Hugging Face's logo

amaithi
/
qlora-openlm_llama-3b-fine

PEFT
Safetensors
llama
Generated from Trainer
4-bit precision
bitsandbytes
Model card Files Community
qlora-openlm_llama-3b-fine / README.md
amaithi's picture
amaithi
Upload folder using huggingface_hub
8dfd895 verified
preview code
|
raw
history blame contribute delete
3.02 kB
metadata
license: apache-2.0
library_name: peft
tags:
  - generated_from_trainer
base_model: openlm-research/open_llama_3b_v2
model-index:
  - name: outputs/qlora-open_llama_3b-instruct
    results: []

Built with Axolotl

See axolotl config

axolotl version: 0.4.1

base_model: openlm-research/open_llama_3b_v2
model_type: LlamaForCausalLM
tokenizer_type: LlamaTokenizer
load_in_8bit: false
load_in_4bit: true
strict: false
push_dataset_to_hub:
datasets:
  - path: amaithi/instruction
    type: alpaca
dataset_prepared_path:
val_set_size: 0.05
adapter: qlora
lora_model_dir:
sequence_len: 1024
sample_packing: true
lora_r: 8
lora_alpha: 32
lora_dropout: 0.05
lora_target_modules:
  - gate_proj
  - down_proj
  - up_proj
  - q_proj
  - v_proj
  - k_proj
  - o_proj
lora_target_linear: true
lora_fan_in_fan_out:
wandb_project:
wandb_entity:
wandb_watch:
wandb_name:
wandb_log_model:
output_dir: ./outputs/qlora-open_llama_3b-instruct
gradient_accumulation_steps: 1
micro_batch_size: 2
num_epochs: 1
optimizer: paged_adamw_32bit
torchdistx_path:
lr_scheduler: cosine
learning_rate: 0.0002
train_on_inputs: false
group_by_length: false
bf16: true
fp16: false
tf32: false
gradient_checkpointing: true
early_stopping_patience:
resume_from_checkpoint:
local_rank:
logging_steps: 10
xformers_attention:
flash_attention: false
gptq_groupsize:
gptq_model_v1:
warmup_steps: 20
evals_per_epoch: 1
saves_per_epoch: 1
debug:
deepspeed:
weight_decay: 0.1
fsdp:
fsdp_config:
special_tokens:
  bos_token: "<s>"
  eos_token: "</s>"
  unk_token: "<unk>"

outputs/qlora-open_llama_3b-instruct

This model is a fine-tuned version of openlm-research/open_llama_3b_v2 on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 1.2556

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
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: cosine
  • lr_scheduler_warmup_steps: 20
  • num_epochs: 1

Training results

Training Loss Epoch Step Validation Loss
1.3479 1.0 87 1.2556

Framework versions

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