Mirage Pipeline#1

Open

DavidBert wants to merge 70 commits intomainfrom

Commits on Sep 25, 2025

Support both huggingface_hub `v0.x` and `v1.x` (#12389)
Wauplin
authored

Commits on Sep 26, 2025

mirage pipeline first commit
David Bertoin
committed
use attention processors
David Bertoin
committed
use diffusers rmsnorm
David Bertoin
committed
use diffusers timestep embedding method
David Bertoin
committed
[CI] disable installing transformers from main in ci for now. (#12397)
sayakpaul
authored
remove MirageParams
David Bertoin
committed
checkpoint conversion script
David Bertoin
committed
ruff formating
David Bertoin
committed
[docs] slight edits to the attention backends docs. (#12394)

sayakpaul
and
stevhliu
authored
[docs] remove docstrings from repeated methods in `lora_pipeline.py` (#12393)
sayakpaul
authored

Commits on Sep 29, 2025

Commits on Sep 30, 2025

[Tests] Add single file tester mixin for Models and remove unittest dependency (#12352)
DN6
authored
fix 3 xpu failures uts w/ latest pytorch (#12408)
yao-matrix
authored
Install latest prerelease from huggingface_hub when installing transformers from main (#12395)
Wauplin
authored
[docs] CP (#12331)
stevhliu
authored
[docs] Migrate syntax (#12390)
stevhliu
authored
remove dependencies to old checkpoints
David Bertoin
committed
remove old checkpoints dependency
David Bertoin
committed
move default height and width in checkpoint config
David Bertoin
committed
add docstrings
David Bertoin
committed
if conditions and raised as ValueError instead of asserts
David Bertoin
committed
small fix
David Bertoin
committed
nit remove try block at import
David Bertoin
committed
mirage pipeline doc
David Bertoin
committed

Commits on Oct 1, 2025

[tests] cache non lora pipeline outputs. (#12298)
sayakpaul
authored

Commits on Oct 2, 2025

Commits on Oct 3, 2025

[training-scripts] Make more examples UV-compatible (follow up on #12000) (#12407)

linoytsaban
and
sayakpaul
authored

Commits on Oct 5, 2025

Commits on Oct 7, 2025

handle offload_state_dict when initing transformers models (#12438)
sayakpaul
authored
update doc
David Bertoin
committed
rename model to photon
David Bertoin
committed
[Qwen LoRA training] fix bug when offloading (#12440)
linoytsaban
authored

Commits on Oct 8, 2025

Align Flux modular more and more with Qwen modular (#12445)
sayakpaul
authored
fix dockerfile definitions. (#12424)
sayakpaul
authored
fix more torch.distributed imports (#12425)
sayakpaul
authored
mirage pipeline first commit
David Bertoin
committed
use attention processors
David Bertoin
committed
use diffusers rmsnorm
David Bertoin
committed
use diffusers timestep embedding method
David Bertoin
committed
remove MirageParams
David Bertoin
committed
checkpoint conversion script
David Bertoin
committed
ruff formating
David Bertoin
committed
remove dependencies to old checkpoints
David Bertoin
committed
remove old checkpoints dependency
David Bertoin
committed
move default height and width in checkpoint config
David Bertoin
committed
add docstrings
David Bertoin
committed
if conditions and raised as ValueError instead of asserts
David Bertoin
committed
small fix
David Bertoin
committed
nit remove try block at import
David Bertoin
committed
mirage pipeline doc
David Bertoin
committed
update doc
David Bertoin
committed
rename model to photon
David Bertoin
committed
add text tower and vae in checkpoint
David Bertoin
committed
update doc
David Bertoin
committed
Merge branch 'mirage' of https://github.com/Photoroom/diffusers into mirage
David Bertoin
committed
update photon doc
David Bertoin
committed
ruff fixes
David Bertoin
committed