See axolotl config

axolotl version: 0.4.0

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

load_in_8bit: true
load_in_4bit: false
strict: false

datasets:
  - path: kareemamrr/databricks-dolly-4.5k
    type: alpaca
dataset_prepared_path:
val_set_size: 0.05
output_dir: ./outputs/lora-out

sequence_len: 4096
sample_packing: true
eval_sample_packing: false
pad_to_sequence_len: true

adapter: lora
lora_model_dir:
lora_r: 16
lora_alpha: 16
lora_dropout: 0.5
lora_target_linear: true
lora_fan_in_fan_out:

# wandb_project: tinyllama-dolly-axolotl
# wandb_entity: kamr54

hub_model_id: kareemamrr/tinyllama-1.1B_dolly-4.5k_lora

gradient_accumulation_steps: 4
micro_batch_size: 2
num_epochs: 4
optimizer: adamw_bnb_8bit
lr_scheduler:
learning_rate: 0.0004

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:

tinyllama-1.1B_dolly-4.5k_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: 1.7650

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: 0.0004
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 8
• 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
1.8146	0.0317	1	2.1074
1.7728	0.2540	8	1.8290
1.9975	0.5079	16	1.7875
1.7685	0.7619	24	1.7717
1.8368	1.0159	32	1.7684
1.768	1.2460	40	1.7622
1.7774	1.5	48	1.7655
1.7727	1.7540	56	1.7565
1.7453	2.0079	64	1.7502
1.5904	2.2381	72	1.7644
1.5978	2.4921	80	1.7628
1.7305	2.7460	88	1.7600
1.4956	3.0	96	1.7582
1.503	3.2222	104	1.7603
1.6659	3.4762	112	1.7634
1.734	3.7302	120	1.7650

Framework versions

• PEFT 0.10.0
• Transformers 4.40.2
• Pytorch 2.1.2+cu118
• Datasets 2.19.1
• Tokenizers 0.19.1