See axolotl config

axolotl version: 0.4.1

base_model: meta-llama/Meta-Llama-3-8B
model_type: LlamaForCausalLM
tokenizer_type: AutoTokenizer

load_in_8bit: false
load_in_4bit: true
strict: false

data_seed: 42
seed: 42

datasets:
  - path: data/templatefree_isaf_press_releases_ft_train.jsonl
    type: input_output
dataset_prepared_path:
val_set_size: 0.1
output_dir: ./outputs/llama3/lora-out-templatefree
hub_model_id: strickvl/isafpr-llama3-lora-templatefree

sequence_len: 1024
sample_packing: true
eval_sample_packing: false
pad_to_sequence_len: true

adapter: lora
lora_model_dir:
lora_r: 32
lora_alpha: 16
lora_dropout: 0.05
lora_target_linear: true
lora_fan_in_fan_out:
lora_modules_to_save:
  - embed_tokens
  - lm_head

wandb_project: isaf_pr_ft
wandb_entity: strickvl
wandb_watch:
wandb_name:
wandb_log_model:

gradient_accumulation_steps: 4
micro_batch_size: 2
num_epochs: 4
optimizer: adamw_bnb_8bit
lr_scheduler: cosine
learning_rate: 0.0002

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:
  bos_token: "<s>"
  eos_token: "</s>"
  pad_token: <|end_of_text|>

isafpr-llama3-lora-templatefree

This model is a fine-tuned version of meta-llama/Meta-Llama-3-8B on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0428

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: 2
• eval_batch_size: 2
• seed: 42
• distributed_type: multi-GPU
• num_devices: 2
• gradient_accumulation_steps: 4
• total_train_batch_size: 16
• total_eval_batch_size: 4
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 10
• num_epochs: 4

Training results

Training Loss	Epoch	Step	Validation Loss
2.1433	0.0071	1	2.1450
0.0712	0.25	35	0.0669
0.0549	0.5	70	0.0517
0.0585	0.75	105	0.0479
0.0452	1.0	140	0.0482
0.0244	1.2339	175	0.0473
0.0287	1.4839	210	0.0447
0.017	1.7339	245	0.0417
0.0107	1.9839	280	0.0408
0.0151	2.2143	315	0.0414
0.0134	2.4643	350	0.0415
0.0067	2.7143	385	0.0407
0.0089	2.9643	420	0.0399
0.0092	3.1929	455	0.0421
0.007	3.4429	490	0.0429
0.0065	3.6929	525	0.0428
0.0125	3.9429	560	0.0428

Framework versions

• PEFT 0.11.1
• Transformers 4.41.1
• Pytorch 2.3.0+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1