比较是否使用 image loss对编码器效果的影响。
使用generator_encoder来进行测试
对比的encoder为：
1： /home/joselyn/workspace/0419-course/logs/StyleNeRF-main/train_encoder_1GPU/checkpoints/network-snapshot-000900.pkl
2： /home/joselyn/workspace/0419-course/logs/StyleNeRF-main/train_encoder/use_image_loss/checkpoints/network-snapshot-000900.pkl
命令1：python generate_encoder.py --outdir=/home/joselyn/workspace/0419-course/logs/StyleNeRF-main/generate_encoder/wo_image_loss --encoder=/home/joselyn/workspace/0419-course/logs/StyleNeRF-main/train_encoder_1GPU/checkpoints/network-snapshot-000900.pkl --trunc=0.7 --seeds=5 --network=/home/joselyn/workspace/0419-course/stylenerf_pkl/ffhq_256.pkl --render-program="rotation_camera3:/home/joselyn/workspace/0419-course/my_images/aframes_256"
命令2：python generate_encoder.py --outdir=/home/joselyn/workspace/0419-course/logs/StyleNeRF-main/generate_encoder/use_image_loss --encoder=/home/joselyn/workspace/0419-course/logs/StyleNeRF-main/train_encoder/use_image_loss/checkpoints/network-snapshot-000900.pkl --trunc=0.7 --seeds=5 --network=/home/joselyn/workspace/0419-course/stylenerf_pkl/ffhq_256.pkl --render-program="rotation_camera3:/home/joselyn/workspace/0419-course/my_images/aframes_256"
效果对比：左边是使用了 image_loss的效果，右边是没有使用image_loss的效果。 使用image loss的话， 对背景和头发要学得像一些。
