nicolabs/Open-Sora
1
0
Fork 0
mirror of https://github.com/hpcaitech/Open-Sora.git synced 2026-04-11 05:13:31 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions 1

add more instructions

Browse source
This commit is contained in:
Shen-Chenhui 2024-06-07 02:12:11 +00:00
parent 7e0f9b4b34
commit bad212c328
2 changed files with 12 additions and 12 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

10
docs/installation.md
View file

@ -82,7 +82,7 @@ export PATH="/path/to/vbench:$PATH"
You need to install VBench mannually by:
```bash
# first clone their repo
cd ..
cd .. # assume you are in the Open-Sora root folder, you may install at other location but make sure the soft link paths later are correct
git clone https://github.com/Vchitect/VBench.git
cd VBench
git checkout v0.1.2
@ -92,10 +92,10 @@ vim vbench2_beta_i2v/utils.py
# find `image_root` in the `load_i2v_dimension_info` function, change it to point to your appropriate image folder
# last, create softlinks
cd Open-Sora # or `cd Open-Sora-dev` for development
ln -s ../VBench/vbench vbench
ln -s ../VBench/vbench2_beta_i2v vbench2_beta_i2v
# later you need to make sure to call evaluatio from your Open-Sora folder, else vbench, vbench2_beta_i2v cannot be found
cd ../Open-Sora # or `cd ../Open-Sora-dev` for development
ln -s ../VBench/vbench vbench # you may need to change ../VBench/vbench to your corresponding path
ln -s ../VBench/vbench2_beta_i2v vbench2_beta_i2v # you may need to change ../VBench/vbench_beta_i2v to your corresponding path
# later you need to make sure to run evaluation from your Open-Sora folder, else vbench, vbench2_beta_i2v cannot be found
```

14
eval/README.md
View file

@ -46,9 +46,9 @@ python eval/loss/tabulate_rl_loss.py --log_dir path/to/log/dir
First, generate the relevant videos with the following commands:
```bash
# vbench tasks (4a 4b 4c ...)
bash eval/sample.sh /path/to/ckpt num_frames model_name_for_log -4a
# launch 8 jobs at once (you must read the script to understand the details)
# vbench task, if evaluation all set start_index to 0, end_index to 2000
bash eval/sample.sh /path/to/ckpt num_frames model_name_for_log -4 start_index end_index
# Alternatively, launch 8 jobs at once (you must read the script to understand the details)
bash eval/vbench/launch.sh /path/to/ckpt num_frames model_name
```
@ -70,13 +70,13 @@ python eval/vbench/tabulate_vbench_scores.py --score_dir path/to/score/dir
## VBench-i2v
[VBench-i2v](https://github.com/Vchitect/VBench/tree/master/vbench2_beta_i2v) is a benchmark for short image to video generation (beta version).
Similarly, install the VBench package following our [installation](../docs/installation.md)'s sections of "Evaluation Dependencies".
```bash
# Step 1: generate the relevant videos
# vbench i2v tasks (5a 5b 5c ...)
bash eval/sample.sh /path/to/ckpt num_frames model_name_for_log -5a
# launch 8 jobs at once
# vbench i2v tasks, if evaluation all set start_index to 0, end_index to 2000
bash eval/sample.sh /path/to/ckpt num_frames model_name_for_log -5 start_index end_index
# Alternatively, launch 8 jobs at once
bash eval/vbench_i2v/launch.sh /path/to/ckpt num_frames model_name
# Step 2: run vbench to evaluate the generated samples