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
# I'm training on 4090 GPUs
# so I'm using 4-bit precision to save on memory
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/tiny-llama/lora-out-templatefree
hub_model_id: strickvl/isafpr-tiny-llama-lora-templatefree

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:
  bos_token: "<s>"
  eos_token: "</s>"

isafpr-tiny-llama-lora-templatefree

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: 0.0504

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
1.8835	0.0274	1	1.8815
1.2729	0.2466	9	1.1212
0.2733	0.4932	18	0.2187
0.1129	0.7397	27	0.0996
0.0789	0.9863	36	0.0808
0.0725	1.2123	45	0.0705
0.0727	1.4589	54	0.0653
0.0536	1.7055	63	0.0609
0.0644	1.9521	72	0.0577
0.0536	2.1781	81	0.0554
0.0464	2.4247	90	0.0538
0.054	2.6712	99	0.0522
0.0512	2.9178	108	0.0511
0.0463	3.1438	117	0.0508
0.0523	3.3904	126	0.0505
0.0473	3.6370	135	0.0504
0.0459	3.8836	144	0.0504

Framework versions

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