File size: 1,026 Bytes
dab24b4
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
41
42
43
44
45
46
47
### model
model_name_or_path: hfl/llama-3-chinese-8b-instruct-v3

### method
stage: sft
do_train: true
finetuning_type: lora
lora_target: all
# quantization_bit: 4                     # use 4-bit QLoRA
loraplus_lr_ratio: 16.0                 # use LoRA+ with lambda=16.0
# use_unsloth: true                       # use UnslothAI's LoRA optimization for 2x faster training
upcast_layernorm: true

### dataset
dataset: alpaca_mgtv_p1
template: llama3
cutoff_len: 4096
max_samples: 25000
overwrite_cache: true
preprocessing_num_workers: 16

### output
output_dir: saves/llama3-8b/lora/sft_bf16_p1_full_r3
logging_steps: 10
save_steps: 35
plot_loss: true
# overwrite_output_dir: true

### train
per_device_train_batch_size: 16
gradient_accumulation_steps: 8
learning_rate: 1.0e-4
num_train_epochs: 1.0
lr_scheduler_type: cosine
warmup_ratio: 0.1
bf16: true
ddp_timeout: 180000000

### eval
val_size: 0.1
per_device_eval_batch_size: 1
eval_strategy: steps
eval_steps: 35

report_to: wandb
run_name: llama3_8b_p1_full_r3 # optional