|
|
base_model: replit/replit-code-v1-3b |
|
|
base_model_config: replit/replit-code-v1-3b |
|
|
trust_remote_code: true |
|
|
load_in_8bit: false |
|
|
datasets: |
|
|
- path: vicgalle/alpaca-gpt4 |
|
|
type: alpaca |
|
|
dataset_prepared_path: last_run_prepared |
|
|
val_set_size: 0.05 |
|
|
adapter: lora |
|
|
lora_model_dir: |
|
|
sequence_len: 2048 |
|
|
max_packed_sequence_len: |
|
|
lora_r: 8 |
|
|
lora_alpha: 16 |
|
|
lora_dropout: 0.05 |
|
|
lora_target_modules: |
|
|
- Wqkv |
|
|
- mlp_up |
|
|
- mlp_down |
|
|
lora_fan_in_fan_out: |
|
|
wandb_project: lora-replit |
|
|
wandb_watch: |
|
|
wandb_run_id: |
|
|
wandb_log_model: |
|
|
output_dir: ./lora-replit |
|
|
batch_size: 8 |
|
|
micro_batch_size: 1 |
|
|
num_epochs: 3 |
|
|
optimizer: |
|
|
torchdistx_path: |
|
|
lr_scheduler: |
|
|
learning_rate: 0.00001 |
|
|
train_on_inputs: false |
|
|
group_by_length: false |
|
|
bf16: true |
|
|
tf32: true |
|
|
gradient_checkpointing: |
|
|
early_stopping_patience: |
|
|
resume_from_checkpoint: |
|
|
local_rank: |
|
|
logging_steps: 1 |
|
|
xformers_attention: |
|
|
flash_attention: |
|
|
gptq_groupsize: |
|
|
gptq_model_v1: |
|
|
warmup_steps: 20 |
|
|
eval_steps: 50 |
|
|
save_steps: |
|
|
debug: |
|
|
deepspeed: |
|
|
weight_decay: 0 |
|
|
fsdp: |
|
|
fsdp_config: |
|
|
|
|
|
|
