Skip to content

qwix for WAN 2.1 FP8 recipe #293

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
susanbao wants to merge 3 commits into from
Closed

qwix for WAN 2.1 FP8 recipe #293

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
a83bb1c
qwix
Dec 11, 2025
6aee280
add qwix version
susanbao Dec 12, 2025
22a87b8
fix unit test
susanbao Dec 12, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion requirements.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -36,4 +36,4 @@ aqtp
imageio==2.37.0
imageio-ffmpeg==0.6.0
hf_transfer>=0.1.9
qwix@git+https://github.com/google/qwix.git
qwix==0.1.5
2 changes: 1 addition & 1 deletion requirements_with_jax_ai_image.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -37,4 +37,4 @@ aqtp
imageio==2.37.0
imageio-ffmpeg==0.6.0
hf_transfer>=0.1.9
qwix@git+https://github.com/google/qwix.git
qwix==0.1.5
8 changes: 4 additions & 4 deletions src/maxdiffusion/pipelines/wan/wan_pipeline.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -302,8 +302,8 @@ def get_fp8_config(cls, config: HyperParameters):
act_qtype=jnp.float8_e4m3fn,
bwd_qtype=jnp.float8_e5m2,
disable_channelwise_axes=True, # per_tensor calibration
weight_calibration_method=config.quantization_calibration_method,
act_calibration_method=config.quantization_calibration_method,
weight_calibration_method="fixed,-224,224",
act_calibration_method="fixed,-224,224",
bwd_calibration_method=config.quantization_calibration_method,
op_names=("dot_general", "einsum"),
),
Expand All @@ -313,8 +313,8 @@ def get_fp8_config(cls, config: HyperParameters):
act_qtype=jnp.float8_e4m3fn,
bwd_qtype=jnp.float8_e4m3fn,
disable_channelwise_axes=True, # per_tensor calibration
weight_calibration_method=config.quantization_calibration_method,
act_calibration_method=config.quantization_calibration_method,
weight_calibration_method="fixed,-224,224",
act_calibration_method="fixed,-224,224",
bwd_calibration_method=config.quantization_calibration_method,
op_names=("conv_general_dilated"),
),
Expand Down
8 changes: 4 additions & 4 deletions src/maxdiffusion/tests/wan_transformer_test.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -343,8 +343,8 @@ def create_real_rule_instance(*args, **kwargs):
act_qtype=jnp.float8_e4m3fn,
bwd_qtype=jnp.float8_e5m2,
disable_channelwise_axes=True, # per_tensor calibration
weight_calibration_method=config_fp8_full.quantization_calibration_method,
act_calibration_method=config_fp8_full.quantization_calibration_method,
weight_calibration_method="fixed,-224,224",
act_calibration_method="fixed,-224,224",
bwd_calibration_method=config_fp8_full.quantization_calibration_method,
op_names=("dot_general", "einsum"),
),
Expand All @@ -354,8 +354,8 @@ def create_real_rule_instance(*args, **kwargs):
act_qtype=jnp.float8_e4m3fn,
bwd_qtype=jnp.float8_e4m3fn,
disable_channelwise_axes=True, # per_tensor calibration
weight_calibration_method=config_fp8_full.quantization_calibration_method,
act_calibration_method=config_fp8_full.quantization_calibration_method,
weight_calibration_method="fixed,-224,224",
act_calibration_method="fixed,-224,224",
bwd_calibration_method=config_fp8_full.quantization_calibration_method,
op_names=("conv_general_dilated"),
),
Expand Down
Loading