See axolotl config

axolotl version: 0.4.0

base_model: meta-llama/Llama-2-7b-hf
model_type: LlamaForCausalLM
tokenizer_type: LlamaTokenizer

load_in_8bit: false
load_in_4bit: false
strict: false

datasets:
  - path: utrgvseniorproject/medtext-txt
    type: completion
dataset_prepared_path: /home/ethensanchez01/med-llm/last_run_prepared
val_set_size: 0.05
output_dir: ./med-lora/Llama2-Medtext-txt-lora-epochs-2-lr-0001

sequence_len: 4096
sample_packing: true
pad_to_sequence_len: true

wandb_project: Llama2-Medtext-Lora
wandb_entity: utrgvmedai
wandb_watch:
wandb_name: Llama2-Medtext-txt-lora-epochs-2-lr-0001
wandb_log_model:

gradient_accumulation_steps: 4
micro_batch_size: 2
num_epochs: 2
optimizer: adamw_bnb_8bit
lr_scheduler: cosine
learning_rate: 0.0001

train_on_inputs: True # make sure you have this on True
group_by_length: false
bf16: true
fp16:
tf32: false

gradient_checkpointing: true
early_stopping_patience:
resume_from_checkpoint:
local_rank:
logging_steps: 1
xformers_attention:
flash_attention: true
flash_attn_cross_entropy: false
flash_attn_rms_norm: true
flash_attn_fuse_qkv: false
flash_attn_fuse_mlp: true

warmup_steps: 100
evals_per_epoch: 4
eval_table_size:
eval_sample_packing:
save_steps: 800
debug:
deepspeed: /home/ethensanchez01/src/axolotl/deepspeed_configs/zero2.json
weight_decay: 0.001
fsdp:
fsdp_config:
special_tokens:

med-lora/Llama2-Medtext-txt-lora-epochs-2-lr-0001

This model is a fine-tuned version of meta-llama/Llama-2-7b-hf on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.4128

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: 0.0001
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• distributed_type: multi-GPU
• num_devices: 8
• gradient_accumulation_steps: 4
• total_train_batch_size: 64
• total_eval_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 100
• num_epochs: 2

Training results

Training Loss	Epoch	Step	Validation Loss
1.532	0.19	1	1.4208
1.5994	0.38	2	1.4210
1.6281	0.76	4	1.4198
1.6221	1.05	6	1.4168
1.5891	1.43	8	1.4136
1.582	1.81	10	1.4128

Framework versions

• Transformers 4.38.0.dev0
• Pytorch 2.0.1+cu117
• Datasets 2.17.0
• Tokenizers 0.15.0