No save needed

Fix example

Start & end frame mode

Better allocation estimation

Image at any position

30 steps

Optimize GPU time

Improve examples

Video extension optimization

Merge code

, progress = None

Last frame

Generation time

Improve time

Mistake

Display resolution and save preferences

New features

Re-commit the right code

Revert

Timed prompt

Links

Adapt the code

Do not cache the examples

free_mem_gb = 0

high_vram = False

from diffusers_helper.memory import cpu, gpu, get_cuda_free_memory_gb, move_model_to_device_with_memory_preservation, offload_model_from_device_for_memory_preservation, fake_diffusers_current_device, DynamicSwapInstaller, unload_complete_models, load_model_as_complete

handle_timed_prompt_change

Timed prompts

Use TeaCache', value=False

Import time

Other time

import time

Compute time

Add check

Upload 4 files

Upload 4 files

Finish merge

Merge examples

Merge interface

use_teacache=True

prompt = gr.Textbox(label="Prompt", value='')

Upload 2 files

Upload 5 files

Upload 4 files

80 -> 60

Merge code

block.launch(mcp_server=False, ssr_mode=False)

Upload 5 files

Upload 4 files

gr.Markdown('# Framepack F1 with Image Input or with Video Input (Video Extension)')