Cosmos 3 FP8: wrapping pipe.transformer (e.g. torch.compile) silently disables the checkpoint's W8A16 step policy
Les mainteneurs répondent en général sous 1 jour
Personne n'a encore pris cette issue.
Évaluation
- Difficulté
- 4/5
- Temps estimé
- 3-5 jours
- Accessibilité débutants
- 42/100
- Type d'issue
- Bug
- Clarté
- Plutôt claire
- Activité
- Active
- Domaine
- ai-infra-agents
Piste de recherche
Start in src/diffusers/pipelines/cosmos/mixed_precision.py: the loader lookup getattr(type(module), "load_config", None) in _on_disk_quantization_config() is the cause, and resolve() falls back to format="none" when it fails. Try the suggested fix of looking up load_config on the instance, or unwrapping with unwrap_module in src/diffusers/utils/torch_utils.py. Done when Cosmos3MixedPrecisionConfig.resolve(torch.compile(model)) equals resolve(model) on a small model whose config declares the ModelOpt FP8 policy, with a regression test added.
Rédigé par le modèle d'indexation à partir du texte de l'issue.
Description
Describe the bug
For the ModelOpt FP8 Cosmos 3 checkpoints, the per-step W8A16/W8A8 policy (quantization_config.runtime.diffusion_step_policy in transformer/config.json, added in #14664) is resolved from self.transformer on every pipeline call:
Cosmos3MixedPrecisionConfig.resolve() → quantization_config_from_module() → _on_disk_quantization_config(), which looks up the config loader on the module's class:
loader = getattr(type(module), "load_config", None)
The in-memory config does not carry quantization_config.runtime (see the docstring of quantization_config_from_module, L409-L426), so this on-disk read is the only source of the policy. When pipe.transformer is a wrapper, type(module) is the wrapper class, which has no load_config. The on-disk config is never read, _parsed_checkpoint_policy() returns None, and resolve() returns format="none" (L157-L158). Every step then runs W8A8 instead of the checkpoint's W8A16 first/last steps, with no warning. The Cosmos 3 docs note that all-W8A8 multi-step video can flicker (cosmos3.md L50-L54).
pipe.transformer = torch.compile(pipe.transformer), the pattern shown in several docs pages (e.g. optimization/cache_dit.md L260, quantization/autoround.md L102), is such a wrapper (OptimizedModule). In-place pipe.transformer.compile() and compile_repeated_blocks() (used on the Cosmos 3 page) keep the class and are not affected. The line is unchanged on main: mixed_precision.py#L450.
Reproduction
The policy is printed with the same resolve() call the pipeline makes, before any generation:
import modelopt.torch.quantization.backends.fp8_per_tensor_gemm # noqa: F401 - registers the FP8 GEMM
import torch
from diffusers import Cosmos3OmniPipeline
from diffusers.pipelines.cosmos.mixed_precision import Cosmos3MixedPrecisionConfig
from modelopt.torch.opt import enable_huggingface_checkpointing
enable_huggingface_checkpointing()
pipe = Cosmos3OmniPipeline.from_pretrained("nvidia/Cosmos3-Nano", revision="fp8", dtype=torch.bfloat16)
pipe.vae.to("cuda")
pipe.transformer.enable_group_offload(
onload_device=torch.device("cuda"), offload_device=torch.device("cpu"),
offload_type="block_level", num_blocks_per_group=1, use_stream=True,
)
pipe.transformer = torch.compile(pipe.transformer, backend="eager") # the only difference between the two runs
print(f"type(pipe.transformer) = {type(pipe.transformer).__name__}")
print(f"resolved policy = {Cosmos3MixedPrecisionConfig.resolve(pipe.transformer)}")
Logs
Without the torch.compile line:
type(pipe.transformer) = Cosmos3OmniTransformer
resolved policy = Cosmos3MixedPrecisionConfig(format='fp8', first_steps=3, last_steps=3, reasoner_policy='high_precision', overlap='a16')
With it:
type(pipe.transformer) = OptimizedModule
resolved policy = Cosmos3MixedPrecisionConfig(format='none', first_steps=3, last_steps=3, reasoner_policy='high_precision', overlap='a16')
Effect on generation (T2V 848×480, 49 frames, 8 steps, seed 123, transformer with enable_group_offload(offload_type="block_level", num_blocks_per_group=1, use_stream=True) to fit a 16 GB GPU):
pipe.transformer |
call | resolved policy | mixed-precision runtime installed on the transformer |
|---|---|---|---|
Cosmos3OmniTransformer |
default | fp8, first/last 3 steps W8A16 |
yes (252 reasoner + 252 generation linears) |
Cosmos3OmniTransformer |
mixed_precision_format="none" |
none |
no; different video from the default run |
OptimizedModule (torch.compile(pipe.transformer, backend="eager")) |
default | none |
— (see below) |
With the compiled ModelOpt FP8 transformer, generation currently stops earlier inside Dynamo, so the end-to-end output of the compiled run cannot be shown; the resolved policy above is what the pipeline applies to it. Any wrapper that forwards attribute access and does run reaches the none schedule.
Dynamo traceback of the compiled run (separate issue, context only)
Traceback (most recent call last):
File "repro.py", line 73, in <module>
video = pipe(
^^^^^
File "/usr/local/lib/python3.12/dist-packages/torch/utils/_contextlib.py", line 120, in decorate_context
return func(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.12/dist-packages/diffusers/pipelines/cosmos/pipeline_cosmos3_omni.py", line 1756, in __call__
preds_vision, preds_sound, preds_action = self.transformer(
^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.12/dist-packages/torch/_dynamo/eval_frame.py", line 414, in __call__
return super().__call__(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.12/dist-packages/torch/nn/modules/module.py", line 1775, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.12/dist-packages/torch/nn/modules/module.py", line 1786, in _call_impl
return forward_call(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.12/dist-packages/torch/_dynamo/eval_frame.py", line 832, in compile_wrapper
return fn(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.12/dist-packages/torch/nn/modules/module.py", line 1775, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.12/dist-packages/torch/nn/modules/module.py", line 1786, in _call_impl
return forward_call(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.12/dist-packages/diffusers/hooks/hooks.py", line 217, in new_forward
output = function_reference.forward(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.12/dist-packages/diffusers/hooks/hooks.py", line 216, in new_forward
args, kwargs = function_reference.pre_forward(module, *args, **kwargs)
[... torch/_dynamo frames ...]
File "/usr/local/lib/python3.12/dist-packages/modelopt/torch/opt/dynamic.py", line 81, in __len__
return len(self._callbacks)
^^^^^^^^^^^^^^^
torch._dynamo.exc.InternalTorchDynamoError: AttributeError: '_FoldedCallback' object has no attribute '_callbacks'
from user code:
File "/usr/local/lib/python3.12/dist-packages/diffusers/hooks/hooks.py", line 217, in torch_dynamo_resume_in_new_forward_at_216
output = function_reference.forward(*args, **kwargs)
File "/usr/local/lib/python3.12/dist-packages/diffusers/models/transformers/transformer_cosmos3.py", line 745, in forward
packed_tokens_vision = self.proj_in(packed_tokens_vision)
File "/usr/local/lib/python3.12/dist-packages/modelopt/torch/quantization/nn/modules/quant_linear.py", line 240, in forward
return super().forward(input, *args, **kwargs)
File "/usr/local/lib/python3.12/dist-packages/modelopt/torch/quantization/nn/modules/quant_module.py", line 240, in forward
return super().forward(input, *args, **kwargs)
File "/usr/local/lib/python3.12/dist-packages/modelopt/torch/quantization/nn/modules/quant_module.py", line 192, in forward
output = super().forward(input, *args, **kwargs)
File "/usr/local/lib/python3.12/dist-packages/torch/nn/modules/linear.py", line 134, in forward
return F.linear(input, self.weight, self.bias)
File "/usr/local/lib/python3.12/dist-packages/modelopt/torch/opt/dynamic.py", line 782, in __getattr__
return manager.get_da_cb(name)(self, value)
Suggested fix
Either of these, plus a regression test (resolve(torch.compile(model)) == resolve(model) on a small model whose config declares the ModelOpt FP8 policy):
- Look the loader up on the instance:
loader = getattr(module, "load_config", None).load_configis aConfigMixinclassmethod, so this returns the same bound classmethod, and wrappers that forward attribute access (OptimizedModule.__getattr__→_orig_mod) work too. - Or unwrap first with the existing helper
unwrap_module, as other Diffusers code does for compiled modules. This coverstorch.compileonly.
Optionally, a warning when a ModelOpt FP8 backend is detected but the on-disk policy could not be read, so the fallback to all-W8A8 is visible. Happy to open a PR.
System Info
- diffusers 0.41.0, nvidia-modelopt 0.44.0, transformers 5.20.2, accelerate 1.15.0
- torch 2.9.0a0+145a3a7 (container
nvcr.io/nvidia/pytorch:25.10-py3), CUDA 13.0, Python 3.12 - NVIDIA GeForce RTX 5080 16 GB, driver 620.62
- Langage dominant
- Python
- Étoiles
- 34.9k
- Forks
- 7.5k
- Merge moyen
- 3 j 16 h
- PR mergées (30 j)
- 65
Préparer son environnement
- Aucun Dockerfile ni fichier Docker Compose
- Propose un modèle de pull request
- Lire le guide de contribution
Par où commencer
- Lisez l'issue en entier, puis le guide de contribution du projet.
- Signalez en commentaire que vous la prenez — cela évite que deux personnes fassent le même travail.
- Forkez le dépôt et travaillez sur une branche.
- Ouvrez une pull request qui référence le numéro de l'issue.
Autres issues de huggingface/diffusers
-
Difficulté 2/5 1-3 heures Accessibilité débutants 84/100
huggingface/diffusers#15363 ·
Les mainteneurs répondent en général sous 1 jour
-
`ZImageImg2ImgPipeline` / `ZImageInpaintPipeline`: explicit height/width different from the input image raises RuntimeErrorPeut-être pris @chessing234 l’a pris il y a 1 jour. Ouvertebug
Difficulté 2/5 1-3 heures Accessibilité débutants 70/100
huggingface/diffusers#15337 · 2 commentaires ·
Les mainteneurs répondent en général sous 1 jour
-
Multi-vector textual inversion uses only the first vector when the token has uppercase letters (transformers v5)Peut-être pris @sunhouy l’a pris il y a 2 jours. Ouverte
Difficulté 2/5 1-3 heures Accessibilité débutants 75/100
huggingface/diffusers#15313 ·
Les mainteneurs répondent en général sous 1 jour
-
TCDScheduler advances the generator on the final step of a strength-truncated custom schedulePeut-être pris @SIDDARTHAREDDY8 l’a pris il y a 5 jours. Ouvertebug
Difficulté 2/5 1-3 heures Accessibilité débutants 74/100
huggingface/diffusers#15288 · 2 commentaires ·
Les mainteneurs répondent en général sous 1 jour
-
unload_textual_inversion(token) leaves the extra tokens of a multi-vector embedding behind, so it can't be loaded againPeut-être pris @sohailshk l’a pris il y a 11 jours. Ouverte
Difficulté 2/5 1-3 heures Accessibilité débutants 72/100
huggingface/diffusers#15211 · 3 commentaires ·
Les mainteneurs répondent en général sous 1 jour
Toutes les issues de huggingface/diffusers
Issues similaires
-
Difficulté 2/5 1-3 heures Accessibilité débutants 73/100
Les mainteneurs répondent en général sous 1 jour
-
Unused loot bag group names left in dreamcraft en_US.langPeut-être pris Une pull request liée à cette issue est ouverte ou déjà fusionnée. OuverteStatus: Triage Suggestion
Difficulté 1/5 Moins d'une heure Accessibilité débutants 90/100
GTNewHorizons/GT-New-Horizons-Modpack#27789 ·
Les mainteneurs répondent en général sous 1 jour
-
bug frontend sev-5
Difficulté 2/5 1-3 heures Accessibilité débutants 83/100
-
core good first issue
Difficulté 2/5 1-3 heures Accessibilité débutants 65/100
-
b4b code health next
Difficulté 2/5 1-3 heures Accessibilité débutants 68/100
ESCOMP/CTSM#4330 · 1 commentaire ·
Les mainteneurs répondent en général sous 1 jour