Home Creators Posts Import Register Favorites Logout

Videos

  • 542649573.mp4

Downloads

Content

Ovi SECourses Premium app generates 121-frame videos with real audio from text or images, like VEO 3 and SORA 2, with block swapping for low VRAM GPUs.

Supports all GPUs including RTX 5000 Series and runs on as low as 8 GB VRAM (6 GB with block swapping and tiled VAE), with Flash Attention, FP8 scaled models and batch processing. 1-click install on Windows, RunPod and Massed Compute, with presets for every GPU from 6 GB to 96 GB.

Patreon exclusive posts index to find our scripts easily, Patreon scripts updates history to see which updates arrived to which scripts and amazing Patreon special generative scripts list that you can use in any of your task.

Join discord to get help, chat, discuss and also tell me your discord username to get your special rank : SECourses Discord

Please also Star, Watch and Fork our Stable Diffusion & Generative AI  GitHub repository and join our Reddit subreddit and follow me on LinkedIn (my real profile)

=======

Latest zip file : Ovi_Pro_v9.zip

Full scale ultra advanced app for Ovi - an open source project that can generate videos from both text prompts and image + text prompts with real audio.

When Clear All Memory is selected (default in 32 GB and below presets) make sure to click Cancel button first and then close CMD or it will continue working as a subprocess

  • Project page is here : https://aaxwaz.github.io/Ovi/

  • I have developed an ultra advanced and easy to use Gradio app and much better pipeline that fully supports block swapping

    • Our block swapping is based on Kohya Musubi tuner implemention thus it is the best in the world right now

  • Our app also supports Block Based FP8 Scaling which is also based on Kohya Musubi tuner and it is also the best in the world right now from quality point

    • So we are not using base FP8 but using FP8_Scaled when enabled with higher quality

    • With intelliengt Block Based Scaling there isn't almost any quality loss

    • Our FP8_Scaled Base model reduces VRAM like 10 GB and its safentesors file will be auto downloaded as well

  • Now we can generate full quality videos with as low as 6 GB VRAM with Block Swapping + Tiled VAE

    • Our implemented tiled-VAE is same as how ComfyUI does so it is perfect quality and best out there

  • The 1-click installer will install into Python 3.10.11 venv and will auto download models as well so it is literally 1-click

    • My installer auto installs with Torch 2.8, CUDA 12.9, Flash Attention 2.8.3 and it supports literally all GPUs like RTX 3000 series, 4000 series, 5000 series, H100, B200, etc

  • All generations will be saved inside outputs folder and we support so many features like batch folder processing, number of generations, full preset save and load

    • All generations will have metadata txt files saved as well

  • Look the examples to understand how to prompt the model that is extremely important

    • You can use Google Studio AI and Gemini for free to write new amazing prompts hopefully I will show in upcoming tutorial

  • Look our below screenshots to see the app features

  • 50 Steps recommended but you can do low too like 20

  • 1-Click to install on Windows, RunPod and Massed Compute

  • Optimized presets for literally every GPU (starting from 6 GB to 96 GB)

9 November 2025 V8.4

  • New feature Optimized Block Swap added

  • It is supposed to speed up on Linux systems, sadly on Windows we are blocked by NVIDIA drivers

15 October 2025 V8.3

  • New checkbox Merge LoRAs on GPU added

  • This is for cloud services faster LoRA merge

  • This is auto enabled in 80 gb and 96 gb configs but should work with 48 gb GPUs as well

9 October 2025 V8.1

  • Lots of amazing new examples added

  • Added new example tabs T2V Video Extend Examples and I2V Video Extend Examples

    • These tabs will auto set Video Extend examples, and duration to 4 second for each clip - total 12 second

    • Now when you click load example, it will auto switch back to generate tab

  • New zip file has prompt_generate_guide_in_Gemini which you can use to generate prompts

  • Automatic set of Aspect Ratio with different base resolutions bug fixed

  • Manually set of Video Width and Video Height bugs fixed

  • Please run Windows_Install_and_Update.bat to very quickly update to latest version and see on your Gradio top left 8.1

Windows Requirements

Massed Compute (Recommend Cloud) :

RunPod (Cloud):

9 October 2025 V7.6

  • This is a massively bug fix and performance improvement update

  • This time we fixed for real finally :D

  • Now we have prompt caching feature - auto enabled

    • It will generate hash value of your Video Prompt + Video Negative Prompt + Audio Negative Prompt + FP8_Scaled enabled or not

    • With this hash value, it will check if T5 encoding exists in prompt_cache or not

    • If exists, it will skip T5 encoding, will speed up immersely, if not it will cache and save

    • This should work in all cases we have like single processing, batch processing, multi line, video extend, etc

  • I have entirely re-factored T5 system for above change and hopefully now we have finally fixed T5 infinite loop error, so update and try again

    • The loop error was caused auto enabling Delete T5 After Encoding and
      Clear All Memory at the same time and now issue fixed

  • I have fixed the RAM leak that was happening when Clear All Memory was not enabled

    • Therefore now 96 GB RAM PCs can disable Clear All Memory, It may work with 64 GB RAM too so test

    • Still if you get error enable Clear All Memory

  • When Clear All Memory was not enabled, changing video duration after first run was not working since it was initializing and never changing again

    • This bug fixed and changing video duration should work in all cases now

  • Auto pad for 32px divisibility checkbox added

    • When enable, it won't crop any part of image, it will auto downscale to target resolution and it will fill missing parts with black pixels to make it divisible to 32

    • Use with Auto Crop Image - don't disable it

  • Do not auto enforce validation check was not working in all cases but this issue is fixed and should work now

  • Prompt validation system improved

    • Now it will show you errors of single generation on Gradio

    • So you will see your error

    • When doing batch processing, it wont start the batch and show errors on cmd

      • You can always enable Do not auto enforce validation check and skip auto enforce

  • Sorry for the errors

8 October 2025 V6.4

  • LoRAs were not working accurately with FP8_Scaled base model and this issue fixed

  • However, for LoRAs to work, we have to each time re-scale on the fly, so cached scale won't be used

    • This adds like 10 seconds delay

    • This LoRA having version F5_Scaled won't be saved, we can add save feature if you wish but that means 11.5 GB model file for every LoRA combination

    • I will look further if there is any way to apply LoRA to F5_Scaled base model cache

  • In some cases, users reported that it was infinite looping T5 encoding, this issue hopefully fixed

  • Hopefully I will add T5 text encoder caching to cache directory, so same prompt will use cache directly not re-cache

  • Hopefully I will add load background sound, so you can upload background music and auto added

  • Just run Windows_Install_and_Update.bat to very quickly update to latest version and see on your Gradio top left 6.4

6 October 2025 V6.3

  • Exact resolution issue fixed and the model will use exactly the resolution you give on interface Video Width and Video Height

  • When doing batch processing, based on your base Width and Height, it will auto crop and resize your input folder images accurately to exact resolutions

    • e.g. when base resolution is 960x960, 1152x1728px image will be generated as 768x1184

  • Be careful bigger resolution uses more VRAM

  • When doing video extension, it was not keeping Sage Attention selection, now will respect your selection

  • First example prompt issue fixed

  • Do not auto enforce validation check added

    • So you can generate videos without any speech tags

  • Now we are supporting up to 4 LoRAs

  • Just run Windows_Install_and_Update.bat to very quickly update to latest version and see on your Gradio top left 6.3

6 October 2025 V5.9

  • This is a super important major update that almost completes our app into maximum quality

  • Now you can use both videos and images as input

    • When you upload a video as input, it will get the last frame, auto crop it if enabled, use it as a reference image, then it will generate your video and merge back with your input video, so you have basically video extension of existing videos feature right now

    • If you don't want auto combine, enable Don't auto combine video input checkbox and it won't auto combine just use last frame of video as an input

  • Now we have Multi-line Prompts feature

    • When this is enabled, the prompt box input will be seperated into lines and every new line prompt will become an individual prompt and it will generate a video for each prompt

    • Lines lesser than 3 characters will be ignored so you can put 2 new line spaces and write if you wish

    • This will work with batch processing as well, just have your prompts multiple lines in your batch processing folder

    • Don't enable Multi-line Prompts and Video Extension at the same time

  • Now we have Video Extension (Last Frame Based) feature

    • When this is enabled, it will extend your video based on the number of lines you have

    • Lets say you have a prompt that is 3 lines

    • So first line will be base prompt and will generate 0001.mp4

    • The second line will be second prompt, it will get last frame of 0001.mp4 and use it as an input image, use second line prompt and will generate 0001_ext1.mp4

    • The third line will be third prompt, it will get last frame of 0001_ext1.mp4 and use it as an input image, use third line prompt and will generate 0001_ext2.mp4

    • After all generations done it will merge all generations and generate 0001_final.mp4

    • You can extend as many as times you want with number of lines, fully automatic and working pretty good

    • I will hopefully add example for this into examples tab soon

    • Don't enable Multi-line Prompts and Video Extension at the same time

  • Now we have presets for all lower VRAM GPUs for Scaled FP8 Base Model

    • Scaled FP8 Base Model is working perfect and 24 GB GPUs can generate 5 second videos without any Block Swap, thus ultra fast

  • I have implemented Sage Attention and working perfect

    • It did speed up 15% during inference and now it is auto enabled in all presets

  • Auto cropping logic improved, some bugs fixed and made more robust

  • Automatic prompt format validation system implemented

    • When you click Generate button it will check and throw error if not valid

    • There is also Validate Prompt Format button now for you to validate and see errors

  • Entire Gradio app font changed to Tahoma for better readibility

  • New prompting feature setting duration in prompts

    • If you write in beginning of your prompt like {2} it will make that video generation as 2 seconds

    • This feature is useful for multi line generation and video extension features

    • So the format is {x}, if there is no such format, it will use duration slider set value

    • Example prompt

      • {4} A man is doing a podcast video. He is saying <S> Hi guys! How are you! Did you know, I am not real? <E> He continues to talk.

      • {2} A pod cast making man talking. He is saying <S> Like for real, I just found out. I was made by Furkan's Ovi app! <E> He then giggles <S> Hi-hi-hi! <E>

  • You can also write speaking prompts like this

    • <S>[strong sound] I am an artificial intelligence android robot.<E>

    • <S>[soft whisper] I am an artificial intelligence android robot.<E>

  • Don't modify our presets and save as your modified presets since when you update, they will be overwritten back to originals

  • Don't close running cmd window immediately, first click cancel and then close cmd

5 October 2025 V4.4

  • This is a super important update from performance wise

  • Now Delete T5 After Encoding will be auto enabled if your RAM is under 64 GB but if you don't have 96 GB RAM i really recommend to enable

  • Now when Delete T5 After Encoding is enabled, it will start it as a sub-process therefore there will be 0 RAM and VRAM leakage

  • Now CPU-Only T5 will load like 2x faster

  • New option Scaled FP8 Base Model added

    • This saves like 10 GB VRAM

    • If you had generated with V4.0 delete it from Ovi_Pro\ckpts\Ovi folder and regenerate

  • With FP8 Base Model, 24 GB GPUs can generate without any Block Swap

    • It uses like 17 GB VRAM at the moment during inference without Block Swap

  • T5 loading speed increased for all BF16, FP8_Scaled and CPU

  • FP8 Scaled T5 and Base model will be auto downloaded now

  • Working on fixing longer generation - reported to be broken

  • Just run Windows_Install_and_Update.bat to very quickly update to latest version

5 October 2025 V3.8

  • This is a big update and I am still testing so many new amazing features

    • Still in testing so report errors

  • Fully automatic aspect ratio resolutions and detection based on your entered Base Width and Base Height

  • Like set 960x960 and you will get 1280x704 automatically for 16:9

  • Moreover now you can generate bigger resolution videos like 1280x704 but remember it would use more VRAM than 960x544 - base resolution which is 720x720

  • Longer generations are also available so give them a try

  • Now it will show uploaded image resolution

  • Now it will auto crop to the new desired resolution you give immediately and show cropped image resolution too

  • Batch processing fixed and auto cropping perfected

    • It will auto recognize your images in your folder and process them with their closest aspect ratio based on your base resolution

  • Just run Windows_Install_and_Update.bat to very quickly update to latest version

5 October 2025 V3.4

  • Ok this is a massive update

  • We have added full tiled-VAE same as ComfyUI and working amazing

  • Now with Block Swapping + tiled-VAE + T5 Text Encoding on CPU (still super fast) we can generate 121 frames 5 second videos as low as on 6 GB GPUs

  • I have added presets for every GPU out there and the app will automatically detect your GPU and select your preset when you first time install and start

  • Cancel button was not working properly and now working perfect

  • 3:4 and 4:3 aspect ratios added as well

  • Original repo was forcing all resolution to be 720x720, I have added a new feature called as Force Exact Resolution and with this you can generate with higher resolution like 1280x704

    • It must be divisible to 32

    • Auto crop will auto handle this

    • All presets have this feature enabled by default but remember VRAM presets are made for 720x720 base resolution, higher resolution uses more VRAM

  • Much more robust preset system developed to not have any errors when saving or loading older presets

  • If you are low on both VRAM and RAM try this Delete T5 After Encoding + Scaled FP8 T5 + CPU-Only T5

  • Inaccurately showing previous generation result on interface fixed

  • Just run Windows_Install_and_Update.bat to very quickly update to latest version

4 October 2025 V2.9

  • Delete Text Encoder After Encoding : Now you can enable or disable

  • Now will load Text Encoder directly into VRAM, encode and then delete or move according to your selection - Therefore it is was faster than before

  • Clear All Memory added and recommended - 0 VRAM and RAM leak

  • Scaled FP8 T5 - reduces T5 VRAM usage but slower tto load - quality same

    • It will be auto enabled when your VRAM is below 23 GB and you don't load a preset

  • Delete T5 After Encoding - enable if low on RAM

  • Preset save and load fully working now

  • Consequent generation error fixed with Clear All Memory

  • With CPU Offload now we will move VAE to RAM while not needed and thus with 29 Block Swap now it uses only 6 GB VRAM during inference

  • Upcoming tiled VAE hopefully and FP8_Scaled model loading and video extending - loop

  • With v2.9 it will save FP8_Scaled version of T5 and use it when you next time used Scaled FP8 T5 to speed up

    • The file will be saved inside ckpts\Wan2.2-TI2V-5B

  • Just run Windows_Install_and_Update.bat to very quickly update to latest version

Full tutorial video coming soon hopefully

Related posts

All posts in this topic: AI Video Generation and Animation Apps · New to SECourses? Start Here: Best and Most Popular SECourses Posts

Files

Comments

DanO..

Just gave it a try. Ran out of memory with RTX 3060 12GB VRAM. I lowered resolution and maxed out swapped blocks but wouldn't generate. It appears as though it leaves system memory (40GB) and VRAM maxed out after failed attempt. Needs to clean it out after failure or add some mechanism to clear out memory. Restarting to try with lower resolution and maxed block swapping.

DanO..

Looks like it's not going to work for me. Started fresh. 512X512. Max block swapping. Ran out of memory.

Hipno

You are amazing Dr! thank you once again!!

SECourses

thanks a lot. working to add more features today hopefully

SECourses

yes you need more RAM. today I will work on FP8_Scaled If i can implement it may fix your issue. but you need more RAM

b

rtx 4060 with 8gb gpu and 64gb ram memory not working, seems like memory issue

SECourses

8 GB memory is too low. so far i got as low as 8.2 GB. I am trying to add FP8_Scaled right now lets see if that can fix your issue

JP LONB

Thanks for your amazing work, dear Furkan. Unfortunately, it did not work for me, I will wait for the next version. Here is the error I got, in case you're interested: "INFERENCE STARTING - VRAM: 16.58 GB Block swap active: 12/30 blocks on CPU ================================================================================ 2it [08:37, 258.78s/it] ERROR:root:Traceback (most recent call last): File "E:\AI\Ovi_Pro\Ovi_Pro\ovi\ovi_fusion_engine.py", line 455, in generate pred_vid_pos, pred_audio_pos = self.model( File "E:\AI\Ovi_Pro\Ovi_Pro\venv\lib\site-packages\torch\nn\modules\module.py", line 1773, in _wrapped_call_impl return self._call_impl(*args, **kwargs) File "E:\AI\Ovi_Pro\Ovi_Pro\venv\lib\site-packages\torch\nn\modules\module.py", line 1784, in _call_impl return forward_call(*args, **kwargs) File "E:\AI\Ovi_Pro\Ovi_Pro\ovi\modules\fusion.py", line 310, in forward vid, audio = gradient_checkpointing( File "E:\AI\Ovi_Pro\Ovi_Pro\ovi\modules\model.py", line 22, in gradient_checkpointing return module(*args, **kwargs) File "E:\AI\Ovi_Pro\Ovi_Pro\ovi\modules\fusion.py", line 223, in single_fusion_block_forward assert not torch.equal(og_audio, audio), "Audio should be changed after cross-attention!" torch.AcceleratorError: CUDA error: an illegal memory access was encountered CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect. For debugging consider passing CUDA_LAUNCH_BLOCKING=1 Compile with `TORCH_USE_CUDA_DSA` to enable device-side assertions. Error during video generation: cannot unpack non-iterable NoneType object"

macmotu

Hi Furkan, thanks a lot for your great work on Ovi Pro Fusion! I found a bug on my RTX 4090: With block swap = 12 and CPU offload = true, the first generation works perfectly. But the second generation with the same settings always fails. Error message: RuntimeError: Input type (CUDABFloat16Type) and weight type (CPUBFloat16Type) should be the same It looks like the patch_embedding layer stays on CPU after the first run, while the inputs are already on CUDA. Without block swap/offload it doesn’t run well on 24 GB, so this is important. Restarting the app always fixes it for one run. Maybe the block swap step needs to re-sync that layer on every run. Thanks again for this amazing project!

SECourses

fixed with v2.5 just run installer again and make sure Clear All Memory enabled

SECourses

fixed with v2.5 just run installer again and make sure Clear All Memory enabled

SECourses

With CPU Offload now we will move VAE to RAM while not needed and thus with 29 Block Swap now it uses only 6 GB VRAM during inference update latest version and try read changelogs

SECourses

With CPU Offload now we will move VAE to RAM while not needed and thus with 29 Block Swap now it uses only 6 GB VRAM during inference update latest version and try read changelogs

DanO..

I'm checking it now, buddy! By the way, 3:4 and 4:3 are handy aspect ratios to include. The two you have along with square are very narrow.

REGINALDO BARBOSA

[GENERATE_VIDEO] Called with clear_all=False, num_generations=1 ================================================================================ INITIALIZING OVI FUSION ENGINE IN MAIN PROCESS Block Swap: 8 blocks (0 = disabled) CPU Offload: True Image Generation: False No Block Prep: False Note: Models will be loaded in main process (Clear All Memory disabled) ================================================================================ [OK] OviFusionEngine initialized successfully (models will load on first generation) [GENERATION 1/1] Starting with seed: 99 ================================================================================ STEP 1/2: Loading T5 text encoder FIRST to minimize RAM usage ================================================================================ ================================================================================ Loading OVI models for first generation... Block Swap: 8 blocks CPU Offload: True ================================================================================ Initial VRAM: 0.00 GB Removing weight norm... Loading T5 text encoder directly to GPU (BEFORE fusion model to save RAM)... Error during video generation: CUDA out of memory. Tried to allocate 160.00 MiB. GPU 0 has a total capacity of 12.00 GiB of which 0 bytes is free. Of the allocated memory 18.23 GiB is allocated by PyTorch, and 361.40 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True to avoid fragmentation. See documentation for Memory Management (https://pytorch.org/docs/stable/notes/cuda.html#environment-variables) [SINGLE-GEN] Failed - no result returned [SUBPROCESS] Generation failed with return code: 1 [GENERATION 1/1] Failed in subprocess

DanO..

Didn't work for me. All VRAM savings activated. 480X640. RTX 3060 12GB VRAM. 40GB RAM. 250GB Virtual Memory on SSD. (I bought an SSD just for virtual memory and I have two 3060s installed so one is for AI stuff exclusively.) We'll see if others get it to work.

SECourses

Update to latest version and try block swap 12

SECourses

Update to latest version and try T5 scaled and block swap 12

SECourses

At which stage you get error can you send me entire cmd as email monstermmorpg@gmail.com I think for you we need tiled vae + fp8 model which I will work hopefully today and let you know

DanO..

Already moved on to something else! I never reached any sort of progress bar if that helps. VAE encoding is after that normally? I can normally get most stuff working. We'll get there!

macmotu

Thanks Furkan, fix works perfectly now on my 4090! Second runs are stable, no more device mismatch. Really appreciate the quick update 🙏

Damjan Žakelj

Amazing work, really. Thank you so much :)

SECourses

you are welcome thanks for comment

SECourses

VAE decoding is at the end. lets see wait me to add FP8 Scaled and VAE decoding hopefully

ranjeet

What are the generation time? For example on rtx 3080ti 16gb VRAM, 32 gb RAM?

SECourses

5090 really fast but i dont know 3080 ti. give it a try. also making improvements. you can also generate in 20 steps to speed up

Realtoonai

Will this work with 4090?

Lou

4090 Just under 10min at default settings using a sample provided in app. 30 steps shaves about 2-3min from 50 steps. Total Vram usage is around 18g with 2g allocated to "other" background programs. I know it's not a 3080ti but I though I would share.

SECourses

yes block swap really slows down. with 12 block swap it is 6.75 second it and without it, it is like 3.5 second it on rtx 5090.

SECourses

yes it works perfect. you can use default config

mastermk

Hi Furkan. Thanks for this fantastic work. I tried generating videos in Italian, but the audio is horrible. Is it possible to improve this language? Can videos be generated that are at least 10 seconds long?

SECourses

I don't know both of them. I am working on improving the VAE for lower VRAM gpus but i will check them later.

SECourses

I don't know sadly. I dont have AMD card to test.

SECourses

with v3 update i added presets for every GPU

SECourses

with v3 update i added presets for every GPU - works great on 16 GB GPUs now

SECourses

with v3 update i added presets for every GPU - works great on 12 GB GPUs now

SECourses

with v3 update i added presets for every GPU - works great on 24 GB GPUs now

SECourses

with v3 update i added presets for every GPU - works great on 8 GB GPUs now

SECourses

with v3 update i added presets for every GPU - works great on 12 GB GPUs now

James Charleston II

It won’t run in text to video mode for me. I have to start with an image or nothing happens.

SECourses

run installer bat and see v3.1. and test one of the example text to video. i just tested and it works. please also copy paste your cmd here and also preset name. but first update

DanO..

Stuck at "encoding text prompts" each time. This time I let it go for an hour and a half. Just sits there. I turned off CPU Only T5 and it ran out of memory pretty quickly. I selected 8GB set up and turned on other memory saving settings. No help. Stuck each time at text prompts. If it's working for everyone else it's likely a problem unique to me and I'll check out later versions. Thanks for trying to get it to work! In case it's helpful: [GENERATION 1/1] Starting with seed: 99 ================================================================================ STEP 1/2: Loading T5 text encoder FIRST to minimize RAM usage ================================================================================ ================================================================================ Loading OVI models for first generation... Block Swap: 29 blocks CPU Offload: True ================================================================================ Initial VRAM: 0.00 GB Removing weight norm... ================================================================================ T5 CPU-ONLY MODE: Loading T5 on CPU for CPU inference This saves VRAM but text encoding will be slower ================================================================================ [T5 CPU-ONLY MODE] Loading T5 text encoder on CPU for CPU inference [T5 CPU-ONLY MODE] This saves VRAM but encoding will be slower [T5 LOAD][FP8] Encoder structure created in 78.81s [FP8 CACHE] Cached FP8 checkpoint not found at D:\Auto1111\OVI\Ovi_Pro_v1\Ovi_Pro\ckpts\Wan2.2-TI2V-5B\models_t5_umt5-xxl-enc-fp8_scaled.safetensors [T5 LOAD][FP8] BF16 weights loaded from disk in 75.28s Quantizing to FP8: 100%|█████████████████████████████████████████████████████████████| 168/168 [00:44<00:00, 3.77it/s] [T5 LOAD][FP8] FP8 quantization completed in 44.69s [T5 LOAD][FP8] FP8 monkey patching completed in 0.00s [T5 LOAD][FP8] FP8 layer prep completed in 0.00s [FP8 CACHE] Saved FP8 T5 checkpoint to D:\Auto1111\OVI\Ovi_Pro_v1\Ovi_Pro\ckpts\Wan2.2-TI2V-5B\models_t5_umt5-xxl-enc-fp8_scaled.safetensors [T5 LOAD][FP8] Cached FP8 checkpoint written in 48.54s [T5 LOAD][FP8] Total FP8 load time 247.42s [SCALED FP8] T5 text encoder loaded in Scaled FP8 format [SCALED FP8] Expected VRAM savings: ~50% (~2.5GB saved) [T5 CPU-ONLY MODE] T5 encoder ready on CPU ================================================================================ T5 loaded. Fusion model will load AFTER text encoding. ================================================================================ STEP 2/2: Encoding text and optionally deleting T5 before loading fusion model ================================================================================ Encoding text prompts...

SECourses

I will try fp8 + cpu combination. Looks like your ram is being big limitation. What do you see on ram usage when it is set to cpu? And cpu usage

Agino Terra

im getting this error with the 3.1 - Initial VRAM: 0.02 GB Step 1/6: Creating model structure on meta device... Score model (Fusion) all parameters:11660753108 Step 2/6: Loading checkpoint weights to CPU... [SUBPROCESS] Generation failed with return code: 3221225477 [GENERATION 1/1] No output file found

DanO..

All 40GB of RAM in use. CPU not really tasked. 25% at the most. Virtual RAM doesn't look like it's being touched. My SSD E drive is normally maxed out if Virtual RAM is getting used a lot. But E doesn't look like it's being touched. VRAM not touched at all when stuck at prompt text.

Richard Nagy

Thanks for that! Few thing to mention: (I used the update batch file to update to 3.0. But I'll try to delete venv and hit it again just to make sure the update really happened) - Cancelling doesn't work. - Saving new preset also not working Traceback (most recent call last): File "P:\AI\Ovi_Pro_v1\Ovi_Pro\venv\lib\site-packages\gradio\queueing.py", line 759, in process_events response = await route_utils.call_process_api( File "P:\AI\Ovi_Pro_v1\Ovi_Pro\venv\lib\site-packages\gradio\route_utils.py", line 354, in call_process_api output = await app.get_blocks().process_api( File "P:\AI\Ovi_Pro_v1\Ovi_Pro\venv\lib\site-packages\gradio\blocks.py", line 2112, in process_api inputs = await self.preprocess_data( File "P:\AI\Ovi_Pro_v1\Ovi_Pro\venv\lib\site-packages\gradio\blocks.py", line 1774, in preprocess_data processed_input.append(block.preprocess(inputs_cached)) File "P:\AI\Ovi_Pro_v1\Ovi_Pro\venv\lib\site-packages\gradio\components\dropdown.py", line 206, in preprocess raise Error( gradio.exceptions.Error: "Value: True is not in the list of choices: ['BlahBlah']" Traceback (most recent call last): File "P:\AI\Ovi_Pro_v1\Ovi_Pro\venv\lib\site-packages\gradio\queueing.py", line 759, in process_events response = await route_utils.call_process_api( File "P:\AI\Ovi_Pro_v1\Ovi_Pro\venv\lib\site-packages\gradio\route_utils.py", line 354, in call_process_api output = await app.get_blocks().process_api( File "P:\AI\Ovi_Pro_v1\Ovi_Pro\venv\lib\site-packages\gradio\blocks.py", line 2112, in process_api inputs = await self.preprocess_data( File "P:\AI\Ovi_Pro_v1\Ovi_Pro\venv\lib\site-packages\gradio\blocks.py", line 1774, in preprocess_data processed_input.append(block.preprocess(inputs_cached)) File "P:\AI\Ovi_Pro_v1\Ovi_Pro\venv\lib\site-packages\gradio\components\dropdown.py", line 206, in preprocess raise Error( gradio.exceptions.Error: "Value: True is not in the list of choices: ['BlahBlah']"

b

updated to v3, rtx 4060 8gb gpu and 64gb ram, got this error [OK] OviFusionEngine initialized successfully (models will load on first generation) [GENERATION 1/1] Starting with seed: 99 ================================================================================ STEP 1/2: Loading T5 text encoder FIRST to minimize RAM usage ================================================================================ ================================================================================ Loading OVI models for first generation... Block Swap: 29 blocks CPU Offload: True ================================================================================ Initial VRAM: 0.00 GB Removing weight norm... ================================================================================ T5 CPU-ONLY MODE: Loading T5 on CPU for CPU inference This saves VRAM but text encoding will be slower ================================================================================ [T5 CPU-ONLY MODE] Loading T5 text encoder on CPU for CPU inference [T5 CPU-ONLY MODE] This saves VRAM but encoding will be slower [T5 LOAD][BF16] Encoder structure created in 33.61s [T5 LOAD][BF16] Weights loaded in 6.59s (total 40.20s) [T5 CPU-ONLY MODE] T5 encoder ready on CPU ================================================================================ T5 loaded. Fusion model will load AFTER text encoding. ================================================================================ STEP 2/2: Encoding text and optionally deleting T5 before loading fusion model ================================================================================ Encoding text prompts... Text embeddings encoded on CPU and moved to GPU Keeping T5 on CPU (already in CPU-only mode) ================================================================================ STEP 3/3: Loading fusion model (T5 already deleted if enabled) ================================================================================ ================================================================================ Loading OVI models for first generation... Block Swap: 29 blocks CPU Offload: True ================================================================================ Initial VRAM: 0.00 GB Step 1/6: Creating model structure on meta device... Score model (Fusion) all parameters:11660753108 Step 2/6: Loading checkpoint weights to CPU... Error during video generation: The paging file is too small for this operation to complete. (os error 1455) [SINGLE-GEN] Failed - no result returned [SUBPROCESS] Generation failed with return code: 1 [GENERATION 1/1] No output file found

SECourses

how much RAM VRAM you have and what preset?

SECourses

i just made another update for you. please update and see v3.2 and try this : CPU-Only T5 + Scaled FP8 T5 + Delete T5 After Encoding and let me know how far it goes

SECourses

Just updated app to v3.3 for you and made preset more robust. please email me logs of Windows_Install_and_Update.bat file : monstermmorpg@gmail.com and try again

Fco Muñoz

I mostly of times has a static video with sound, which parameter i have to change to avoid it?

SECourses

i think you need to change prompt. did you test example prompts in example tab? they are perfectly animated with default presets.

Richard Nagy

Ah, you are too kind! :) I’ll check ot later today. Many Thanks!

SECourses

you are welcome. i am also adding more updates about to push a new one :D

Agino Terra

I have 24 GB of VRAM and 48 GB of RAM. The funny thing I asked Crok how to fix the error, and his says to make a virtual disk. I did, and it worked one time, then stopped again. The old version works it's just 3.1 that has this problem

SECourses

no newer version should be even better. do you have 100 gb virtual RAM? also try Delete T5 After Encoding + Scaled FP8 T5 + CPU-Only T5 and make sure to update and have 3.8 on the gradio. i just made more fixes

FalconBravery

Thank you for this awesome release. I'm currently testing on H100 and 60 steps (3 min generation/video). Ive maybe found a bug, no matter which video length I set, the videos are always 5 seconds long, even if I set 10 seconds or any other number

Alexandre Rangel

I ran the installer ok on Windows 64 but when trying to run I get the error: The system cannot find the path specified. Traceback (most recent call last): File "C:\Ovi_Pro_v3\Ovi_Pro\premium.py", line 1, in import gradio as gr ModuleNotFoundError: No module named 'gradio'

FalconBravery

i had the same problem on runpod. Yesterday I installed the 2.x version, and it worked nice, today I updated it to 3.x with git pull but gradio missing, after 1 hours of wasting time trying to fix it I just deleted the entire folder and reinstalled it

SECourses

I am pretty confident V4 will work update and try :) made huge RAM improvement

SECourses

I am pretty confident V4 will work update and try :) made huge RAM improvement

SECourses

I am pretty confident V4 will work update and try :) made huge RAM improvement

SECourses

I am pretty confident V4 will work update and try :) made huge RAM improvement

SECourses

I am pretty confident V4 will work update and try :) made huge RAM improvement

DanO..

Sorry for not getting back to you. I'm on the other side of the planet! XD I did try the previous version and let it run while I did other things and it was still stuck at Encoding Text Prompts for almost two hours. I kept an eye on the CPU this time and it sat around 40-50% the whole time. System memory was maxed for quite a while but slowly diminished to 50%. My giant 250GB virtual memory was never touched. (I have it on a separate SSD drive.) And my 12GB VRAM was never touched during that phase. I'll try the new version! (I'm not sure memory is the problem.)

Agino Terra

Yeah, it has the same problem . This is how my virtual memory looks - 16384 Initial size 32768 memory size, i did try Delete T5 After Encoding + Scaled FP8 T5 + CPU-Only T5

SECourses

upgrade to 4.4 and try this : Delete T5 After Encoding + CPU-Only T5 dont enable Scaled FP8 T5 and let me know how far it goes

SECourses

can you try this? upgrade to 4.4 and try this : Delete T5 After Encoding + CPU-Only T5 dont enable Scaled FP8 T5 and let me know how far it goes

SECourses

yes on runpod it may get broken easily. best to do reinstall. and always use torch 2.2.0 template.

SECourses

just tested with v4.4 and worked perfect. it took 515 seconds on rtx 5090 . shared result in discord

DanO..

OK. The last one I tried, same thing. Virtual memory hit a bit prior to text encoding. None of it used during. Hung up for over two hours. I'll try the newest version. (I'm just updating by the way.)

SECourses

ye let me know i am hopeful. it is v 4.5

Accessible Game Mods by James

Hello, I got everything installed and running, but running this even with the 1tgb preset it slows down my computer a lot and a few times it slowed it down so much I had to hold in the button to turn it off. I'm not sure what to really do about it.

SECourses

what is 1tgb? are you on v 4.5? how much RAM you have and what GPU

Accessible Game Mods by James

Meant 12gb. don't know how I messed that up lol. I have 32gb ram, 12gb vram 4070 super and on latest version I think as I just ran the installer about an hour ago.

SECourses

ye did a lot of updates. run installer again and try again please. and tell me where it is stuck or logs email me : monstermmorpg@gmail.com

DanO..

No good. Locked up again for 40 minutes. I think I'll need to wait for it to become more optimized. From what I understand people are saying a 5090 is the minimum to get it to work. I appreciate your attempts to make it work for most people but that may not be possible at this time. Unless other people are reporting it's working. Thanks for the effort!

SECourses

people knows nothing. it is working as low as 6 GB GPUs at the moment i did huge optimization. your RAM is the biggest issue. are you on v 4.5? can you send me entire cmd logs : monstermmorpg@gmail.com and screenshot of how you tested?

DanO..

No, thank you. I'm out. I'm enjoying WAN 2.2 Animate with a context window that lets me replace characters in video 1024X720 for up to 20 seconds right now. Could be as long as I want if I'm willing to wait. If people have more than 40GB of RAM and only a 6GB video card I'd think that's extremely rare. I'm not joking about having 250GB of virtual RAM but as I said, it doesn't touch that. Thanks again for your herculean effort! I'll try again later. Keep up the great work.

James Charleston II

I’ve updated to version 3 and v2 worked better. It did find my card but not building. It gets stuck.

SECourses

try 4.5 and email me cmd logs : monstermmorpg@gmail.com

James Charleston II

Hello, I have a busy start to the week, but I will do my best to address this promptly. I will note that on v3.8, I put in the sample Priest image and prompt, I can see the priest prompt and it shows the video of the two back women in the kitchen. Ah, this is very unexpected behavior. LOL

SECourses

must be related to training dataset :D

Daniel Smith

Hi, great work. Everything worked and I managed to create a video just fine. I have two GPU's a 4060 8gb and 5060ti 16gb, yet It ignores the 16gb even though I change the REM SET CUDA_VISIBLE_DEVICES=0 as it was =1. (or set CUDA_VISIBLE_DEVICES=0) The 8gb is plugged to monitor so I use the 5060 in OVI/ComfyUI etc for rendering. What can I do to fix this?

SECourses

i have dual GPUs too. REM means it is commented. so make a copy of bat file and add this line before call line as SET CUDA_VISIBLE_DEVICES=1 - message me from discord i will send you bat file

leem0nchu

New version is lighting up my antivirus "Threat found - action needed. 10/6/2025 9:38 PM Severe Detected: Trojan:Script/Wacatac.C!ml Status: Active Active threats have not been remediated and are running on your device. Date: 10/6/2025 9:38 PM Details: This program is dangerous and executes commands from an attacker. Affected items: file: C:\Users\ Downloads\Ovi_Pro_v5.zip Learn more"

The Catz Home

In the "how to use" windows, you said : Check <s>...</s> tag format Add ... descriptions ect… but in the examples and on the page of the model they use the tag formats <s> (for speech) (for audio description). which ones are the good ones ?</s>

The Catz Home

The tags are not allow on the comment lol so I will use * instead of the tag symbols :p In the "how to use" windows, you said : Check *S*...*/S* tag format Add *AUDCAP*...*/ENDAUDCAP* descriptions ect… but in the examples and on the page of the model they use the tag formats *S* *E* (for speech) *AUDCAP* *ENDAUDCAP* (for audio description). which ones are the good ones ?

The Catz Home

I tested with "your" tags S /S and I have an error message about "Unknown tags found:" . So I think there is an error in the "how to use" windows ^^

Neil Rhodes

In the samples on the project page https://aaxwaz.github.io/Ovi/ the voices vary, mine sound all the same. Is there a way to vary the voice when using an image as a starting point for the video? A prompt trigger word or something

SECourses

you can change seed. currently it is 99 unless you enable randomize seed

SECourses

just click load example and generate you will see how it works.

Oliver

Greetings. I did a clean install from v3 (working) to v5. Now my generations are consistently stuck in a loop back into the T5 encoding process. Only way to stop the process is closing Python. Thanks in advance

SECourses

clear your output folder. i need to make a fix for this. after doing that restart and let me know. also try v 6.2

Diggy Dre

I wasn't able to get this to work on my 4080. Also, for some reason, it kept trying to offload the text encoder to my 9950x.

SECourses

Update latest version, delete outputs and send me entire cmd log as email : monstermmorpg@gmail.com You can use T5 cpu almost same speed

mastermk

Hi Furkan. Can I generate videos without speech or background sounds, or with one or the other?

SECourses

Well you can mute sound. I added checkbox to generate without speech tags option. Try that too . But will check if no audio option available or not

Justin

Same thing is happening to me. Fresh install with nothing in the output folder on v6.2. It worked fine on 2.5

SECourses

please make a generation, copy entire cmd and send me logs : monstermmorpg@gmail.com

SECourses

just made a fix. please try again with v6.4. update and make sure you see 6.4 on gradio left top

SECourses

just made a fix. please try again with v6.4. update and make sure you see 6.4 on gradio left top

SECourses

just made a fix of example prompt. you were right. now fixed

SECourses

just made a fix. please try again with v6.4. update and make sure you see 6.4 on gradio left top

SECourses

just made a fix. please try again with v6.4. update and make sure you see 6.4 on gradio left top

SECourses

just made a fix. please try again with v6.4. update and make sure you see 6.4 on gradio left top

SECourses

just made a fix. please try again with v6.4. update and make sure you see 6.4 on gradio left top

SECourses

just made a fix. please try again with v6.4. update and make sure you see 6.4 on gradio left top

b

thank you for updating me , yes installing right now

Pratik Rasam

Hi!I am running into the same issue, using 6.4 and running on H200 SXM using runpod

carlos chavez

The latest version does not allow downloading, the Windows system marks it as a virus. I was able to download it without any problem. Thank you very much.

b

'DOSKEY' is not recognized as an internal or external command, operable program or batch file. [DEBUG] Starting premium.py with args: ['premium.py'] [DEBUG] Parsed args: single_generation=False, single_generation_file=False, test=False, test_subprocess=False Starting Gradio interface with lazy loading... Share mode: DISABLED (local only) Use --share flag to enable public access with a shareable URL [DEBUG] Main block: single_generation_file=None, single_generation=None, encode_t5_only=False, test=False, test_subprocess=False * Running on local URL: http://127.0.0.1:7860 * To create a public link, set `share=True` in `launch()`. Auto-loading last used preset: 6-GB GPUs ============================================================ GPU DETECTION RESULTS: GPU Model: NVIDIA GeForce RTX 4060 GPU Count: 1 VRAM Size: 8.00 GB ============================================================ ============================================================ SYSTEM RAM DETECTION RESULTS: Total RAM: 63.76 GB ============================================================ ✓ VRAM optimization: Enabled Scaled FP8 T5 + Tiled VAE (VRAM: 8.0GB < 23GB) [PRESET] Applied automatic optimizations for '6-GB GPUs': VRAM 8.0GB < 23GB → Enabled Scaled FP8 T5 + Tiled VAE ============================================================ GPU DETECTION RESULTS: GPU Model: NVIDIA GeForce RTX 4060 GPU Count: 1 VRAM Size: 8.00 GB ============================================================ ============================================================ SYSTEM RAM DETECTION RESULTS: Total RAM: 63.76 GB ============================================================ ✓ VRAM optimization: Enabled Scaled FP8 T5 + Tiled VAE (VRAM: 8.0GB < 23GB) [PRESET] Applied automatic optimizations for '6-GB GPUs': VRAM 8.0GB < 23GB → Enabled Scaled FP8 T5 + Tiled VAE ============================================================ GPU DETECTION RESULTS: GPU Model: NVIDIA GeForce RTX 4060 GPU Count: 1 VRAM Size: 8.00 GB ============================================================ ============================================================ SYSTEM RAM DETECTION RESULTS: Total RAM: 63.76 GB ============================================================ ✓ VRAM optimization: Enabled Scaled FP8 T5 + Tiled VAE (VRAM: 8.0GB < 23GB) [PRESET] Applied automatic optimizations for '6-GB GPUs': VRAM 8.0GB < 23GB → Enabled Scaled FP8 T5 + Tiled VAE [AUTO-DETECT] Image 1424×736 (aspect 1.935) → Closest ratio: 16:9 → 960×544 [AUTO-DETECT] Image 1424×736 (aspect 1.935) → Closest ratio: 16:9 → 960×544 [LORA DEBUG] Received from UI: lora_1='None', lora_1_scale=1, lora_1_layers='Video Layers' lora_2='None', lora_2_scale=1, lora_2_layers='Video Layers' lora_3='None', lora_3_scale=1, lora_3_layers='Video Layers' lora_4='None', lora_4_scale=1, lora_4_layers='Video Layers' lora_specs=None [LORA STATUS] UI selections will be processed: 4 LoRA(s) [1] None (scale: 1) [2] None (scale: 1) [3] None (scale: 1) [4] None (scale: 1) [RESOLUTION] Using exact user-specified resolution: 960x544 ================================================================================ VIDEO GENERATION STARTED enable_multiline_prompts: False enable_video_extension: False Text prompt: A stylish radio host in a professional studio lean... Image path: C:\FFF\Ovi_Pro_v1\Ovi_Pro\temp\cropped_20251008_113012_982984.png Resolution: 544x960 Base Resolution: 720x720 Duration: 5 seconds Seed: 99 Num generations per prompt: 1 Video extensions: 0 Valid prompt lines detected: 1 ================================================================================ ================================================================================ CLEAR ALL MEMORY ENABLED Main process will NOT load any models All generations will run in separate subprocesses VRAM/RAM will be completely cleared between generations ================================================================================ [DEBUG] Raw image path from upload: C:\FFF\Ovi_Pro_v1\Ovi_Pro\temp\cropped_20251008_113012_982984.png [AUTO-CROP] Auto-cropping image for generation... [AUTO-CROP] Using provided resolution: 960x544 (16:9) [AUTO-CROP] Cropped image saved to: C:\FFF\Ovi_Pro_v1\Ovi_Pro\temp\cropped_20251008_113021_860217.png [DEBUG] Final image file exists: Yes (843368 bytes) [PROMPT 1/1] Processing: A stylish radio host in a professional studio lean... [GENERATION 1/1] Starting with seed: 99 [GENERATION LORA] No LoRAs applied (using base model only) [SUBPROCESS DEBUG] Passing lora_specs to subprocess: [] [SUBPROCESS DEBUG] lora_specs type: , length: 0 [SUBPROCESS] Running generation in subprocess... [SUBPROCESS] Command: C:\FFF\Ovi_Pro_v1\Ovi_Pro\venv\Scripts\python.exe C:\FFF\Ovi_Pro_v1\Ovi_Pro\premium.py --single-generation-file C:\FFF\Ovi_Pro_v1\Ovi_Pro\tmp13rqzoz6.json [SUBPROCESS] Params file: C:\FFF\Ovi_Pro_v1\Ovi_Pro\tmp13rqzoz6.json [DEBUG] Starting premium.py with args: ['C:\\FFF\\Ovi_Pro_v1\\Ovi_Pro\\premium.py', '--single-generation-file', 'C:\\FFF\\Ovi_Pro_v1\\Ovi_Pro\\tmp13rqzoz6.json'] [DEBUG] Parsed args: single_generation=False, single_generation_file=True, test=False, test_subprocess=False Starting Gradio interface with lazy loading... Share mode: DISABLED (local only) Use --share flag to enable public access with a shareable URL [DEBUG] Main block: single_generation_file=C:\FFF\Ovi_Pro_v1\Ovi_Pro\tmp13rqzoz6.json, single_generation=None, encode_t5_only=False, test=False, test_subprocess=False [DEBUG] Taking single_generation_file path [SINGLE-GEN] Loaded params from file: C:\FFF\Ovi_Pro_v1\Ovi_Pro\tmp13rqzoz6.json [SINGLE-GEN] Starting generation with params: ['text_prompt', 'image', 'video_frame_height', 'video_frame_width', 'video_seed', 'solver_name', 'sample_steps', 'shift', 'video_guidance_scale', 'audio_guidance_scale', 'slg_layer', 'blocks_to_swap', 'video_negative_prompt', 'audio_negative_prompt', 'use_image_gen', 'cpu_offload', 'delete_text_encoder', 'fp8_t5', 'cpu_only_t5', 'fp8_base_model', 'use_sage_attention', 'no_audio', 'no_block_prep', 'num_generations', 'randomize_seed', 'save_metadata', 'aspect_ratio', 'clear_all', 'vae_tiled_decode', 'vae_tile_size', 'vae_tile_overlap', 'base_resolution_width', 'base_resolution_height', 'duration_seconds', 'auto_crop_image', 'base_filename', 'output_dir', 'text_embeddings_cache', 'enable_multiline_prompts', 'enable_video_extension', 'disable_auto_prompt_validation', 'force_exact_resolution', 'lora_specs'] [SINGLE-GEN] Text prompt: A stylish radio host in a professional studio lean... [LORA DEBUG] Received from UI: lora_1=None, lora_1_scale=1.0, lora_1_layers='Video Layers' lora_2=None, lora_2_scale=1.0, lora_2_layers='Video Layers' lora_3=None, lora_3_scale=1.0, lora_3_layers='Video Layers' lora_4=None, lora_4_scale=1.0, lora_4_layers='Video Layers' lora_specs=[] [LORA STATUS] No LoRAs selected [LORA] Using pre-built lora_specs from subprocess: 0 LoRA(s) [RESOLUTION] Using exact user-specified resolution: 960x544 ================================================================================ VIDEO GENERATION STARTED enable_multiline_prompts: False enable_video_extension: False Text prompt: A stylish radio host in a professional studio lean... Image path: C:\FFF\Ovi_Pro_v1\Ovi_Pro\temp\cropped_20251008_113021_860217.png Resolution: 544x960 Base Resolution: 720x720 Duration: 5 seconds Seed: 99 Num generations per prompt: 1 Video extensions: 0 Valid prompt lines detected: 1 ================================================================================ ================================================================================ INITIALIZING OVI FUSION ENGINE IN MAIN PROCESS Block Swap: 29 blocks (0 = disabled) CPU Offload: True Image Generation: False No Block Prep: False Note: Models will be loaded in main process (Clear All Memory disabled) ================================================================================ [OK] OviFusionEngine initialized successfully (models will load on first generation) [DEBUG] Raw image path from upload: C:\FFF\Ovi_Pro_v1\Ovi_Pro\temp\cropped_20251008_113021_860217.png [AUTO-CROP] Auto-cropping image for generation... [AUTO-CROP] Using provided resolution: 960x544 (16:9) [AUTO-CROP] Cropped image saved to: C:\FFF\Ovi_Pro_v1\Ovi_Pro\temp\cropped_20251008_113027_976916.png [DEBUG] Final image file exists: Yes (843368 bytes) [PROMPT 1/1] Processing: A stylish radio host in a professional studio lean... [GENERATION 1/1] Starting with seed: 99 [GENERATION LORA] No LoRAs applied (using base model only) [DEBUG] No text_embeddings_cache available - T5 will be loaded in-process [SAGE ATTENTION] Enabled - using Sage Attention for ~10% speedup & lower VRAM [DEBUG] text_embeddings_cache is None - will load T5 ================================================================================ STEP 1/2: Loading T5 text encoder FIRST to minimize RAM usage ================================================================================ ================================================================================ Loading OVI models for first generation... Block Swap: 29 blocks CPU Offload: True ================================================================================ Initial VRAM: 0.00 GB Removing weight norm... ================================================================================ T5 CPU-ONLY + SCALED FP8: Loading FP8 T5 on CPU for maximum memory efficiency Expected RAM savings: ~50% (~2.5GB saved) compared to full precision on CPU Text encoding will be slower but uses minimal RAM ================================================================================ [T5 CPU-ONLY MODE] Loading T5 text encoder on CPU for CPU inference [T5 CPU-ONLY MODE] This saves VRAM but encoding will be slower [FP8 CACHE] Found cached FP8 checkpoint: C:\FFF\Ovi_Pro_v1\Ovi_Pro\ckpts\Wan2.2-TI2V-5B\models_t5_umt5-xxl-enc-fp8_scaled.safetensors [FP8 CACHE] Creating structure on CPU first (avoids BF16 VRAM allocation) [T5 LOAD][FP8] Structure created on CPU in 21.92s (FP8 cached path) [SUBPROCESS] Generation failed with return code: 3221225477 [GENERATION 1/1] No output file found in C:\FFF\Ovi_Pro_v1\Ovi_Pro\outputs after retries Total generation time: 42.15 seconds ================================================================================ VIDEO GENERATION COMPLETED Final output path: None File exists: No ================================================================================

SECourses

Windows error. You can see every file content with notepad. You have to allow it to download.

Agino Terra

I'm having a problem; the program just keeps looping and never starts. Ovi Pro SECourses Premium App v6.4, this is the one I have

b

updated to latest version above and tried Total generation time: 42.15 seconds ================================================================================ VIDEO GENERATION COMPLETED Final output path: None File exists: No ================================================================================ can you check error log above i sent 5hrs back

SECourses

I have uploaded v7 zip file can you test it and let me know, i guess it was due to i just changed file name without anything else and it triggered windows somehow

SECourses

Sorry for this error. Finally i figured out the reason. Please read V7 logs and update to V7. this should fix it

SECourses

Sorry for this error. Finally i figured out the reason. Please read V7 logs and update to V7. this should fix it

SECourses

I have uploaded v7 zip file can you test it and let me know, i guess it was due to i just changed file name without anything else and it triggered windows somehow

SECourses

Sorry for this error. Finally i figured out the reason. Please read V7 logs and update to V7. this should fix it

SECourses

Sorry for this error. Finally i figured out the reason. Please read V7 logs and update to V7. this should fix it

SECourses

Sorry for this error. Finally i figured out the reason. Please read V7 logs and update to V7. this should fix it

SECourses

Sorry for this error. Finally i figured out the reason. Please read V7 logs and update to V7. this should fix it

SECourses

Sorry for this error. Finally i figured out the reason. Please read V7 logs and update to V7. this should fix it

SECourses

Sorry for this error. Finally i figured out the reason. Please read V7 logs and update to V7. this should fix it

b

updated to v7 can you please check? is it due to disk space ? i have virtual disk like 25gb only 'DOSKEY' is not recognized as an internal or external command, operable program or batch file. [DEBUG] Starting premium.py with args: ['premium.py'] [DEBUG] Parsed args: single_generation=False, single_generation_file=False, test=False, test_subprocess=False Starting Gradio interface with lazy loading... Share mode: DISABLED (local only) Use --share flag to enable public access with a shareable URL [DEBUG] Main block: single_generation_file=None, single_generation=None, encode_t5_only=False, test=False, test_subprocess=False * Running on local URL: http://127.0.0.1:7860 * To create a public link, set `share=True` in `launch()`. Auto-loading last used preset: 6-GB GPUs ============================================================ GPU DETECTION RESULTS: GPU Model: NVIDIA GeForce RTX 4060 GPU Count: 1 VRAM Size: 8.00 GB ============================================================ ============================================================ SYSTEM RAM DETECTION RESULTS: Total RAM: 63.76 GB ============================================================ ✓ VRAM optimization: Enabled Scaled FP8 T5 + Tiled VAE (VRAM: 8.0GB < 23GB) [PRESET] Applied automatic optimizations for '6-GB GPUs': VRAM 8.0GB < 23GB → Enabled Scaled FP8 T5 + Tiled VAE ============================================================ GPU DETECTION RESULTS: GPU Model: NVIDIA GeForce RTX 4060 GPU Count: 1 VRAM Size: 8.00 GB ============================================================ ============================================================ SYSTEM RAM DETECTION RESULTS: Total RAM: 63.76 GB ============================================================ ✓ VRAM optimization: Enabled Scaled FP8 T5 + Tiled VAE (VRAM: 8.0GB < 23GB) [PRESET] Applied automatic optimizations for '6-GB GPUs': VRAM 8.0GB < 23GB → Enabled Scaled FP8 T5 + Tiled VAE ============================================================ GPU DETECTION RESULTS: GPU Model: NVIDIA GeForce RTX 4060 GPU Count: 1 VRAM Size: 8.00 GB ============================================================ ============================================================ SYSTEM RAM DETECTION RESULTS: Total RAM: 63.76 GB ============================================================ ✓ VRAM optimization: Enabled Scaled FP8 T5 + Tiled VAE (VRAM: 8.0GB < 23GB) [PRESET] Applied automatic optimizations for '6-GB GPUs': VRAM 8.0GB < 23GB → Enabled Scaled FP8 T5 + Tiled VAE [AUTO-DETECT] Image 1424×736 (aspect 1.935) → Closest ratio: 16:9 → 960×544 [AUTO-DETECT] Image 1424×736 (aspect 1.935) → Closest ratio: 16:9 → 960×544 Traceback (most recent call last): File "C:\FFF\Ovi_Pro_v1\Ovi_Pro\venv\lib\site-packages\gradio\queueing.py", line 759, in process_events response = await route_utils.call_process_api( File "C:\FFF\Ovi_Pro_v1\Ovi_Pro\venv\lib\site-packages\gradio\route_utils.py", line 354, in call_process_api output = await app.get_blocks().process_api( File "C:\FFF\Ovi_Pro_v1\Ovi_Pro\venv\lib\site-packages\gradio\blocks.py", line 2112, in process_api inputs = await self.preprocess_data( File "C:\FFF\Ovi_Pro_v1\Ovi_Pro\venv\lib\site-packages\gradio\blocks.py", line 1774, in preprocess_data processed_input.append(block.preprocess(inputs_cached)) File "C:\FFF\Ovi_Pro_v1\Ovi_Pro\venv\lib\site-packages\gradio\components\number.py", line 142, in preprocess self.raise_if_out_of_bounds(payload, self.minimum, self.maximum) File "C:\FFF\Ovi_Pro_v1\Ovi_Pro\venv\lib\site-packages\gradio\components\number.py", line 129, in raise_if_out_of_bounds raise Error(f"Value {num} is less than minimum value {minimum}.") gradio.exceptions.Error: 'Value 1 is less than minimum value 128.' Traceback (most recent call last): File "C:\FFF\Ovi_Pro_v1\Ovi_Pro\venv\lib\site-packages\gradio\queueing.py", line 759, in process_events response = await route_utils.call_process_api( File "C:\FFF\Ovi_Pro_v1\Ovi_Pro\venv\lib\site-packages\gradio\route_utils.py", line 354, in call_process_api output = await app.get_blocks().process_api( File "C:\FFF\Ovi_Pro_v1\Ovi_Pro\venv\lib\site-packages\gradio\blocks.py", line 2112, in process_api inputs = await self.preprocess_data( File "C:\FFF\Ovi_Pro_v1\Ovi_Pro\venv\lib\site-packages\gradio\blocks.py", line 1774, in preprocess_data processed_input.append(block.preprocess(inputs_cached)) File "C:\FFF\Ovi_Pro_v1\Ovi_Pro\venv\lib\site-packages\gradio\components\number.py", line 142, in preprocess self.raise_if_out_of_bounds(payload, self.minimum, self.maximum) File "C:\FFF\Ovi_Pro_v1\Ovi_Pro\venv\lib\site-packages\gradio\components\number.py", line 129, in raise_if_out_of_bounds raise Error(f"Value {num} is less than minimum value {minimum}.") gradio.exceptions.Error: 'Value 1 is less than minimum value 128.' [LORA DEBUG] Received from UI: lora_1='None', lora_1_scale=1, lora_1_layers='Video Layers' lora_2='None', lora_2_scale=1, lora_2_layers='Video Layers' lora_3='None', lora_3_scale=1, lora_3_layers='Video Layers' lora_4='None', lora_4_scale=1, lora_4_layers='Video Layers' lora_specs=None [LORA STATUS] UI selections will be processed: 4 LoRA(s) [1] None (scale: 1) [2] None (scale: 1) [3] None (scale: 1) [4] None (scale: 1) [RESOLUTION] Using exact user-specified resolution: 832x608 ================================================================================ VIDEO GENERATION STARTED enable_multiline_prompts: False enable_video_extension: False Text prompt: A stylish radio host in a professional studio lean... Image path: C:\FFF\Ovi_Pro_v1\Ovi_Pro\temp\cropped_20251008_194453_073331.png Resolution: 608x832 Base Resolution: 720x720 Duration: 3 seconds Seed: 99 Num generations per prompt: 1 Video extensions: 0 Valid prompt lines detected: 1 ================================================================================ ================================================================================ CLEAR ALL MEMORY ENABLED Main process will NOT load any models All generations will run in separate subprocesses VRAM/RAM will be completely cleared between generations ================================================================================ [DEBUG] Raw image path from upload: C:\FFF\Ovi_Pro_v1\Ovi_Pro\temp\cropped_20251008_194453_073331.png [AUTO-CROP] Auto-cropping image for generation... [AUTO-CROP] Using provided resolution: 832x608 (4:3) [AUTO-CROP] Cropped image saved to: C:\FFF\Ovi_Pro_v1\Ovi_Pro\temp\cropped_20251008_194509_603784.png [DEBUG] Final image file exists: Yes (802466 bytes) [PROMPT 1/1] Processing: A stylish radio host in a professional studio lean... [GENERATION 1/1] Starting with seed: 99 [GENERATION LORA] No LoRAs applied (using base model only) [SUBPROCESS DEBUG] Passing lora_specs to subprocess: [] [SUBPROCESS DEBUG] lora_specs type: , length: 0 [SUBPROCESS] Running generation in subprocess... [SUBPROCESS] Command: C:\FFF\Ovi_Pro_v1\Ovi_Pro\venv\Scripts\python.exe C:\FFF\Ovi_Pro_v1\Ovi_Pro\premium.py --single-generation-file C:\FFF\Ovi_Pro_v1\Ovi_Pro\tmpqvmmmszo.json [SUBPROCESS] Params file: C:\FFF\Ovi_Pro_v1\Ovi_Pro\tmpqvmmmszo.json [DEBUG] Starting premium.py with args: ['C:\\FFF\\Ovi_Pro_v1\\Ovi_Pro\\premium.py', '--single-generation-file', 'C:\\FFF\\Ovi_Pro_v1\\Ovi_Pro\\tmpqvmmmszo.json'] [DEBUG] Parsed args: single_generation=False, single_generation_file=True, test=False, test_subprocess=False Starting Gradio interface with lazy loading... Share mode: DISABLED (local only) Use --share flag to enable public access with a shareable URL [DEBUG] Main block: single_generation_file=C:\FFF\Ovi_Pro_v1\Ovi_Pro\tmpqvmmmszo.json, single_generation=None, encode_t5_only=False, test=False, test_subprocess=False [DEBUG] Taking single_generation_file path [SINGLE-GEN] Loaded params from file: C:\FFF\Ovi_Pro_v1\Ovi_Pro\tmpqvmmmszo.json [SINGLE-GEN] Starting generation with params: ['text_prompt', 'image', 'video_frame_height', 'video_frame_width', 'video_seed', 'solver_name', 'sample_steps', 'shift', 'video_guidance_scale', 'audio_guidance_scale', 'slg_layer', 'blocks_to_swap', 'video_negative_prompt', 'audio_negative_prompt', 'use_image_gen', 'cpu_offload', 'delete_text_encoder', 'fp8_t5', 'cpu_only_t5', 'fp8_base_model', 'use_sage_attention', 'no_audio', 'no_block_prep', 'num_generations', 'randomize_seed', 'save_metadata', 'aspect_ratio', 'clear_all', 'vae_tiled_decode', 'vae_tile_size', 'vae_tile_overlap', 'base_resolution_width', 'base_resolution_height', 'duration_seconds', 'auto_crop_image', 'base_filename', 'output_dir', 'text_embeddings_cache', 'enable_multiline_prompts', 'enable_video_extension', 'disable_auto_prompt_validation', 'force_exact_resolution', 'lora_specs'] [SINGLE-GEN] Text prompt: A stylish radio host in a professional studio lean... [LORA DEBUG] Received from UI: lora_1=None, lora_1_scale=1.0, lora_1_layers='Video Layers' lora_2=None, lora_2_scale=1.0, lora_2_layers='Video Layers' lora_3=None, lora_3_scale=1.0, lora_3_layers='Video Layers' lora_4=None, lora_4_scale=1.0, lora_4_layers='Video Layers' lora_specs=[] [LORA STATUS] No LoRAs selected [LORA] Using pre-built lora_specs from subprocess: 0 LoRA(s) [RESOLUTION] Using exact user-specified resolution: 832x608 ================================================================================ VIDEO GENERATION STARTED enable_multiline_prompts: False enable_video_extension: False Text prompt: A stylish radio host in a professional studio lean... Image path: C:\FFF\Ovi_Pro_v1\Ovi_Pro\temp\cropped_20251008_194509_603784.png Resolution: 608x832 Base Resolution: 720x720 Duration: 3 seconds Seed: 99 Num generations per prompt: 1 Video extensions: 0 Valid prompt lines detected: 1 ================================================================================ ================================================================================ INITIALIZING OVI FUSION ENGINE IN MAIN PROCESS Block Swap: 29 blocks (0 = disabled) CPU Offload: True Image Generation: False No Block Prep: False Note: Models will be loaded in main process (Clear All Memory disabled) ================================================================================ [OK] OviFusionEngine initialized successfully (models will load on first generation) [DEBUG] Raw image path from upload: C:\FFF\Ovi_Pro_v1\Ovi_Pro\temp\cropped_20251008_194509_603784.png [AUTO-CROP] Auto-cropping image for generation... [AUTO-CROP] Using provided resolution: 832x608 (4:3) [AUTO-CROP] Cropped image saved to: C:\FFF\Ovi_Pro_v1\Ovi_Pro\temp\cropped_20251008_194515_533184.png [DEBUG] Final image file exists: Yes (802466 bytes) [PROMPT 1/1] Processing: A stylish radio host in a professional studio lean... [GENERATION 1/1] Starting with seed: 99 [GENERATION LORA] No LoRAs applied (using base model only) [DEBUG] No text_embeddings_cache available - T5 will be loaded in-process [SAGE ATTENTION] Enabled - using Sage Attention for ~10% speedup & lower VRAM [T5 CACHE] Cache key: 9bd1dd7f0bf175613a44bc4f5a57a45aaf23bd8e09332c94fa42c921e80da0cf [DEBUG] No cached embeddings - will load T5 ================================================================================ STEP 1/2: Loading T5 text encoder FIRST to minimize RAM usage ================================================================================ ================================================================================ Loading OVI models for first generation... Block Swap: 29 blocks CPU Offload: True ================================================================================ Initial VRAM: 0.00 GB Removing weight norm... ================================================================================ T5 CPU-ONLY + SCALED FP8: Loading FP8 T5 on CPU for maximum memory efficiency Expected RAM savings: ~50% (~2.5GB saved) compared to full precision on CPU Text encoding will be slower but uses minimal RAM ================================================================================ [T5 CPU-ONLY MODE] Loading T5 text encoder on CPU for CPU inference [T5 CPU-ONLY MODE] This saves VRAM but encoding will be slower [FP8 CACHE] Found cached FP8 checkpoint: C:\FFF\Ovi_Pro_v1\Ovi_Pro\ckpts\Wan2.2-TI2V-5B\models_t5_umt5-xxl-enc-fp8_scaled.safetensors [FP8 CACHE] Creating structure on CPU first (avoids BF16 VRAM allocation) [T5 LOAD][FP8] Structure created on CPU in 21.00s (FP8 cached path) [SUBPROCESS] Generation failed with return code: 3221225477 [GENERATION 1/1] No output file found in C:\FFF\Ovi_Pro_v1\Ovi_Pro\outputs after retries Total generation time: 44.64 seconds ================================================================================ VIDEO GENERATION COMPLETED Final output path: None File exists: No ================================================================================ [MEMORY CLEANUP] Final cleanup completed - all generation memory freed

SECourses

ok dont enable FP8 scaled T5. make it cpu T5 and test again. your system doesnt show any obvious error so weird

SECourses

actually best is message me from discord it is faster there to debug issue

SECourses

hi are you on 7.2? can you copy cmd logs starting from first line copy into txt file and email me? monstermmorpg@gmail.com

Juanmyth

Hi. Sorry to say I have the same problem with v7.2. V3 was working fine in 5090. Deleted venv to upgrade from v3 and deleted outputs folder. It generates endless json files in Ovi_Pro directory. The files are automatically deleted when closing the CMD window.

supersonix

Where you ever able to resolve this? I get the same gradio error and have python 3.10.11 installed and working.

SECourses

there is no error atm. delete venv, run installer and send me install logs. entire cmd output : monstermmorpg@gmail.com so i can see why your install fails. actually make a fresh install and you can move ckpt folder to avoid download models

supersonix

Will do so and thank you. Here's a snippet from the install log, which could be the issue? HEAD is now at 23d89cf 7.42 Already up to date. Python launcher is available. Generating Python 3.10 VENV Error: Command '['F:\\Ovi_Pro_v7\\Ovi_Pro\\venv\\Scripts\\python.exe', '-Im', 'ensurepip', '--upgrade', '--default-pip']' returned non-zero exit status 1. '.\venv\Scripts\activate.bat' is not recognized as an internal or external command, operable program or batch file.

Diggy Dre

I am now able to get video, which is pretty cool, but I can't get the voice part to actually read through the script. It randomly bounces back and forth between previous words within the S and E speech brackets.

SECourses

because it is wrong. it is not s s. we even added check :D it is <s> </s>

Diggy Dre

I got it figured out. Actually, updating it again did the trick. Thank you so much! Is there a way to control the camera movement as well? If so, that would be awesome. I tried to do a locked-down shot, but it wouldn't stay put. I also tried to do a forward dolly shot. It kind of got that. I am having fun experimenting with it, but I am finding that it does benefit from the prompts being concise and to the point.

SECourses

ye try some different prompting. preparing a tutorial will show some tricks hopefully

Diggy Dre

Awesome! Thank you so much!

Abu Bhakar

its getting flagged as a virus when downloading the installer

SECourses

Did you try v8 zip file? it only has txt and bat files you can see what is inside. so please allow and download it

Justin

Hey Furkan, thanks for fixing the looping bug. I am able to get video output now. Is there a flag to fun the Gradio server on the local network?

rSandor

Hello there. I have just tried one of the examples and this is the error im getting: traceback (most recent call last): File "C:\Python310\lib\asyncio\events.py", line 80, in _run self._context.run(self._callback, *self._args) File "C:\Python310\lib\asyncio\proactor_events.py", line 165, in _call_connection_lost self._sock.shutdown(socket.SHUT_RDWR) ConnectionResetError: [WinError 10054]

SECourses

that is gradio error. not important. what else do you see after this?

Alexandre Rangel

Great, thank you! Worked, after following the full install tutorial.

SECourses

i dont know if any flags sadly. but --share will start on gradio live so you can use from anywhere

Diggy Dre

I have an interesting question. Suppose you had some pre-recorded audio that you wanted this thing to animate to, would that ever be a feature? Almost like giving it something to lip-sync to?

Mimic

Thank you always! If I may express a personal wish, it would be great if Korean and Japanese voice options were also available. And if there were features like generating videos that lip-sync to the input voice, or generating voices that match sample voices (e.g., GPT-SoVITS), that would be truly amazing. I believe that someday, something even better encompassing these features will emerge. :) Anyway, thank you so much!!

SECourses

thanks a lot for comment. i believe what you want and what i also want will become available soon. currently vibevoice can generate other languages. hopefully i will publish an app for it soon

cool1

Would you be able to make it use all .safetensors files instead of it having some .pt and .pth files (that I think are riskier because those could contain pickled content in theory)?

SECourses

nah they are official sources no issues. possibly safe tensor can be used if there is accurate version

Dan

Hi, I get this error when trying to run the app, I have downloaded everything and run the update. Do you know this error? [STARTUP] Set MKL/OMP threads to 20 for optimal CPU performance Traceback (most recent call last): File "C:\OVI\Ovi_Pro\premium.py", line 21, in from ovi.utils.io_utils import save_video File "C:\OVI\Ovi_Pro\ovi\utils\io_utils.py", line 5, in from moviepy.editor import ImageSequenceClip, AudioFileClip ModuleNotFoundError: No module named 'moviepy' Press any key to continue . . .

SECourses

hi we have it in requirements. that means your install failed for some reason. can you run installer again and email me logs? please delete venv before : monstermmorpg@gmail.com

Taiga

This is very nice. I can't wait to try the batch feature. I'm trying to use an anime-style character. How can I get the mouth movements to be more accurate (this is for pronunciation teaching)

SECourses

i suppose closer shot is better. also it has 2 options you can try and see : Audio Guidance Scale - SLG Layer (Skip Layer Guidance layer - affects audio-video synchronization)

Diggy Dre

Would it ever be possible to set a starting frame and ending frame with this? Also, what all are the types of LoRA that can be used?

SECourses

Wan 2.2 5b loras tested and working. i didnt test others. sadly you cant set ending frame. only beginning frame

Musclelogs

I encounter the same problem. Were you able to get this fixed?

Pedro Burle

I've got this erro on first generation. What it could be? Initial VRAM: 0.00 GB Removing weight norm... ================================================================================ SCALED FP8 T5: Loading T5 in Scaled FP8 format Expected VRAM savings: ~50% (~5-6GB saved) ================================================================================ [FP8 CACHE] Found cached FP8 checkpoint: E:\AI\Ovi_Pro_v8\Ovi_Pro\ckpts\Wan2.2-TI2V-5B\models_t5_umt5-xxl-enc-fp8_scaled.safetensors [FP8 CACHE] Creating structure on CPU first (avoids BF16 VRAM allocation) [T5 LOAD][FP8] Structure created on CPU in 32.66s (FP8 cached path) [SUBPROCESS] Generation failed with return code: 3221225477 [GENERATION 1/1] No output file found in E:\AI\Ovi_Pro_v8\Ovi_Pro\outputs after retries Total generation time: 63.85 seconds ================================================================================ VIDEO GENERATION COMPLETED Final output path: None File exists: No

Dan

Not yet, are looking into it.

SECourses

please message me from discord so we can check better

SECourses

please message me from discord so we can check better

SECourses

out of RAM. how much RAM you have? did you set 100 gb virtual disk?

FalconBravery

If I load a lora, the process requires quite some (big) time: [VIDEO MODEL] Merging 1 LoRA(s)... Merging LoRA layers: 79%|█ It takes about 5 minutes, but I see my runpod (H100 + cpu xeon platinum 8352Y), the cpu is stuck at almost 100%, and gpu not running. Maybe there's space for some optimization on lora loading (maybe now device is set on cpu). Hope this can help!

SECourses

which preset you using? changed any settings? it should be instant normally

Hockey

I would like this feature too if possible.

Pedro Burle

I have 16 gb of VRAM and i gave the windows permission to manage the amout of virtual disk, do i must set it to 100 gb?

SECourses

i will test today hopefully sorry for delay. i will rent h100 to test

FalconBravery

I did some more testing. Some pods are extremely slow (and take even 2 hours extimated time to load a single lora). So I just stop and start a new pod, and, if pod is a good one with no issues, average time to load lora is 3-5 minutes (using H100). Thanks for ur message!

SECourses

you were right. i made an update with v 8.3 and now it will auto enable Merge LoRAs on GPU for 80 gb and 96 gb configs and ultra fast

SECourses

you were right. i made an update with v 8.3 and now it will auto enable Merge LoRAs on GPU for 80 gb and 96 gb configs and ultra fast

SECourses

yes 100 gb is recommended. the lesser RAM you have more virtual disk helps

thom mick

Thank you very much for this great work. I've already done some tests and am happy about the possibility of working locally. For the first time, I have video and sound, and it works with German voice. I hope there will soon be an option to achieve significantly reduced generation times with a 4- or 8-step LORA. Are you working on this?

SECourses

i am looking as well. i hope there will be speed loras

Accessible Game Mods by James

So after trying again this isn't for me especially how you can't let the app generate what it wants to say for you instead of you telling it everything. Anyway great job on the app

SECourses

thanks. next gen will be hopefully even better model and app :)

Ant-2014

Hello, I have all python, cuda etc locally, but always get this error , whatever t2v, i2v : Can you help me, please? VAE DECODE COMPLETE VRAM after: 2.38 GB Peak during decode: 7.51 GB VRAM used by decode: 5.46 GB ================================================================================ Error during video generation: [WinError 2] The system cannot find the file specified [SINGLE-GEN] Failed - no result returned [SUBPROCESS] Generation failed with return code: 1 [GENERATION 1/1] No output file found in N:\AI\AI_Video\Ovi_Pro_v8\Ovi_Pro\outputs after retries Total generation time: 188.18 seconds ================================================================================ VIDEO GENERATION COMPLETED Final output path: None File exists: No ================================================================================ [MEMORY CLEANUP] Final cleanup completed - all generation memory freed

SECourses

yes out of RAM. how much RAM you have?

Ant-2014

Hi. I have VRAM 24Gb, all parameters as you write. I checked all items and then I reinstalled all. Again, The app calculates/generates, but in the end I always have only this message : "Error during video generation: [WinError 2] The system cannot find the file specified" without any video in the Output. only wav founded in Output folder. ================================================================================ STARTING VAE DECODE - VRAM before: 2.07 GB ================================================================================ VAE DECODE PROGRESS: Decoding video (standard mode)... VAE DECODE PROGRESS: Standard decode completed ================================================================================ VAE DECODE COMPLETE VRAM after: 2.81 GB Peak during decode: 14.33 GB VRAM used by decode: 12.25 GB ================================================================================ Error during video generation: [WinError 2] The system cannot find the file specified [SINGLE-GEN] Failed - no result returned [SUBPROCESS] Generation failed with return code: 1 [GENERATION 1/1] No output file found in N:\AI\AI_Video\Ovi_Pro_v8\Ovi_Pro\outputs after retries Total generation time: 332.54 seconds ================================================================================ VIDEO GENERATION COMPLETED Final output path: None File exists: No ================================================================================ [MEMORY CLEANUP] Final cleanup completed - all generation memory freed Please I need help. Thank you for your job and other programs and scripts.

Ant-2014

Hi. I have win 10, python etc, 64-RAM. 24-VRAM

Ant-2014

I checked also - about (this MAX peak) 60 Gb of RAM used. But execution is going normaly. Only in final I have the message of this error. I use your preset for 24 VRAM and fp8 scaled. I think it is normal for my configuration.

Ant-2014

Thank you very much. I will try it. Virtual memory expanding..You do good job and nice that have the some time to check and answer for us. Forces for you!

SECourses

thanks let me know after 100 gb virtual disk setup and restart

Ant-2014

Good morning, sir. I got the 100 gb virtual disk setup on the nvme disk D: , (and C: system disk also is nvme) restarted win, check. Then app and i have the same message as before. [STARTUP] Set MKL/OMP threads to 20 for optimal CPU performance [DEBUG] Parsed args: single_generation=False, single_generation_file=False, test=False, test_subprocess=False Starting Gradio interface with lazy loading... Share mode: DISABLED (local only) Use --share flag to enable public access with a shareable URL Ovi Pro SECourses Premium App v8.3 [DEBUG] Main block: single_generation_file=None, single_generation=None, encode_t5_only=False, test=False, test_subprocess=False * Running on local URL: http://127.0.0.1:7860 * To create a public link, set `share=True` in `launch()`. Auto-loading last used preset: 24-GB GPUs - Faster FP8 Scaled ============================================================ GPU DETECTION RESULTS: GPU Model: NVIDIA GeForce RTX 4090 GPU Count: 1 VRAM Size: 23.99 GB ============================================================ ============================================================ SYSTEM RAM DETECTION RESULTS: Total RAM: 63.79 GB ============================================================ ✓ RAM optimization: Enabled Clear All Memory (RAM: 63.8GB < 128GB) [PRESET] Applied automatic optimizations for '24-GB GPUs - Faster FP8 Scaled': RAM 63.8GB < 128GB → Enabled Clear All Memory ============================================================ GPU DETECTION RESULTS: GPU Model: NVIDIA GeForce RTX 4090 GPU Count: 1 VRAM Size: 23.99 GB ============================================================ ============================================================ SYSTEM RAM DETECTION RESULTS: Total RAM: 63.79 GB ============================================================ ✓ RAM optimization: Enabled Clear All Memory (RAM: 63.8GB < 128GB) [PRESET] Applied automatic optimizations for '24-GB GPUs - Faster FP8 Scaled': RAM 63.8GB < 128GB → Enabled Clear All Memory ============================================================ GPU DETECTION RESULTS: GPU Model: NVIDIA GeForce RTX 4090 GPU Count: 1 VRAM Size: 23.99 GB ============================================================ ============================================================ SYSTEM RAM DETECTION RESULTS: Total RAM: 63.79 GB ============================================================ ✓ RAM optimization: Enabled Clear All Memory (RAM: 63.8GB < 128GB) [PRESET] Applied automatic optimizations for '24-GB GPUs - Faster FP8 Scaled': RAM 63.8GB < 128GB → Enabled Clear All Memory [LORA DEBUG] Received from UI: lora_1='None', lora_1_scale=1, lora_1_layers='Video Layers' lora_2='None', lora_2_scale=1, lora_2_layers='Video Layers' lora_3='None', lora_3_scale=1, lora_3_layers='Video Layers' lora_4='None', lora_4_scale=1, lora_4_layers='Video Layers' lora_specs=None [LORA STATUS] UI selections will be processed: 4 LoRA(s) [1] None (scale: 1) [2] None (scale: 1) [3] None (scale: 1) [4] None (scale: 1) [RESOLUTION] Using exact user-specified resolution: 992x512 ================================================================================ VIDEO GENERATION STARTED enable_multiline_prompts: False enable_video_extension: False Text prompt: A man in a library, surrounded by old book... Image path: None Resolution: 512x992 Base Resolution: 720x720 Duration: 1 seconds Seed: 99 Num generations per prompt: 1 Video extensions: 0 Valid prompt lines detected: 1 ================================================================================ ================================================================================ CLEAR ALL MEMORY ENABLED Main process will NOT load any models All generations will run in separate subprocesses VRAM/RAM will be completely cleared between generations ================================================================================ [PROMPT 1/1] Processing: A man in a library, surrounded by old book... [GENERATION 1/1] Starting with seed: 99 [GENERATION LORA] No LoRAs applied (using base model only) [SUBPROCESS DEBUG] Passing lora_specs to subprocess: [] [SUBPROCESS DEBUG] lora_specs type: , length: 0 [SUBPROCESS] Running generation in subprocess... [SUBPROCESS] Command: N:\AI\AI_Video\Ovi_Pro_v8\Ovi_Pro\venv\Scripts\python.exe N:\AI\AI_Video\Ovi_Pro_v8\Ovi_Pro\premium.py --single-generation-file N:\AI\AI_Video\Ovi_Pro_v8\Ovi_Pro\tmpphf63n06.json [SUBPROCESS] Params file: N:\AI\AI_Video\Ovi_Pro_v8\Ovi_Pro\tmpphf63n06.json [DEBUG] Parsed args: single_generation=False, single_generation_file=True, test=False, test_subprocess=False Starting Gradio interface with lazy loading... Share mode: DISABLED (local only) Use --share flag to enable public access with a shareable URL Ovi Pro SECourses Premium App v8.3 [DEBUG] Main block: single_generation_file=N:\AI\AI_Video\Ovi_Pro_v8\Ovi_Pro\tmpphf63n06.json, single_generation=None, encode_t5_only=False, test=False, test_subprocess=False [DEBUG] Taking single_generation_file path [SINGLE-GEN] Loaded params from file: N:\AI\AI_Video\Ovi_Pro_v8\Ovi_Pro\tmpphf63n06.json [SINGLE-GEN] Starting generation with params: ['text_prompt', 'image', 'video_frame_height', 'video_frame_width', 'video_seed', 'solver_name', 'sample_steps', 'shift', 'video_guidance_scale', 'audio_guidance_scale', 'slg_layer', 'blocks_to_swap', 'video_negative_prompt', 'audio_negative_prompt', 'use_image_gen', 'cpu_offload', 'delete_text_encoder', 'fp8_t5', 'cpu_only_t5', 'fp8_base_model', 'use_sage_attention', 'no_audio', 'no_block_prep', 'num_generations', 'randomize_seed', 'save_metadata', 'aspect_ratio', 'clear_all', 'vae_tiled_decode', 'vae_tile_size', 'vae_tile_overlap', 'base_resolution_width', 'base_resolution_height', 'duration_seconds', 'auto_crop_image', 'base_filename', 'output_dir', 'text_embeddings_cache', 'enable_multiline_prompts', 'enable_video_extension', 'disable_auto_prompt_validation', 'force_exact_resolution', 'lora_specs', 'merge_loras_on_gpu'] [SINGLE-GEN] Text prompt: A man in a library, surrounded by old book... [LORA DEBUG] Received from UI: lora_1=None, lora_1_scale=1.0, lora_1_layers='Video Layers' lora_2=None, lora_2_scale=1.0, lora_2_layers='Video Layers' lora_3=None, lora_3_scale=1.0, lora_3_layers='Video Layers' lora_4=None, lora_4_scale=1.0, lora_4_layers='Video Layers' lora_specs=[] [LORA STATUS] No LoRAs selected [LORA] Using pre-built lora_specs from subprocess: 0 LoRA(s) [RESOLUTION] Using exact user-specified resolution: 992x512 ================================================================================ VIDEO GENERATION STARTED enable_multiline_prompts: False enable_video_extension: False Text prompt: A man in a library, surrounded by old book... Image path: None Resolution: 512x992 Base Resolution: 720x720 Duration: 1 seconds Seed: 99 Num generations per prompt: 1 Video extensions: 0 Valid prompt lines detected: 1 ================================================================================ ================================================================================ INITIALIZING OVI FUSION ENGINE IN MAIN PROCESS Block Swap: 0 blocks (0 = disabled) CPU Offload: True Image Generation: False No Block Prep: False Note: Models will be loaded in main process (Clear All Memory disabled) ================================================================================ Removing weight norm... [OK] OviFusionEngine initialized successfully (models will load on first generation) [PROMPT 1/1] Processing: A man in a library, surrounded by old book... [GENERATION 1/1] Starting with seed: 99 [GENERATION LORA] No LoRAs applied (using base model only) [DEBUG] No text_embeddings_cache available - T5 will be loaded in-process [SAGE ATTENTION] Enabled - using Sage Attention for ~10% speedup & lower VRAM [T5 CACHE] Cache key: 7978af2284f337605d136dda349bf671cc0864255b1c7f22ff389b2945a437b7 [DEBUG] No cached embeddings - will load T5 ================================================================================ STEP 1/2: Loading T5 text encoder FIRST to minimize RAM usage ================================================================================ ================================================================================ Loading OVI models for first generation... Block Swap: 0 blocks CPU Offload: True ================================================================================ Initial VRAM: 0.00 GB ================================================================================ SCALED FP8 T5: Loading T5 in Scaled FP8 format Expected VRAM savings: ~50% (~5-6GB saved) ================================================================================ [FP8 CACHE] Found cached FP8 checkpoint: N:\AI\AI_Video\Ovi_Pro_v8\Ovi_Pro\ckpts\Wan2.2-TI2V-5B\models_t5_umt5-xxl-enc-fp8_scaled.safetensors [FP8 CACHE] Creating structure on CPU first (avoids BF16 VRAM allocation) [T5 LOAD][FP8] Structure created on CPU in 17.97s (FP8 cached path) [T5 LOAD][FP8] Cached weights file read in 0.05s [T5 LOAD][FP8] Cached weights applied to model in 4.41s [T5 LOAD][FP8] Cached model moved to device in 7.70s [T5 LOAD][FP8] Cached model monkey patching in 0.00s [T5 LOAD][FP8] Cached model layer prep in 0.00s [FP8 CACHE] Loaded cached FP8 T5 encoder successfully [T5 LOAD][FP8] Total time: 25.95s (structure: 17.97s, cache load: 7.97s) [SCALED FP8] T5 text encoder loaded in Scaled FP8 format [SCALED FP8] Expected VRAM savings: ~50% (~2.5GB saved) ================================================================================ T5 loaded. Fusion model will load AFTER text encoding. ================================================================================ STEP 2/2: Encoding text and optionally deleting T5 before loading fusion model ================================================================================ Encoding text prompts... Keeping T5 and offloading to CPU for future reuse... T5 text encoder offloaded to CPU [T5 CACHE SAVE] Saved embeddings to: 7978af2284f337605d136dda349bf671cc0864255b1c7f22ff389b2945a437b7.pt ================================================================================ Loading fusion model (T5 skipped or deleted) ================================================================================ ================================================================================ Loading OVI models for first generation... Block Swap: 0 blocks CPU Offload: True ================================================================================ Initial VRAM: 0.02 GB Step 1/6: Creating model structure on meta device... Score model (Fusion) all parameters:11660753108 Step 2/6: Loading checkpoint weights to CPU... Successfully loaded fusion checkpoint from ./ckpts\Ovi\model.safetensors ================================================================================ APPLYING FP8 OPTIMIZATION TO FUSION MODEL Target: Video + Audio Transformer Blocks Format: FP8 E4M3 with per-block scaling (block_size=64) Expected VRAM savings: ~50% for transformer weights ================================================================================ [FP8 CACHE] Loading cached FP8 checkpoint (no LoRAs applied)... [FP8 CACHE] Loading cached FP8 checkpoint from ./ckpts\Ovi\model_fp8_scaled.safetensors... [FP8 CACHE DEBUG] Found 720 FP8 tensors in cached checkpoint [FP8 CACHE DEBUG] After loading: 720 Linear layers have FP8 weights [FP8 CACHE] Loaded cached FP8 checkpoint successfully [FP8 CACHE] Loaded and patched cached FP8 checkpoint in 0.28s [FP8] Fusion model FP8 optimization complete (from cache) VRAM after loading to CPU: 0.02 GB (should be ~0) Step 4/6: Moving entire model to cpu... VRAM after full model load: 0.02 GB Peak VRAM during loading: 0.02 GB ================================================================================ MODEL LOADING COMPLETE Total parameters: 11,660,753,108 GPU parameters: 0 (0.0%) CPU parameters: 11,660,753,108 (100.0%) Final VRAM allocated: 0.02 GB Final VRAM reserved: 0.05 GB Peak VRAM usage: 0.02 GB CPU offload active: Text encoder on CPU ================================================================================ Pure T2V mode: calculated video latent size: 32 x 62 [FP8] Moved model to cuda:0 with 720 FP8 weights preserved [CPU Offload] Moving model to GPU for inference (no block swap) ================================================================================ INFERENCE STARTING - VRAM: 12.83 GB ================================================================================ 0it [00:00, ?it/s]================================================================================ [ATTENTION BACKEND] >>> SAGE ATTENTION <<< [ATTENTION BACKEND] Sage Attention is ACTIVE for all attention operations ================================================================================ 20it [00:21, 1.06s/it] ================================================================================ INFERENCE COMPLETE - Final VRAM: 0.04 GB (Peak: 13.44 GB) ================================================================================ ================================================================================ STARTING VAE DECODE - VRAM before: 2.04 GB ================================================================================ VAE DECODE PROGRESS: Decoding video (standard mode)... VAE DECODE PROGRESS: Standard decode completed ================================================================================ VAE DECODE COMPLETE VRAM after: 2.17 GB Peak during decode: 13.67 GB VRAM used by decode: 11.63 GB ================================================================================ Error during video generation: [WinError 2] The system cannot find the file specified [SINGLE-GEN] Failed - no result returned [SUBPROCESS] Generation failed with return code: 1 [GENERATION 1/1] No output file found in N:\AI\AI_Video\Ovi_Pro_v8\Ovi_Pro\outputs after retries Total generation time: 89.04 seconds ================================================================================ VIDEO GENERATION COMPLETED Final output path: None File exists: No ================================================================================ [MEMORY CLEANUP] Final cleanup completed - all generation memory freed

Ant-2014

Hi again! I founded the solution: I hope it helps you (and save your time and health for many good prjects) for the next versions. Video generation always failed with error: "Error during video generation: [WinError 2] The system cannot find the file specified" However, all preprocessing steps completed successfully: ✓ GPU detection: RTX 4090 with 24GB VRAM ✓ RAM available: 64GB ✓ T5 text encoder loaded ✓ Inference completed ✓ VAE decode completed successfully The error occurred specifically at the final step: saving MP4 file using moviepy. =============================================================================== ROOT CAUSE ANALYSIS =============================================================================== The issue was in: ovi/utils/io_utils.py The save_video() function uses moviepy.editor to write MP4 files, which internally calls FFmpeg as a subprocess. However, moviepy could not locate the FFmpeg executable because: 1. FFmpeg WAS installed on the system: - Location: C:\ffmpeg\bin\ffmpeg.exe - Installation: Verified with "ffmpeg -version" ✓ 2. BUT moviepy did NOT know where to find it: - The FFMPEG_BINARY environment variable was NOT set - moviepy.config.FFMPEG_BINARY was NOT configured - When moviepy tried subprocess.run("ffmpeg", ...) it failed - Windows returned [WinError 2] - file not found 3. Root cause: Missing FFmpeg path initialization - No code existed to locate and configure FFmpeg before use - Default moviepy behavior expects FFmpeg in system PATH or standard locations - User's installation was in a custom location outside PATH =============================================================================== SOLUTION =============================================================================== I add automatic FFmpeg path detection at module import time in io_utils.py The new code: 1. Searches common FFmpeg installation locations: SIC - In MY CASE!!! - C:\Ovi_Pro_v8\Ovi_Pro\ffmpeg.exe (bundled with app) - C:\ffmpeg\bin\ffmpeg.exe (standard Windows installation) - System PATH (Linux/macOS) - Default Linux/macOS locations 2. Tests if FFmpeg executable is accessible: - Checks file existence - Runs "ffmpeg -version" to verify it works - Handles errors gracefully 3. Configures moviepy before any video processing: - Sets os.environ["FFMPEG_BINARY"] = ffmpeg_path - Sets moviepy.config.FFMPEG_BINARY = ffmpeg_path - Enables subprocess to find FFmpeg correctly 4. Provides detailed diagnostics: - Prints exact FFmpeg path found - Clear error messages if not found - Full traceback logging for debugging =============================================================================== CODE CHANGES =============================================================================== File: ovi/utils/io_utils.py ADD at the top of the file (after imports): def _setup_ffmpeg(): """Find and configure FFmpeg for moviepy.""" ffmpeg_paths = [ os.path.join(os.path.dirname(__file__), "..", "..", "ffmpeg.exe"), "C:\\ffmpeg\\bin\\ffmpeg.exe", "ffmpeg.exe", "/usr/bin/ffmpeg", "/usr/local/bin/ffmpeg", ] ffmpeg_found = None for path in ffmpeg_paths: if os.path.exists(path): ffmpeg_found = path print(f"[FFMPEG] Found FFmpeg at: {ffmpeg_found}") break elif path in ["ffmpeg.exe", "/usr/bin/ffmpeg", "/usr/local/bin/ffmpeg"]: try: result = subprocess.run([path, "-version"], capture_output=True, timeout=5) if result.returncode == 0: ffmpeg_found = path print(f"[FFMPEG] Found FFmpeg in system PATH: {ffmpeg_found}") break except: pass if not ffmpeg_found: raise FileNotFoundError( "FFmpeg not found! Please install FFmpeg or place ffmpeg.exe in:\n" " 1. C:\\Ovi_Pro_v8\\Ovi_Pro\\ffmpeg.exe\n" " 2. C:\\ffmpeg\\bin\\ffmpeg.exe" ) os.environ["FFMPEG_BINARY"] = ffmpeg_found try: import moviepy.config moviepy.config.FFMPEG_BINARY = ffmpeg_found except: pass return ffmpeg_found # Initialize FFmpeg on module import _setup_ffmpeg() =============================================================================== VERIFICATION RESULTS =============================================================================== After applying the fix it generated videos !!! : Test Case 1: Input: Text prompt, 512x992 resolution, 24 FPS, 1 second duration GPU: RTX 4090 (24GB VRAM) Result: ✅ SUCCESS Output file: 0001.mp4 (117,979 bytes) Time: 42.25 seconds Logs show: [SAVE_VIDEO] Starting video save to: C:\Ovi_Pro_v8\Ovi_Pro\outputs\0001.mp4 [SAVE_VIDEO] ImageSequenceClip created successfully [SAVE_VIDEO] Audio WAV created successfully [SAVE_VIDEO] AudioFileClip loaded successfully [SAVE_VIDEO] Video written successfully! [SAVE_VIDEO] Output file verified! Size: 117979 bytes [SINGLE-GEN] Success: C:\Ovi_Pro_v8\Ovi_Pro\outputs\0001.mp4 =============================================================================== 1. Bundle FFmpeg with the application - Include ffmpeg.exe in: Ovi_Pro/ffmpeg.exe - Users won't need separate FFmpeg installation - Works out of the box on Windows 2. Add FFmpeg detection on first run - Show user if FFmpeg is found or missing - Provide download link if missing - Store result in config file 3. Improve error messages - Show which FFmpeg locations were checked - Provide system-specific installation instructions - Include download links in error output 4. Add logging option - Let users enable verbose moviepy logging - Capture full subprocess output - Help debug any future encoding issues This fix is backward compatible - No changes needed to existing code using save_video() - Works on Windows, Linux, and macOS - Minimal performance impact (FFmpeg detection happens once at import) - Includes comprehensive error handling and diagnostics The fix ensures that moviepy can always locate FFmpeg, regardless of system installation or user configuration. =============================================================================== With best regards A :)

SECourses

it fully worked. you dont have ffmpeg VAE DECODE PROGRESS: Decoding video (standard mode)... VAE DECODE PROGRESS: Standard decode completed

SECourses

i am glad you solved because i just had replied your older message you dont have ffmpeg :D that is why requirements tutorial super important

TokyoIdolsAFK

Hi!, everything is great, the samples work I'm using 3080Ti 12GB . However, I'm having issues with LORA, it will not apply, please assist: [VIDEO MODEL] Merging 1 LoRA(s)... WARNING:root:⚠ DETECTED: bfloat16 CPU matmul is catastrophically slow on this system! WARNING:root: Automatically enabling float32 workaround for LoRA merging WARNING:root: Recommendation: Downgrade PyTorch to 2.4.x or 2.5.x (current: 2.8.0+cu129) Merging LoRA layers: 0%| | 0/1035 [00:00CPU transfers for each layer! WARNING:root:Failed to merge LoRA for layer blocks.0.self_attn.q.weight: shape '[3072, 3072]' is invalid for input of size 26214400 WARNING:root:Failed to merge LoRA for layer blocks.0.self_attn.k.weight: shape '[3072, 3072]' is invalid for input of size 26214400 WARNING:root:Failed to merge LoRA for layer blocks.0.self_attn.v.weight: shape '[3072, 3072]' is invalid for input of size 26214400 Merging LoRA layers: 1%|▊ | 15/1035 [00:00<00:07, 139.50it/s]WARNING:root:Failed to merge LoRA for layer blocks.0.self_attn.o.weight: shape '[3072, 3072]' is invalid for input of size 26214400 WARNING:root:Failed to merge LoRA for layer blocks.0.cross_attn.q.weight: shape '[3072, 3072]' is invalid for input of size 26214400 WARNING:root:Failed to merge LoRA for layer blocks.0.cross_attn.k.weight: shape '[3072, 3072]' is invalid for input of size 26214400 WARNING:root:Failed to merge LoRA for layer blocks.0.cross_attn.v.weight: shape '[3072, 3072]' is invalid for input of size 26214400 WARNING:root:Failed to merge LoRA for layer blocks.0.cross_attn.o.weight: shape '[3072, 3072]' is invalid for input of size 26214400 WARNING:root:Failed to merge LoRA for layer blocks.0.ffn.0.weight: shape '[14336, 3072]' is invalid for input of size 70778880 WARNING:root:Failed to merge LoRA for layer blocks.0.ffn.2.weight: shape '[3072, 14336]' is invalid for input of size 70778880 WARNING:root:Failed to merge LoRA for layer blocks.1.self_attn.q.weight: shape '[3072, 3072]' is invalid for input of size 26214400 Merging LoRA layers: 4%|██ | 37/1035 [00:00<00:05, 170.06it/s]WARNING:root:Failed to merge LoRA for layer blocks.1.self_attn.k.weight: shape '[3072, 3072]' is invalid for input of size 26214400

SECourses

looks like your lora is incompatiable. so far we have verified Wan 2.2 5B model loras working. what lora is you trying?

Christoph Behrmann

is there any chance to calculate how long a sentence will be in seconds? of course, it depends on the wording/phrasing... currently running with some AORUS 3090 with 24gb VRAM is good quality and speed, but it takes too long to re-do or fine-tune 2-3 times again and again...

SECourses

i wish i knew but as far as i know not possible to know

SECourses

I wish it was supporting. This is requested feature. We are waiting official Ovi model developers

Gaurav

OVI stopped working on Massed Compute one week back. Please tell me a fix

Christoph Behrmann

how to freeze voice to have same voice for set of multiple videos? or how to add voice file or preset for voice?

SECourses

sadly not possible atm. but i am looking for it

Evgenij Zirnov

Hi, thank you for the work. I have a question. Does it support other languages like German for example or can it be added somehow?

aprisma

For RUNPOD INSTALL ONLY: Subject: Critical Issues in Premium UI Installer – Please Update Dependencies Hi, I wanted to give some important feedback regarding the Premium OVI UI installation process on Linux/RunPod. The current installer pulls incompatible dependency versions, which makes the UI fail on startup. Specifically: Key Issues The installer installs Gradio 6.x, but your UI requires Gradio 4.x because of the theme= parameter in gr.Blocks(). Gradio 4.x requires Pydantic v2, but the installer forces Pydantic v1, which causes: ImportError: cannot import name 'GetCoreSchemaHandler' FastAPI and Starlette versions are also mismatched, causing additional import errors. These problems make the Premium UI fail immediately after installation on a clean environment (RunPod, Ubuntu, venv). Working Dependency Set (Tested & Verified) You should pin your installer to: gradio==4.44.0 fastapi==0.109.2 starlette==0.36.3 pydantic==2.7.4 pydantic-core==2.18.4 This combination fully supports theme=, is compatible with your UI, and works on RunPod, A40, and Ubuntu. Recommendation Please update: your installer script your README and optionally provide a requirements.txt with the exact versions above This would prevent a lot of frustration for new users and reduce support load for you. The product is great — but the installation experience really suffers from these avoidable dependency issues. Thanks for your work, and I hope this helps improve the tool.

SECourses

thanks alot for reporting. updated zip file and pinned to Gradio 5.50.0

aprisma

I am happy to help. Your GUI is amazing fast. I really love it.

SECourses

thank you so much. hopefully more optimization and better features coming

Rizwan Sheikh

Do you have any examples of Loras that would work against the sound layer or any insights into what it would take to create one?

SECourses

sadly i dont have :( nor i do know

SECourses

yes on my list sorry for delay. right now focused on video and image upscaler but after that Ovi update hopefully

Christoph Behrmann

maybe you can check how voice is being created and if there is some chance to upload/add voice file instead of text input as an option?

Christoph Behrmann

it makes no sense not to be able to add matching audio/voice/speech (except unknown character) -- Dwayne Johnson with any AI male voice, come on, that's poor...

Christoph Behrmann

for me it sounds that easy to bring OVI and TTS2 together :) or to find out how to give OVI some voice/speech input, instead of using voice/speech creation

FalconBravery

Hey man, hope you can help. I used your Ovi installation for few months. Recently I installed it (updated versions) on new server, and I always get this error if I launch batch generation: process_batch_generation() takes 53 positional arguments but 54 were given PS: if I use ovi by manually putting image+prompt in the UI, it's working, but if I use batch mode, it doesnt work. The older version that always worked for me is v8. The one that gives me errors is 8.5

FalconBravery

Bug founded, in the v9, line 5903, too many args are sent to the batch function (54 instead of 53). gr.Checkbox(value=False, visible=False) should be removed

SECourses

hey sorry for this let me check and fix hopefully. so you get error at batch right?

SECourses

thanks again for reporting fixed with v8.6

SECourses

thanks again for reporting fixed with v8.6

FalconBravery

Hey man, yesterday I discovered on web this new ai technology called "Mova". It looks like an improved technology of Ovi, 32B parameters (18 active during inference) based on wan 2.2 MoE. I can't find tutorials about it, maybe because it was released on 29 Jan, and it is hardware hungry (vram). Here some info: https://github.com/OpenMOSS/MOVA, https://mosi.cn/models/mova. I just wanted to ask you, if you will research and create tutorials for us or no. Thx😊

SECourses

thanks. but it looks like worse than LTX2 i am not sure