Upload VideoX-Fun/evaluation_vbench2.0_sd.sh
Browse files
VideoX-Fun/evaluation_vbench2.0_sd.sh
ADDED
|
@@ -0,0 +1,58 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
#!/bin/bash
|
| 2 |
+
|
| 3 |
+
echo "正在启动 8 个并行的独立任务，并为每个任务分配不同参数..."
|
| 4 |
+
# echo "等待 1 小时后启动任务..."
|
| 5 |
+
# sleep 3600 # 等待 3600 秒，即 1 小时
|
| 6 |
+
source activate wan
|
| 7 |
+
export LD_LIBRARY_PATH=/usr/local/cuda/lib64
|
| 8 |
+
|
| 9 |
+
# 1. 在这里预先定义你的参数数组 (a_i)
|
| 10 |
+
# 数组元素的数量应该与你的任务数量(8)相匹配。
|
| 11 |
+
# 参数可以是任何字符串，比如文件名、配置名、数值等。
|
| 12 |
+
save_folder='validation_samples/samples_vbench2_all_wan_aug'
|
| 13 |
+
sampler_name='Flow'
|
| 14 |
+
num_generated_videos=-1
|
| 15 |
+
prompt_list_path1='/nfs/ywang29/Reward_finetuning/VideoX-Fun/VBench/VBench-2.0/prompts/prompt_aug/Wanx_full_text_aug_part1.txt'
|
| 16 |
+
prompt_list_path2='/nfs/ywang29/Reward_finetuning/VideoX-Fun/VBench/VBench-2.0/prompts/prompt_aug/Wanx_full_text_aug_part2.txt'
|
| 17 |
+
|
| 18 |
+
save_folder_videogen='validation_samples/samples_videogen_eval/sd_pretrain'
|
| 19 |
+
prompt_list_path3='/nfs/ywang29/Reward_finetuning/VideoX-Fun/VideoGen-Eval.txt'
|
| 20 |
+
|
| 21 |
+
# enable_teacache=False
|
| 22 |
+
# num_inference_steps=50
|
| 23 |
+
|
| 24 |
+
|
| 25 |
+
save_folder_div='validation_samples/samples_vbench2_all_wan_aug/Diversity'
|
| 26 |
+
prompt_list_path_div='/nfs/ywang29/Reward_finetuning/VideoX-Fun/VBench/VBench-2.0/prompts/prompt_aug/wanx_aug_prompt/Diversity_dup.txt'
|
| 27 |
+
|
| 28 |
+
model_name='sd_7b_3'
|
| 29 |
+
steps=1000
|
| 30 |
+
|
| 31 |
+
|
| 32 |
+
params=(
|
| 33 |
+
"--seed 42 --save_folder ${save_folder_videogen} --prompt_list_path ${prompt_list_path3}"
|
| 34 |
+
)
|
| 35 |
+
# 2. 修改循环以遍历数组的索引
|
| 36 |
+
# ${!params[@]} 会获取数组 params 的所有索引 (0 1 2 3 4 5 6 7)
|
| 37 |
+
for i in "${!params[@]}"
|
| 38 |
+
do
|
| 39 |
+
# 从数组中获取当前索引对应的参数值
|
| 40 |
+
current_param="${params[$i]}"
|
| 41 |
+
|
| 42 |
+
# CUDA_VISIBLE_DEVICES=$i 告诉程序只能“看见”并使用第 i 张 GPU
|
| 43 |
+
# python X.py --p "$current_param" 将当前参数传递给脚本
|
| 44 |
+
# & 让命令在后台运行
|
| 45 |
+
echo "启动任务 $i (GPU $i)，参数为: $current_param"
|
| 46 |
+
CUDA_VISIBLE_DEVICES=$i python ./examples/wan2.1/predict_t2v_val_sd.py --num_generated_videos $num_generated_videos --sampler_name $sampler_name $current_param &
|
| 47 |
+
done
|
| 48 |
+
|
| 49 |
+
# 'wait' 命令会等待所有后台任务都执行完毕
|
| 50 |
+
echo "所有任务已启动。等待它们全部完成..."
|
| 51 |
+
wait
|
| 52 |
+
echo "所有任务已完成。"
|
| 53 |
+
|
| 54 |
+
python /nfs/ywang29/Reward_finetuning/VideoX-Fun/validation_samples/rename_v1.py --model_name ${model_name} --steps ${steps}
|
| 55 |
+
|
| 56 |
+
source activate vbench2
|
| 57 |
+
cd /nfs/ywang29/Reward_finetuning/VideoX-Fun/VBench/VBench-2.0
|
| 58 |
+
bash ./run_evaluate.sh "wan1.3b_rwft_${model_name}_${steps}_wan_aug1"
|