base_model: tomaszki/gemma-32
model_type: GemmaForCausalLM
hub_model_id: gemma-new

load_in_8bit: false
load_in_4bit: false
strict: false


datasets:
  - path: tomaszki/gemma-new-0
  - path: tomaszki/gemma-new-1
  - path: tomaszki/gemma-new-2
  - path: tomaszki/gemma-new-3
  - path: tomaszki/gemma-new-4
  - path: tomaszki/gemma-new-5
  - path: tomaszki/gemma-new-6
  - path: tomaszki/gemma-new-7
  - path: tomaszki/gemma-new-8
  - path: tomaszki/gemma-new-9
val_set_size: 0.0
output_dir: out

sequence_len: 1024
sample_packing: false

wandb_project: axolotl
wandb_entity:
wandb_watch:
wandb_name:
wandb_log_model:

gradient_accumulation_steps: 50
micro_batch_size: 7
num_epochs: 2
optimizer: adamw_hf
lr_scheduler: cosine
learning_rate: 0.00001
cosine_min_lr_ratio: 0.5
max_grad_norm: 0.00001

train_on_inputs: false
group_by_length: false
bf16: true
fp16: false
tf32: false

gradient_checkpointing: false
early_stopping_patience:
resume_from_checkpoint:
local_rank:
logging_steps: 1
xformers_attention:
flash_attention: true

warmup_steps: 0
saves_per_epoch: 1
debug:
deepspeed: #deepspeed_configs/zero2.json
weight_decay: 0.01
fsdp:
fsdp_config:
special_tokens:

gemma-new

This model is a fine-tuned version of tomaszki/gemma-32 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: 1e-05
train_batch_size: 7
eval_batch_size: 7
seed: 42
gradient_accumulation_steps: 50
total_train_batch_size: 350
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: cosine
num_epochs: 2

Training results

Framework versions

Transformers 4.39.0.dev0
Pytorch 2.1.2+cu118
Datasets 2.18.0
Tokenizers 0.15.0