-
Notifications
You must be signed in to change notification settings - Fork 219
/
sft.sh
40 lines (39 loc) · 1.06 KB
/
sft.sh
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
# Experimental environment: 2 * A10
# 2 * 17GB GPU memory
nproc_per_node=2
PYTHONPATH=../../.. \
CUDA_VISIBLE_DEVICES=0,1 \
torchrun \
--nproc_per_node=$nproc_per_node \
--master_port 29500 \
llm_sft.py \
--model_id_or_path baichuan-inc/Baichuan-13B-Base \
--model_revision master \
--sft_type lora \
--tuner_backend peft \
--template_type AUTO \
--dtype AUTO \
--output_dir output \
--ddp_backend nccl \
--dataset blossom-math-zh \
--num_train_epochs 1 \
--max_length 2048 \
--check_dataset_strategy warning \
--quantization_bit 4 \
--bnb_4bit_comp_dtype AUTO \
--lora_rank 8 \
--lora_alpha 32 \
--lora_dropout_p 0.05 \
--lora_target_modules ALL \
--gradient_checkpointing true \
--batch_size 1 \
--weight_decay 0.1 \
--learning_rate 1e-4 \
--gradient_accumulation_steps $(expr 16 / $nproc_per_node) \
--max_grad_norm 0.5 \
--warmup_ratio 0.03 \
--eval_steps 100 \
--save_steps 100 \
--save_total_limit 2 \
--logging_steps 10 \
--deepspeed default-zero2 \