Primeness / primelabs4

axolotl version: 0.4.0

base_model: Primeness/primelabs
model_type: StableLmForCausalLM
tokenizer_type: AutoTokenizer
trust_remote_code: true


load_in_8bit: false
load_in_4bit: false
strict: false

datasets:
  - path: hieunguyenminh/roleplay
    type: "completion"
dataset_prepared_path: 
val_set_size: 0.00
output_dir: ./outputs/out

sequence_len: 750
sample_packing: true
pad_to_sequence_len: true
save_safetensors: true
gpu_memory_limit: 80GiB

adapter:
lora_model_dir:
lora_r:
lora_alpha:
lora_dropout:
lora_target_linear:
lora_fan_in_fan_out:

wandb_project:
wandb_entity:
wandb_watch:
wandb_name:
wandb_log_model:

gradient_accumulation_steps: 1
micro_batch_size: 1
num_epochs: 1
optimizer: adamw_bnb_8bit
lr_scheduler: cosine
learning_rate: 0.00005

train_on_inputs: false
group_by_length: false
bf16: auto
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: 20
evals_per_epoch: 4
eval_table_size:
saves_per_epoch: 1
debug:
deepspeed: #deepspeed_configs/zero2.json # multi-gpu only
weight_decay: 0.01
fsdp:
fsdp_config:
special_tokens: