Skip to content

Commit

Permalink
mv modify 3
Browse files Browse the repository at this point in the history
  • Loading branch information
hywang2002 committed May 17, 2024
1 parent e87e6f0 commit 9ec3f5c
Show file tree
Hide file tree
Showing 3 changed files with 14 additions and 8 deletions.
12 changes: 10 additions & 2 deletions README.md
Original file line number Diff line number Diff line change
Expand Up @@ -49,7 +49,8 @@ conda activate mv-vton

3. Download the pretrained [vgg](https://drive.google.com/file/d/1rvow8jStPt8t2prDcSRlnf8yzXhrYeGo/view?usp=sharing)
checkpoint and put it in `models/vgg/` for Multi-View VTON and `Frontal-View VTON/models/vgg/` for Frontal-View VTON.
4. Download the pretrained [models](https://github.com/shadow2496/VITON-HD) and put `mvg.ckpt` in `checkpoint/` and put `vitonhd.ckpt`
4. Download the pretrained [models](https://github.com/shadow2496/VITON-HD) and put `mvg.ckpt` in `checkpoint/` and
put `vitonhd.ckpt`
in `Frontal-View VTON/checkpoint/`.

### Datasets
Expand Down Expand Up @@ -106,8 +107,15 @@ After these, the folder structure should look like this (the unpaired-cloth* onl

#### MVG

#### VITON-HD
To test on paired settings (using `cp_dataset_mv_paired.py`), you can modify the `configs/viton512.yaml` and `main.py`,
or directly rename `cp_dataset_mv_paired.py` to `cp_dataset.py` (recommended). Then run:
```shell
sh test.sh
```

To test on unpaired settings, rename `cp_dataset_mv_unpaired.py` to `cp_dataset.py`, and do the same operation.

#### VITON-HD

or just simply run:

Expand Down
6 changes: 3 additions & 3 deletions test.sh
Original file line number Diff line number Diff line change
@@ -1,9 +1,9 @@
CUDA_VISIBLE_DEVICES=3 python test.py --gpu_id 0 \
--ddim_steps 50 \
--outdir results/dci-vton-all/2k/ \
--outdir results/try/ \
--config configs/viton512.yaml \
--dataroot /mnt/pfs-mc0p4k/cvg/team/didonglin/why/datasets/mv_1000_split \
--ckpt /mnt/pfs-mc0p4k/cvg/team/didonglin/why/MV-VTON-ori/checkpoints/mvg.ckpt \
--dataroot /datasets/NVG \
--ckpt checkpoints/mvg.ckpt \
--n_samples 1 \
--seed 23 \
--scale 1 \
Expand Down
4 changes: 1 addition & 3 deletions train.sh
Original file line number Diff line number Diff line change
@@ -1,3 +1 @@
# CUDA_VISIBLE_DEVICES=4,5 python -u main.py --logdir models/dci-vton-mv-warp-paper --pretrained_model /mnt/pfs-mc0p4k/cvg/team/didonglin/why/DCI-VTON-Virtual-Try-On-ca/checkpoints/viton512.ckpt --base configs/viton512.yaml --scale_lr False

CUDA_VISIBLE_DEVICES=4,5 python -u main_finetune_part.py --logdir models/oc --pretrained_model /mnt/pfs-mc0p4k/cvg/team/didonglin/why/DCI-VTON-Virtual-Try-On-ca/checkpoints/viton512.ckpt --base configs/viton512.yaml --scale_lr False
CUDA_VISIBLE_DEVICES=4,5 python -u main.py --logdir models/oc --pretrained_model checkpoints/model.ckpt --base configs/viton512.yaml --scale_lr False

0 comments on commit 9ec3f5c

Please sign in to comment.