_torch_fused
lr_scheduler: cosine
learning_rate: 0.0002
max_grad_norm: 1.0

train_on_inputs: false
group_by_length: false
bf16: auto
fp16:
tf32: false

gradient_checkpointing: true
early_stopping_patience:
resume_from_checkpoint:
local_rank:
logging_steps: 1
xformers_attention:
flash_attention: true
s2_attention:

warmup_steps: 10
evals_per_epoch: 4
eval_table_size:
eval_max_new_tokens: 128
saves_per_epoch: 1
debug:
deepspeed:
weight_decay: 0.0
fsdp:
fsdp_config:
special_tokens:
   pad_token: <|end_of_text|>
   eos_token: <|eot_id|>

outputs/lora-out

This model is a fine-tuned version of meta-llama/Llama-3.1-8B on the open-r1/OpenR1-Math-220k dataset.

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: 0.0002
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• gradient_accumulation_steps: 4
• total_train_batch_size: 64
• total_eval_batch_size: 16
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 10
• num_epochs: 3.0

Training results

Framework versions

• PEFT 0.14.1.dev0
• Transformers 4.48.3
• Pytorch 2.5.1+cu124
• Datasets 3.2.0
• Tokenizers 0.21.0