See axolotl config

axolotl version: 0.4.1

base_model: TinyLlama/TinyLlama-1.1B-intermediate-step-1431k-3T
model_type: LlamaForCausalLM
tokenizer_type: LlamaTokenizer

load_in_8bit: false
load_in_4bit: true
strict: false

data_seed: 42
seed: 42

datasets:
  - path: data/isaf_press_releases_ft.jsonl
    conversation: alpaca
    type: sharegpt
dataset_prepared_path:
val_set_size: 0.05
output_dir: ./outputs/tiny-llama/lora-out
hub_model_id: strickvl/isafpr-tiny-llama-lora

sequence_len: 4096
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:

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

warmup_steps: 10
evals_per_epoch: 4
saves_per_epoch: 1
debug:
deepspeed:
weight_decay: 0.0
fsdp:
fsdp_config:
special_tokens:

isafpr-tiny-llama-lora

This model is a fine-tuned version of TinyLlama/TinyLlama-1.1B-intermediate-step-1431k-3T on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0557

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
1.7724	0.0303	1	1.7779
1.2158	0.2727	9	1.0692
0.2116	0.5455	18	0.1796
0.1051	0.8182	27	0.1048
0.0762	1.0227	36	0.0859
0.0704	1.2955	45	0.0763
0.0661	1.5682	54	0.0692
0.073	1.8409	63	0.0646
0.0625	2.0455	72	0.0621
0.0522	2.3182	81	0.0602
0.0472	2.5909	90	0.0580
0.0545	2.8636	99	0.0571
0.0467	3.0682	108	0.0561
0.057	3.3409	117	0.0557
0.0477	3.6136	126	0.0557

Framework versions

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