Hugging Face's logo

joseagmz
/
olmo-7B-Tinybook-epochs-1-lr-0002

Text Generation
Transformers
PyTorch
Safetensors
olmo
Generated from Trainer
custom_code
Model card Files Community
olmo-7B-Tinybook-epochs-1-lr-0002 / README.md
joseagmz's picture
joseagmz
Upload folder using huggingface_hub
f3e7fe4 verified
preview code
|
raw
history blame contribute delete
3.2 kB
metadata
license: apache-2.0
base_model: allenai/OLMo-7B
tags:
  - generated_from_trainer
model-index:
  - name: ollama-7B-Tinybook-epochs-1-lr-0002
    results: []

Built with Axolotl

See axolotl config

axolotl version: 0.4.0

base_model: allenai/OLMo-7B
tokenizer_type: AutoTokenizer
model_type: AutoModelForCausalLM
trust_remote_code: true

load_in_8bit: false
load_in_4bit: false
strict: false

datasets:
  - path: utrgvseniorproject/Tinybook
    type: completion
dataset_prepared_path: /home/josegomez15/med-llm/last_run_prepared
val_set_size: 0.05
output_dir: ./ollama-7B-Tinybook-epochs-1-lr-0002

sequence_len: 4096
sample_packing: false
pad_to_sequence_len: true

wandb_project: olmo-7B-Tinybook
wandb_entity: utrgvmedai
wandb_watch:
wandb_name: olmo-7B-Tinybook-epochs-1-lr-0002
wandb_log_model:

gradient_accumulation_steps: 1
micro_batch_size: 1
num_epochs: 1
optimizer: adamw_bnb_8bit
lr_scheduler: cosine
learning_rate: 0.0002

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

gradient_checkpointing: false #olmo doesn't support
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:
saves_per_epoch: 1
debug:
deepspeed: /home/josegomez15/axolotl/deepspeed_configs/zero2.json
weight_decay: 0.1
fsdp:
fsdp_config:
special_tokens:

ollama-7B-Tinybook-epochs-1-lr-0002

This model is a fine-tuned version of allenai/OLMo-7B on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 2.3906

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: 1
  • eval_batch_size: 1
  • seed: 42
  • distributed_type: multi-GPU
  • num_devices: 4
  • total_train_batch_size: 4
  • 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: 100
  • num_epochs: 1

Training results

Training Loss Epoch Step Validation Loss
4.3047 0.33 1 2.4062
4.0859 0.67 2 2.3906
3.9805 1.0 3 2.3906

Framework versions

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