Framepack F1 not showing generated video after completion.

[Redtash1]

Framepack F1 not showing generated video after completion. I'm using Framepack F1, 2 seconds length, Sage Attention, Teacache and everything else is default. It is being saved in output folder. Here is terminal output, not sure if it will help.

(venv) C:\Ai\H1111>h1111.py

• Running on local URL: http://0.0.0.0:7860

To create a public link, set share=True in launch().
Switching DiT path to: hunyuan/FramePack_F1_I2V_HY_20250503.safetensors
Using base input image for --image_path: C:\Users\jrm19\AppData\Local\Temp\gradio\dbfdcf32c403add8fbde9d89b8d97d87e20f732433ea3c2a5cdb33793903c97a\448432177_10232886237679235_6676698719711904109_n_8192x8192.jpg
Using Target Resolution 640 with UI image aspect. Batch-wide bucket: H=640, W=608
Calculated total sections for UI from duration: 1
Using base input image as the primary for --image_path (item 1): C:\Users\jrm19\AppData\Local\Temp\gradio\dbfdcf32c403add8fbde9d89b8d97d87e20f732433ea3c2a5cdb33793903c97a\448432177_10232886237679235_6676698719711904109_n_8192x8192.jpg
Item 1: Using batch-wide dimensions: H=640, W=608
DEBUG: Enabling latent preview every 1 sections with suffix fpack_1749202618_6997.
Running FramePack Command: C:\Ai\H1111\venv\Scripts\python.exe fpack_generate_video.py --text_encoder1 hunyuan/llava_llama3_fp16.safetensors --text_encoder2 hunyuan/clip_l.safetensors --image_encoder hunyuan/model.safetensors --image_path C:\Users\jrm19\AppData\Local\Temp\gradio\dbfdcf32c403add8fbde9d89b8d97d87e20f732433ea3c2a5cdb33793903c97a\448432177_10232886237679235_6676698719711904109_n_8192x8192.jpg --save_path outputs --prompt girl laughing --video_size 640 608 --video_seconds 1 --infer_steps 25 --seed 204583234 --embedded_cfg_scale 10 --guidance_scale 1 --guidance_rescale 0 --latent_window_size 9 --sample_solver unipc --output_type video --attn_mode sageattn --is_f1 --dit hunyuan/FramePack_F1_I2V_HY_20250503.safetensors --vae hunyuan/pytorch_model.pt --blocks_to_swap 26 --vae_chunk_size 32 --vae_spatial_tile_sample_min_size 128 --preview_latent_every 1 --preview_suffix fpack_1749202618_6997 --use_teacache --teacache_steps 25 --teacache_thresh 0.15
SUBPROCESS: W0606 05:37:05.402000 20664 venv\Lib\site-packages\torch\distributed\elastic\multiprocessing\redirects.py:29] NOTE: Redirects are currently not supported in Windows or MacOs.
SUBPROCESS: Xformers is not installed!
SUBPROCESS: Flash Attn is installed!
SUBPROCESS: Sage Attn is installed!
SUBPROCESS: Trying to import sageattention
SUBPROCESS: Successfully imported sageattention
SUBPROCESS: INFO blissful_tuner.blissful_args | PyTorch: blissful_args.py:66
SUBPROCESS: 2.7.0+cu128, CUDA: 12.8 CC: 8.9 (blissful_prefunc)
SUBPROCESS: INFO:blissful_tuner.blissful_args:PyTorch: 2.7.0+cu128, CUDA: 12.8 CC: 8.9
SUBPROCESS: INFO blissful_tuner.blissful_args | Device: 'NVIDIA blissful_args.py:66
SUBPROCESS: GeForce RTX 4060 Ti', VRAM: '16379MB'
SUBPROCESS: (blissful_prefunc)
SUBPROCESS: INFO:blissful_tuner.blissful_args:Device: 'NVIDIA GeForce RTX 4060 Ti', VRAM: '16379MB'
SUBPROCESS: HunyuanVideoTransformer3DModelPacked: Block swap enabled. Swapping 26 blocks, double blocks: 13, single blocks: 27, supports_backward: False.
SUBPROCESS: TeaCache enabled: num_steps=25, rel_l1_thresh=0.15
SUBPROCESS: 0%| | 0/25 [00:00<?, ?it/s]
SUBPROCESS: 4%|▍ | 1/25 [00:13<05:18, 13.27s/it]
SUBPROCESS: 8%|▊ | 2/25 [00:13<02:07, 5.55s/it]
SUBPROCESS: 12%|█▏ | 3/25 [00:25<03:07, 8.50s/it]
SUBPROCESS: 16%|█▌ | 4/25 [00:25<01:50, 5.28s/it]
SUBPROCESS: 24%|██▍ | 6/25 [00:37<01:47, 5.66s/it]
SUBPROCESS: 28%|██▊ | 7/25 [00:38<01:15, 4.17s/it]
SUBPROCESS: 36%|███▌ | 9/25 [00:49<01:18, 4.91s/it]
SUBPROCESS: 40%|████ | 10/25 [00:49<00:57, 3.80s/it]
SUBPROCESS: 48%|████▊ | 12/25 [01:01<01:00, 4.68s/it]
SUBPROCESS: 52%|█████▏ | 13/25 [01:02<00:44, 3.69s/it]
SUBPROCESS: 60%|██████ | 15/25 [01:14<00:45, 4.56s/it]
SUBPROCESS: 64%|██████▍ | 16/25 [01:14<00:32, 3.62s/it]
SUBPROCESS: 72%|███████▏ | 18/25 [01:26<00:31, 4.55s/it]
SUBPROCESS: 76%|███████▌ | 19/25 [01:26<00:21, 3.62s/it]
SUBPROCESS: 80%|████████ | 20/25 [01:38<00:28, 5.64s/it]
SUBPROCESS: 84%|████████▍ | 21/25 [01:39<00:17, 4.29s/it]
SUBPROCESS: 88%|████████▊ | 22/25 [01:50<00:18, 6.29s/it]
SUBPROCESS: 92%|█████████▏| 23/25 [02:02<00:15, 7.87s/it]
SUBPROCESS: 96%|█████████▌| 24/25 [02:03<00:05, 5.73s/it]
SUBPROCESS: 100%|██████████| 25/25 [02:15<00:00, 7.50s/it]
SUBPROCESS: 100%|██████████| 25/25 [02:15<00:00, 5.41s/it]
SUBPROCESS: INFO blissful_tuner.latent_preview | latent_preview.py:76
SUBPROCESS: LatentPreviewer.write_preview: Input frames
SUBPROCESS: shape: torch.Size([10, 3, 640, 608]), dtype:
SUBPROCESS: torch.float32, device: cuda:0 (write_preview)
SUBPROCESS: INFO:blissful_tuner.latent_preview:LatentPreviewer.write_preview: Input frames shape: torch.Size([10, 3, 640, 608]), dtype: torch.float32, device: cuda:0
SUBPROCESS: INFO blissful_tuner.latent_preview | latent_preview.py:77
SUBPROCESS: LatentPreviewer.write_preview: Target width=608,
SUBPROCESS: height=640, fps=7 (write_preview)
SUBPROCESS: INFO:blissful_tuner.latent_preview:LatentPreviewer.write_preview: Target width=608, height=640, fps=7
SUBPROCESS: INFO blissful_tuner.latent_preview | latent_preview.py:96
SUBPROCESS: LatentPreviewer.write_preview: Attempting to
SUBPROCESS: write MP4 video to
SUBPROCESS: outputs\previews\latent_preview_fpack_1749202618_
SUBPROCESS: 6997.mp4 at 7 FPS with 10 frames. (write_preview)
SUBPROCESS: INFO:blissful_tuner.latent_preview:LatentPreviewer.write_preview: Attempting to write MP4 video to outputs\previews\latent_preview_fpack_1749202618_6997.mp4 at 7 FPS with 10 frames.
SUBPROCESS: INFO blissful_tuner.latent_preview | LatentPreviewer: latent_preview.py:134
DEBUG: Preview file updated: outputs\previews\latent_preview_fpack_1749202618_6997.mp4 (mtime: 1749202846.6902833)
SUBPROCESS: Successfully finished writing preview video:
SUBPROCESS: outputs\previews\latent_preview_fpack_1749202618
SUBPROCESS: 6997.mp4 (write_preview)
SUBPROCESS: INFO:blissful_tuner.latent_preview:LatentPreviewer: Successfully finished writing preview video: outputs\previews\latent_preview_fpack_1749202618_6997.mp4
SUBPROCESS: 0%| | 0/1 [00:00<?, ?it/s]
SUBPROCESS: 100%|██████████| 1/1 [00:17<00:00, 17.40s/it]
SUBPROCESS: 100%|██████████| 1/1 [00:17<00:00, 17.40s/it]
Cleaned up preview file: outputs\previews\latent_preview_fpack_1749202618_6997.mp4
Using base input image for --image_path: C:\Users\jrm19\AppData\Local\Temp\gradio\dbfdcf32c403add8fbde9d89b8d97d87e20f732433ea3c2a5cdb33793903c97a\448432177_10232886237679235_6676698719711904109_n_8192x8192.jpg
Using Target Resolution 640 with UI image aspect. Batch-wide bucket: H=640, W=608
Calculated total sections for UI from duration: 2
Using base input image as the primary for --image_path (item 1): C:\Users\jrm19\AppData\Local\Temp\gradio\dbfdcf32c403add8fbde9d89b8d97d87e20f732433ea3c2a5cdb33793903c97a\448432177_10232886237679235_6676698719711904109_n_8192x8192.jpg
Item 1: Using batch-wide dimensions: H=640, W=608
DEBUG: Enabling latent preview every 1 sections with suffix fpack_1749202956_4574.
Running FramePack Command: C:\Ai\H1111\venv\Scripts\python.exe fpack_generate_video.py --text_encoder1 hunyuan/llava_llama3_fp16.safetensors --text_encoder2 hunyuan/clip_l.safetensors --image_encoder hunyuan/model.safetensors --image_path C:\Users\jrm19\AppData\Local\Temp\gradio\dbfdcf32c403add8fbde9d89b8d97d87e20f732433ea3c2a5cdb33793903c97a\448432177_10232886237679235_6676698719711904109_n_8192x8192.jpg --save_path outputs --prompt girl laughing --video_size 640 608 --video_seconds 2 --infer_steps 25 --seed 1906865157 --embedded_cfg_scale 10 --guidance_scale 1 --guidance_rescale 0 --latent_window_size 9 --sample_solver unipc --output_type video --attn_mode sageattn --is_f1 --dit hunyuan/FramePack_F1_I2V_HY_20250503.safetensors --vae hunyuan/pytorch_model.pt --blocks_to_swap 26 --vae_chunk_size 32 --vae_spatial_tile_sample_min_size 128 --preview_latent_every 1 --preview_suffix fpack_1749202956_4574 --use_teacache --teacache_steps 25 --teacache_thresh 0.15
SUBPROCESS: W0606 05:42:40.898000 8732 venv\Lib\site-packages\torch\distributed\elastic\multiprocessing\redirects.py:29] NOTE: Redirects are currently not supported in Windows or MacOs.
SUBPROCESS: Xformers is not installed!
SUBPROCESS: Flash Attn is installed!
SUBPROCESS: Sage Attn is installed!
SUBPROCESS: Trying to import sageattention
SUBPROCESS: Successfully imported sageattention
SUBPROCESS: INFO blissful_tuner.blissful_args | PyTorch: blissful_args.py:66
SUBPROCESS: 2.7.0+cu128, CUDA: 12.8 CC: 8.9 (blissful_prefunc)
SUBPROCESS: INFO:blissful_tuner.blissful_args:PyTorch: 2.7.0+cu128, CUDA: 12.8 CC: 8.9
SUBPROCESS: INFO blissful_tuner.blissful_args | Device: 'NVIDIA blissful_args.py:66
SUBPROCESS: GeForce RTX 4060 Ti', VRAM: '16379MB'
SUBPROCESS: (blissful_prefunc)
SUBPROCESS: INFO:blissful_tuner.blissful_args:Device: 'NVIDIA GeForce RTX 4060 Ti', VRAM: '16379MB'
SUBPROCESS: HunyuanVideoTransformer3DModelPacked: Block swap enabled. Swapping 26 blocks, double blocks: 13, single blocks: 27, supports_backward: False.
SUBPROCESS: TeaCache enabled: num_steps=25, rel_l1_thresh=0.15
SUBPROCESS: 0%| | 0/25 [00:00<?, ?it/s]
SUBPROCESS: 4%|▍ | 1/25 [00:13<05:12, 13.02s/it]
SUBPROCESS: 8%|▊ | 2/25 [00:13<02:05, 5.45s/it]
SUBPROCESS: 12%|█▏ | 3/25 [00:25<03:05, 8.45s/it]
SUBPROCESS: 16%|█▌ | 4/25 [00:25<01:49, 5.21s/it]
SUBPROCESS: 24%|██▍ | 6/25 [00:37<01:46, 5.58s/it]
SUBPROCESS: 28%|██▊ | 7/25 [00:37<01:14, 4.12s/it]
SUBPROCESS: 36%|███▌ | 9/25 [00:49<01:19, 4.94s/it]
SUBPROCESS: 40%|████ | 10/25 [00:49<00:57, 3.82s/it]
SUBPROCESS: 48%|████▊ | 12/25 [01:01<01:00, 4.68s/it]
SUBPROCESS: 52%|█████▏ | 13/25 [01:01<00:44, 3.69s/it]
SUBPROCESS: 60%|██████ | 15/25 [01:13<00:45, 4.56s/it]
SUBPROCESS: 64%|██████▍ | 16/25 [01:13<00:32, 3.62s/it]
SUBPROCESS: 72%|███████▏ | 18/25 [01:26<00:31, 4.54s/it]
SUBPROCESS: 76%|███████▌ | 19/25 [01:26<00:21, 3.63s/it]
SUBPROCESS: 80%|████████ | 20/25 [01:38<00:27, 5.57s/it]
SUBPROCESS: 84%|████████▍ | 21/25 [01:38<00:16, 4.23s/it]
SUBPROCESS: 88%|████████▊ | 22/25 [01:50<00:18, 6.26s/it]
SUBPROCESS: 92%|█████████▏| 23/25 [02:02<00:15, 7.86s/it]
SUBPROCESS: 96%|█████████▌| 24/25 [02:02<00:05, 5.72s/it]
SUBPROCESS: 100%|██████████| 25/25 [02:14<00:00, 7.50s/it]
SUBPROCESS: 100%|██████████| 25/25 [02:14<00:00, 5.38s/it]
SUBPROCESS: INFO blissful_tuner.latent_preview | latent_preview.py:76
SUBPROCESS: LatentPreviewer.write_preview: Input frames
SUBPROCESS: shape: torch.Size([10, 3, 640, 608]), dtype:
SUBPROCESS: torch.float32, device: cuda:0 (write_preview)
SUBPROCESS: INFO:blissful_tuner.latent_preview:LatentPreviewer.write_preview: Input frames shape: torch.Size([10, 3, 640, 608]), dtype: torch.float32, device: cuda:0
SUBPROCESS: INFO blissful_tuner.latent_preview | latent_preview.py:77
SUBPROCESS: LatentPreviewer.write_preview: Target width=608,
SUBPROCESS: height=640, fps=7 (write_preview)
SUBPROCESS: INFO:blissful_tuner.latent_preview:LatentPreviewer.write_preview: Target width=608, height=640, fps=7
SUBPROCESS: INFO blissful_tuner.latent_preview | latent_preview.py:96
SUBPROCESS: LatentPreviewer.write_preview: Attempting to
SUBPROCESS: write MP4 video to
SUBPROCESS: outputs\previews\latent_preview_fpack_1749202956
SUBPROCESS: 4574.mp4 at 7 FPS with 10 frames. (write_preview)
SUBPROCESS: INFO:blissful_tuner.latent_preview:LatentPreviewer.write_preview: Attempting to write MP4 video to outputs\previews\latent_preview_fpack_1749202956_4574.mp4 at 7 FPS with 10 frames.
SUBPROCESS: INFO blissful_tuner.latent_preview | LatentPreviewer: latent_preview.py:134
DEBUG: Preview file updated: outputs\previews\latent_preview_fpack_1749202956_4574.mp4 (mtime: 1749203181.9928527)
SUBPROCESS: Successfully finished writing preview video:
SUBPROCESS: outputs\previews\latent_preview_fpack_1749202956
SUBPROCESS: 4574.mp4 (write_preview)
SUBPROCESS: INFO:blissful_tuner.latent_preview:LatentPreviewer: Successfully finished writing preview video: outputs\previews\latent_preview_fpack_1749202956_4574.mp4
SUBPROCESS: 0%| | 0/25 [00:00<?, ?it/s]
SUBPROCESS: 4%|▍ | 1/25 [00:12<04:58, 12.43s/it]
SUBPROCESS: 8%|▊ | 2/25 [00:12<02:01, 5.27s/it]
SUBPROCESS: 12%|█▏ | 3/25 [00:24<03:03, 8.36s/it]
SUBPROCESS: 16%|█▌ | 4/25 [00:24<01:48, 5.16s/it]
SUBPROCESS: 24%|██▍ | 6/25 [00:37<01:47, 5.65s/it]
SUBPROCESS: 28%|██▊ | 7/25 [00:37<01:14, 4.16s/it]
SUBPROCESS: 36%|███▌ | 9/25 [00:49<01:19, 4.95s/it]
SUBPROCESS: 40%|████ | 10/25 [00:49<00:57, 3.83s/it]
SUBPROCESS: 48%|████▊ | 12/25 [01:01<01:01, 4.75s/it]
SUBPROCESS: 52%|█████▏ | 13/25 [01:01<00:44, 3.73s/it]
SUBPROCESS: 60%|██████ | 15/25 [01:13<00:45, 4.55s/it]
SUBPROCESS: 64%|██████▍ | 16/25 [01:13<00:32, 3.61s/it]
SUBPROCESS: 72%|███████▏ | 18/25 [01:25<00:31, 4.55s/it]
SUBPROCESS: 76%|███████▌ | 19/25 [01:26<00:21, 3.62s/it]
SUBPROCESS: 80%|████████ | 20/25 [01:38<00:28, 5.60s/it]
SUBPROCESS: 84%|████████▍ | 21/25 [01:38<00:17, 4.26s/it]
SUBPROCESS: 88%|████████▊ | 22/25 [01:50<00:18, 6.25s/it]
SUBPROCESS: 92%|█████████▏| 23/25 [02:02<00:15, 7.79s/it]
SUBPROCESS: 96%|█████████▌| 24/25 [02:02<00:05, 5.67s/it]
SUBPROCESS: 100%|██████████| 25/25 [02:14<00:00, 7.48s/it]
SUBPROCESS: 100%|██████████| 25/25 [02:14<00:00, 5.38s/it]
SUBPROCESS: INFO blissful_tuner.latent_preview | latent_preview.py:76
SUBPROCESS: LatentPreviewer.write_preview: Input frames
SUBPROCESS: shape: torch.Size([19, 3, 640, 608]), dtype:
SUBPROCESS: torch.float32, device: cuda:0 (write_preview)
SUBPROCESS: INFO:blissful_tuner.latent_preview:LatentPreviewer.write_preview: Input frames shape: torch.Size([19, 3, 640, 608]), dtype: torch.float32, device: cuda:0
SUBPROCESS: INFO blissful_tuner.latent_preview | latent_preview.py:77
SUBPROCESS: LatentPreviewer.write_preview: Target width=608,
SUBPROCESS: height=640, fps=7 (write_preview)
SUBPROCESS: INFO:blissful_tuner.latent_preview:LatentPreviewer.write_preview: Target width=608, height=640, fps=7
SUBPROCESS: INFO blissful_tuner.latent_preview | latent_preview.py:96
SUBPROCESS: LatentPreviewer.write_preview: Attempting to
SUBPROCESS: write MP4 video to
DEBUG: Preview file updated: outputs\previews\latent_preview_fpack_1749202956_4574.mp4 (mtime: 1749203316.9358907)
SUBPROCESS: outputs\previews\latent_preview_fpack_1749202956
SUBPROCESS: 4574.mp4 at 7 FPS with 19 frames. (write_preview)
SUBPROCESS: INFO:blissful_tuner.latent_preview:LatentPreviewer.write_preview: Attempting to write MP4 video to outputs\previews\latent_preview_fpack_1749202956_4574.mp4 at 7 FPS with 19 frames.
SUBPROCESS: INFO blissful_tuner.latent_preview | LatentPreviewer: latent_preview.py:134
SUBPROCESS: Successfully finished writing preview video:
SUBPROCESS: outputs\previews\latent_preview_fpack_1749202956
SUBPROCESS: _4574.mp4 (write_preview)
SUBPROCESS: INFO:blissful_tuner.latent_preview:LatentPreviewer: Successfully finished writing preview video: outputs\previews\latent_preview_fpack_1749202956_4574.mp4
SUBPROCESS: 0%| | 0/2 [00:00<?, ?it/s]
SUBPROCESS: 50%|█████ | 1/2 [00:15<00:15, 15.67s/it]
SUBPROCESS: 100%|██████████| 2/2 [00:49<00:00, 26.11s/it]
SUBPROCESS: 100%|██████████| 2/2 [00:49<00:00, 24.54s/it]
Cleaned up preview file: outputs\previews\latent_preview_fpack_1749202956_4574.mp4

[maybleMyers]

Can you try updating with git pull? something is going on with your installation because you should have alot of INFO:main: prints ie:

SUBPROCESS: INFO:main:Enable swap 26 blocks to CPU from device: cuda
SUBPROCESS: INFO:main:Initializing TeaCache: steps=25, threshold=0.15
SUBPROCESS: INFO:main:Calculated total_latent_sections from video_seconds: 2 sections.
SUBPROCESS: INFO:main:Mode: F1, Video size: 640x608@2.00s, fps: 30, num sections: 2, infer_steps: 25, frames per generation step: 33
SUBPROCESS: INFO:main:Using compute dtype: torch.bfloat16
SUBPROCESS: INFO:main:Starting F1 model sampling process.
SUBPROCESS: INFO:main:F1 Mode: Using video dimensions 640x608 for latent operations and generation.
SUBPROCESS: INFO:main:--- F1 Section 1 / 2 ---
Detected/Updated actual total sections: 2
SUBPROCESS: HunyuanVideoTransformer3DModelPacked: Block swap enabled. Swapping 26 blocks, double blocks: 13, single blocks: 27, supports_backward: False.

The ui looks for the final video's location via this print:
SUBPROCESS: INFO:main:Decoded. Pixel shape torch.Size([1, 3, 397, 928, 704])
SUBPROCESS: INFO:main:Video saved to: outputs/20250606-050917_319807433_100.mp4
SUBPROCESS: INFO:main:Done!
Added metadata to outputs/20250606-050917_319807433_100.mp4

which is missing from your log for some reason.

[Redtash1]

I git pull and it updated a couple files but that didn't work. I deleted everything but the hunyuan, wan, sage, flash, and couple .bat files I made. Recloned H1111 again, created venv and installed torch and requirementsTorch27.txt and tried it again and it worked. Then I installed Flash Attention and it worked. Then I installed Triton and it worked. Then I installed Sage Attention and it worked. I was watch a guy on YouTube and when he installed Framepack F1 he also installed Torchao because he said it supported to give a little speed boost and I installed on Regular Framepack F1 no problem. But when I installed it in H1111 it causes the no preview of generated video and the other problems you saw. I Uninstalled Torchao and it's working fine again. 1 thing I did notice when installing everything again is that if I try to use requirementsTorch27.txt to install Torch==2.7.0+cu128 is throws an error that it can't find it. So I had to install it first?

[maybleMyers]

Glad you got it fixed xD...

Yes it is always best to install your version of torch, torchvision first ie:
pip install torch==2.7.0+cu128 torchvision==0.22.0+cu128 --index-url https://download.pytorch.org/whl/cu128
Then install requirements, because pip's dependency resolver works best with torch already installed.

[Redtash1]

Ok but in requirementsTorch27.txt, Torch, Torchvision and Torchaudio is at the top, so I thought they could be installed that way. I don't know if you saw my comment in Discussions but I wondered if you had a Discord channel for people to interact and talk about H1111?