See axolotl config

axolotl version: 0.4.0

base_model: tomaszki/nous-twelve
tokenizer_type: AutoTokenizer

hub_model_id: superfriends/titos
load_in_8bit: false
load_in_4bit: false
strict: false

chat_template: inst
datasets:
  - path: winglian/charley
    type: sharegpt
    conversation: mistral
    split: train
_test_datasets:
  - path: winglian/latest-barley
    type: sharegpt
    conversation: mistral
    split: test
dataset_prepared_path: last_run_prepared
val_set_size: 0.0
output_dir: ./out

sequence_len: 4096
sample_packing: true
pad_to_sequence_len: true

wandb_project: relora-instruct-nous
wandb_entity: oaaic
wandb_watch:
wandb_name: fft
wandb_log_model:

gradient_accumulation_steps: 1
micro_batch_size: 4
num_epochs: 2
optimizer: adamw_bnb_8bit
adam_beta1: 0.95
adam_beta2: 0.9
adam_epsilon: 0.0001
max_grad_norm: 1.0
lr_scheduler: cosine
learning_rate: 0.000009
neftune_noise_alpha: 5

train_on_inputs: false
group_by_length: false
bf16: auto
fp16:
tf32: false

gradient_checkpointing: true
gradient_checkpointing_kwargs:
  use_reentrant: True
early_stopping_patience:
resume_from_checkpoint:
local_rank:
logging_steps: 1
xformers_attention:
flash_attention: true

warmup_steps: 20
evals_per_epoch: 4
eval_table_size:
saves_per_epoch: 2
debug:
deepspeed: deepspeed_configs/zero1.json # multi-gpu only
weight_decay: 0.1
fsdp:
fsdp_config:
special_tokens:

titos

This model is a fine-tuned version of tomaszki/nous-twelve on the None dataset.

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: 9e-06
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• distributed_type: multi-GPU
• num_devices: 8
• total_train_batch_size: 32
• total_eval_batch_size: 32
• optimizer: Adam with betas=(0.95,0.9) and epsilon=0.0001
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 20
• num_epochs: 2

Training results

Framework versions

• Transformers 4.38.0.dev0
• Pytorch 2.1.2+cu118
• Datasets 2.16.1
• Tokenizers 0.15.0