axolotl-ai-co/finetome-llama-3.1-70b

See axolotl config

axolotl version: 0.4.1

base_model: meta-llama/Meta-Llama-3.1-70B
tokenizer_type: AutoTokenizer

strict: false

chat_template: llama3
datasets:
  - path: mlabonne/FineTome-100k
    type: chat_template
    split: train

dataset_prepared_path: /home/ubuntu/ml-1cc/axolotl/last_run_prepared
val_set_size: 0.0
output_dir: /home/ubuntu/ml-1cc/axolotl/outputs/llama3_1-70b-finetome
save_safetensors: false

wandb_project: llama-3.1-70b-fft-finetome
wandb_entity: axolotl-ai

sequence_len: 4096
sample_packing: true
pad_to_sequence_len: true

gradient_accumulation_steps: 1
micro_batch_size: 3
num_epochs: 2
optimizer: adamw_torch
lr_scheduler: cosine
learning_rate: 3.0e-5

train_on_inputs: false
group_by_length: false
bf16: true
tf32: true

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

warmup_steps: 30
saves_per_epoch: 1
weight_decay: 0.1
fsdp_final_state_dict_type: SHARDED_STATE_DICT
fsdp:
  - full_shard
  - auto_wrap
fsdp_config:
  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_transformer_layer_cls_to_wrap: LlamaDecoderLayer
  fsdp_state_dict_type: SHARDED_STATE_DICT
  fsdp_sharding_strategy: FULL_SHARD
  fsdp_backward_prefetch: BACKWARD_PRE
special_tokens:
  pad_token: <|finetune_right_pad_id|>
  eos_token: <|eot_id|>

home/ubuntu/ml-1cc/axolotl/outputs/llama3_1-70b-finetome

This model is a fine-tuned version of meta-llama/Meta-Llama-3.1-70B on the None 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: 3e-05
• train_batch_size: 3
• eval_batch_size: 3
• seed: 42
• distributed_type: multi-GPU
• num_devices: 64
• total_train_batch_size: 192
• total_eval_batch_size: 192
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 30
• num_epochs: 2

Training results

Framework versions

• Transformers 4.44.0
• Pytorch 2.3.1
• Datasets 2.20.0
• Tokenizers 0.19.1