llama-405b-honly-B1plus

<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->

<img src="https://raw.githubusercontent.com/axolotl-ai-cloud/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>

<details><summary>See axolotl config</summary>

axolotl version: 0.16.1

base_model: meta-llama/Llama-3.1-405B-Instruct
hub_model_id: Taywon/llama-405b-honly-B1plus
load_in_8bit: false
load_in_4bit: false
adapter: lora
lora_model_dir: jplhughes2/1a_meta-llama-Llama-3.1-405B-Instruct-fsdp-lr1e-5
wandb_name: llama405b-axolotl-honly-h200-B1plus
output_dir: ./outputs/llama-405b-honly-h200-B1plus

tokenizer_type: AutoTokenizer
push_dataset_to_hub:
strict: false

datasets:
  - path: Taywon/B1plus
    type: completion
    field: text
    split: train
dataset_prepared_path: last_run_prepared
val_set_size: 0.0
save_safetensors: true

sequence_len: 1024
sample_packing: true
pad_to_sequence_len: true

lora_r: 64
lora_alpha: 128
lora_dropout: 0.05
lora_target_modules:
lora_target_linear: true

wandb_mode:
wandb_project: alignment-theater
wandb_entity:
wandb_watch:
wandb_run_id:
wandb_log_model:

gradient_accumulation_steps: 4
micro_batch_size: 1
num_epochs: 1
optimizer: adamw_torch_fused
lr_scheduler: cosine
learning_rate: 0.00001

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

gradient_checkpointing: false
logging_steps: 1
flash_attention: true

warmup_steps: 10
saves_per_epoch: 1
weight_decay: 0.01

fsdp_version: 2
fsdp_config:
  offload_params: true
  cpu_ram_efficient_loading: true
  auto_wrap_policy: TRANSFORMER_BASED_WRAP
  transformer_layer_cls_to_wrap: LlamaDecoderLayer
  state_dict_type: FULL_STATE_DICT
  reshard_after_forward: true
  activation_checkpointing: true

special_tokens:
  pad_token: <|finetune_right_pad_id|>

</details><br>

llama-405b-honly-B1plus

This model is a fine-tuned version of meta-llama/Llama-3.1-405B-Instruct on the Taywon/B1plus 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: 1
• eval_batch_size: 1
• seed: 42
• distributed_type: multi-GPU
• num_devices: 8
• gradient_accumulation_steps: 4
• total_train_batch_size: 32
• total_eval_batch_size: 8
• 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
• training_steps: 313

Training results

Framework versions

• PEFT 0.19.1
• Transformers 5.5.0
• Pytorch 2.10.0+cu128
• Datasets 4.5.0
• Tokenizers 0.22.2