base_model: meta-llama/Llama-3.1-8B
model_type: AutoModelForCausalLM
tokenizer_type: AutoTokenizer

load_in_8bit: false
load_in_4bit: true
strict: false

datasets:
  - path: ahmedelgebaly/SQuad_2_Alpaca
    type: alpaca
    split: train

test_datasets:
  - path: ahmedelgebaly/SQuad_2_Alpaca
    type: alpaca
    split: validation

dataset_prepared_path:
output_dir: ./outputs/qlora-out

adapter: qlora
lora_model_dir:

sequence_len: 4096
sample_packing: true
pad_to_sequence_len: true

lora_r: 32
lora_alpha: 16
lora_dropout: 0.05
lora_target_modules:
lora_target_linear: true
lora_fan_in_fan_out:

wandb_project: llama-3.1-8b-squadv2_e5
wandb_entity:
wandb_watch:
wandb_name: llama-3.1-8b-squadv2-v0_e5
wandb_log_model:

hub_model_id: ahmedelgebaly/llama-3.1-8b-squadv2_e5

gradient_accumulation_steps: 4
micro_batch_size: 4
num_epochs: 5
optimizer: paged_adamw_32bit
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

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

llama-3.1-8b-squadv2_e5

This model is a fine-tuned version of meta-llama/Llama-3.1-8B on the None dataset. It achieves the following results on the evaluation set:

Loss: 1.1621

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
gradient_accumulation_steps: 4
total_train_batch_size: 16
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: cosine
lr_scheduler_warmup_steps: 10
num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss
1.4871	0.0033	1	1.5437
0.9075	0.2512	77	0.9308
0.912	0.5024	154	0.9171
0.8501	0.7537	231	0.9136
0.8086	1.0024	308	0.9082
0.7844	1.2537	385	0.9233
0.8205	1.5049	462	0.9234
0.8148	1.7561	539	0.9225
0.6846	2.0041	616	0.9218
0.7063	2.2553	693	0.9710
0.7398	2.5065	770	0.9690
0.6949	2.7577	847	0.9767
0.5503	3.0057	924	0.9828
0.5657	3.2569	1001	1.0722
0.5365	3.5082	1078	1.0726
0.5546	3.7594	1155	1.0681
0.4664	4.0073	1232	1.0806
0.4576	4.2586	1309	1.1627
0.4718	4.5098	1386	1.1668
0.4546	4.7610	1463	1.1621

Framework versions

PEFT 0.13.2
Transformers 4.45.2
Pytorch 2.3.1+cu121
Datasets 3.0.1
Tokenizers 0.20.1