issues Search Results · repo:FoundationVision/FlashVideo language:Python
Filter by
19 results
(86 ms)19 results
inFoundationVision/FlashVideo (press backspace or delete to remove)I am trying to learn the process of stage2 using flowmatching to learn the 4 step process from LQ to HQ. However, the
results after four steps are not satisfactory, the result looks like incomplete denoising. ...
reproduce
kxding
- 9
- Opened 8 days ago
- #21
I m having problem:
[2025-03-13 08:34:22,169] [INFO] [RANK 0] merge layer 35 lora attention back to linear INFO:sat:[RANK 0] merge layer 35
lora attention back to linear [2025-03-13 08:34:22,170] [INFO] ...
MehmetcanTozlu
- 5
- Opened 13 days ago
- #20
Hi, Thank you for your great work! I’m wondering about the GPU memory requirements for the second stage. I’m
encountering an OOM issue when training at 1080p resolution on an 80GB A100. Do you have any ...
LaVieEnRoseSMZ
- 3
- Opened 14 days ago
- #19
Hello, I would like to ask a few questions about stage2 2B model training:
1. What is the sampling interval during training? That is, what the num_noise_interval parameter setting in the code?
2. How ...
frozoul
- 6
- Opened 22 days ago
- #18
hello, I use the example code to run the example case. but I found the result is very bad, the second stage result is
strange, don t like the project you showed, I don t know where is wrong,I think the ...
help wanted
523997931
- 24
- Opened 27 days ago
- #17
Hi authors, in your paper, the loss target and noise_latent (model input) came to be target = z_HR - z_LR， and the
noise_latent = (1-t)z_LR + t * z_HR
While in the original flow matching paper (or normally ...
question
DDDaxing
- 3
- Opened on Feb 21
- #16
torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 1226.19 GiB. GPU has a total capacity of 63.98 GiB of
which 36.29 GiB is free. Of the allocated memory 24.03 GiB is allocated by PyTorch, ...
help wanted
sankexin
- 9
- Opened on Feb 17
- #15
qo4onFor GPUs with less memory, one might consider increasing the [spatial and temporal slice]
Can you please elaborate a bit on that matter ? For ex on a 4090, what settings should be tweaked ?
Thanks and ...
sirPhoebus
- 1
- Opened on Feb 16
- #13
Jandown
Learn how you can use GitHub Issues to plan and track your work.
Save views for sprints, backlogs, teams, or releases. Rank, sort, and filter issues to suit the occasion. The possibilities are endless.Learn more about GitHub IssuesProTip!
Restrict your search to the title by using the in:title qualifier.Learn how you can use GitHub Issues to plan and track your work.
Save views for sprints, backlogs, teams, or releases. Rank, sort, and filter issues to suit the occasion. The possibilities are endless.Learn more about GitHub IssuesProTip!
Press the /
key to activate the search input again and adjust your query.