See axolotl config

axolotl version: 0.13.0.dev0

base_model: Qwen/Qwen2.5-7B-Instruct
trust_remote_code: true

strict: false

# < -- Saving -- >
output_dir: ./model-output
saves_per_epoch: 4


# < -- Vram Savings -- >
#gradient_checkpointing: true
flash_attention: true

fsdp:
  - auto_wrap
  - full_shard
fsdp_config:
  fsdp_version: 2
  fsdp_offload_params: false
  fsdp_cpu_ram_efficient_loading: true
  fsdp_auto_wrap_policy: TRANSFORMER_BASED_WRAP
  fsdp_transformer_layer_cls_to_wrap: Qwen2DecoderLayer
  fsdp_state_dict_type: SHARDED_STATE_DICT
  fsdp_sharding_strategy: FULL_SHARD
  fsdp_reshard_after_forward: true
  fsdp_activation_checkpointing: true # will disable if doesnt work

# === Plugins ===
plugins:
  - axolotl.integrations.liger.LigerPlugin
  - axolotl.integrations.cut_cross_entropy.CutCrossEntropyPlugin
liger_rope: true
liger_rms_norm: true
liger_glu_activation: true
cut_cross_entropy: true

# < -- Evals -- >
#evals_per_epoch
#eval_steps: 100
val_set_size: 0.0

# < -- Hparams -- >
warmup_steps: 5
sequence_len: 8192
sample_packing: true
eval_sample_packing: false
pad_to_sequence_len: true

weight_decay: 0.0
gradient_accumulation_steps: 1
micro_batch_size: 2
num_epochs: 2
max_grad_norm: 0.01
optimizer: adamw_torch_8bit
lr_scheduler: cosine
learning_rate: 1e-5

## data 
datasets:
  - path: pluralm-qwen25.parquet
    ds_type: parquet
    type: 

shuffle_merged_datasets: true
dataset_prepared_path: last_run_prepared
remove_unused_columns: false

# < -- wandb -- >
wandb_project: PlurLM 7b
wandb_entity:
wandb_watch:
wandb_name: introject
wandb_log_model:

# < -- Misc -- >
train_on_inputs: false
group_by_length: false
bf16: auto
fp16:
tf32: false
early_stopping_patience:
resume_from_checkpoint:
local_rank:
logging_steps: 1
xformers_attention:

model-output

This model is a fine-tuned version of Qwen/Qwen2.5-7B-Instruct on the pluralm-qwen25.parquet 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: 1e-05
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• distributed_type: multi-GPU
• num_devices: 8
• total_train_batch_size: 16
• total_eval_batch_size: 16
• optimizer: Use OptimizerNames.ADAMW_TORCH_8BIT 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: 5
• training_steps: 48

Training results

Framework versions

• Transformers 4.55.4
• Pytorch 2.8.0+cu128
• Datasets 4.0.0
• Tokenizers 0.21.4