Implements per-guide attention attenuation via log-space additive bias
in self-attention. Each guide reference tracks its own strength and
optional spatial mask in conditioning metadata (guide_attention_entries).
* utils: dont use comfy sft loader in aimdo fallback
This was going to the raw command line switch and should respect main.py
probe of whether aimdo actually loaded successfully.
* ops: dont use deferred linear load in Aimdo fallback
Avoid changes of behaviour on --fast dynamic_vram when aimdo doesnt work.
* mp: attach re-construction arguments to model patcher
When making a model-patcher from a unet or ckpt, attach a callable
function that can be called to replay the model construction. This
can be used to deep clone model patcher WRT the actual model.
Originally written by Kosinkadink
f4b99bc623
* mp: Add disable_dynamic clone argument
Add a clone argument that lets a caller clone a ModelPatcher but disable
dynamic to demote the clone to regular MP. This is useful for legacy
features where dynamic_vram support is missing or TBD.
* torch_compile: disable dynamic_vram
This is a bigger feature. Disable for the interim to preserve
functionality.
Integrate comfy-aimdo 0.2 which takes a different approach to
installing the memory allocator hook. Instead of using the complicated
and buggy pytorch MemPool+CudaPluggableAlloctor, cuda is directly hooked
making the process much more transparent to both comfy and pytorch. As
far as pytorch knows, aimdo doesnt exist anymore, and just operates
behind the scenes.
Remove all the mempool setup stuff for dynamic_vram and bump the
comfy-aimdo version. Remove the allocator object from memory_management
and demote its use as an enablment check to a boolean flag.
Comfy-aimdo 0.2 also support the pytorch cuda async allocator, so
remove the dynamic_vram based force disablement of cuda_malloc and
just go back to the old settings of allocators based on command line
input.
This check was far too broad and the dtype is not a reliable indicator
of wanting the requant (as QT returns the compute dtype as the dtype).
So explictly plumb whether fp8mm wants the requant or not.
* lora: add weight shape calculations.
This lets the loader know if a lora will change the shape of a weight
so it can take appropriate action.
* MPDynamic: force load flux img_in weight
This weight is a bit special, in that the lora changes its geometry.
This is rather unique, not handled by existing estimate and doesn't
work for either offloading or dynamic_vram.
Fix for dynamic_vram as a special case. Ideally we can fully precalculate
these lora geometry changes at load time, but just get these models
working first.
Get rid of the cat and unary negation and inplace add-cmul the two
halves of the rope. Precompute -sin once at the start of the model
rather than every transformer block.
This is slightly faster on both GPU and CPU bound setups.
The current behaviour of the default ModelPatcher is to .to a model
only if its fully loaded, which is how random non-leaf weights get
loaded in non-LowVRAM conditions.
The however means they never get loaded in dynamic_vram. In the
dynamic_vram case, force load them to the GPU.
* model_management: lazy-cache aimdo_tensor
These tensors cosntructed from aimdo-allocations are CPU expensive to
make on the pytorch side. Add a cache version that will be valid with
signature match to fast path past whatever torch is doing.
* dynamic_vram: Minimize fast path CPU work
Move as much as possible inside the not resident if block and cache
the formed weight and bias rather than the flat intermediates. In
extreme layer weight rates this adds up.
* Fix bypass dtype/device moving
* Force offloading mode for training
* training context var
* offloading implementation in training node
* fix wrong input type
* Support bypass load lora model, correct adapter/offloading handling
This was missing the stochastic rounding required for fp8 downcast
to be consistent with model_patcher.patch_weight_to_device.
Missed in testing as I spend too much time with quantized tensors
and overlooked the simpler ones.
If there are non-trivial python objects nested in the model_options, this
causes all sorts of issues. Traverse lists and dicts so clones can safely
overide settings and BYO objects but stop there on the deepclone.
* revert threaded model loader change
This change was only needed to get around the pytorch 2.7 mempool bugs,
and should have been reverted along with #12260. This fixes a different
memory leak where pytorch gets confused about cache emptying.
* load non comfy weights
* MPDynamic: Pre-generate the tensors for vbars
Apparently this is an expensive operation that slows down things.
* bump to aimdo 1.8
New features:
watermark limit feature
logging enhancements
-O2 build on linux
Torch has alignment enforcement when viewing with data type changes
but only relative to itself. Do all tensor constructions straight
off the memory-view individually so pytorch doesnt see an alignment
problem.
The is needed for handling misaligned safetensors weights, which are
reasonably common in third party models.
This limits usage of this safetensors loader to GPU compute only
as CPUs kernnel are very likely to bus error. But it works for
dynamic_vram, where we really dont want to take a deep copy and we
always use GPU copy_ which disentangles the misalignment.
This is using a different layers weight with .to(). Change it to use
the ops caster if the original layer is a comfy weight so that it picks
up dynamic_vram and async_offload functionality in full.
Co-authored-by: Rattus <rattus128@gmail.com>
* mp: fix full dynamic unloading
This was not unloading dynamic models when requesting a full unload via
the unpatch() code path.
This was ok, i your workflow was all dynamic models but fails with big
VRAM leaks if you need to fully unload something for a regular ModelPatcher
It also fices the "unload models" button.
* mm: load models outside of Aimdo Mempool
In dynamic_vram mode, escape the Aimdo mempool and load into the regular
mempool. Use a dummy thread to do it.
This function has a dtype argument that allows the caller to set the
dtype in the cast. TIL Some models override this on weight casts, which
means its the highest priority.
Priority scheme is: argument > model dtype > state dict dtype
pinned memory was converted back to pinning the CPU side weight without
any changes. Fix the pinner to use the CPU weight and not the model defined
geometry. This will either save RAM or stop buffer overruns when the types
mismatch.
Fix the model defined weight caster to use the [ s.weight, s.bias ]
interpretation, as xfer_dest might be the flattened pin now. Fix the detection
of needing to cast to not be conditional on !pin.
When a node is declared as dev-only, it doesn't show in the default UI
unless the dev mode is enabled in the settings. The intention is to
allow nodes related to unit testing to be included in ComfyUI
distributions without confusing the average user.
The code throughout is None safe to just skip the feature cache saving
step if none. Set it none in single frame use so qwen doesn't burn VRAM
on the unused cache.
* ops: introduce autopad for conv3d
This works around pytorch missing ability to causal pad as part of the
kernel and avoids massive weight duplications for padding.
* wan-vae: rework causal padding
This currently uses F.pad which takes a full deep copy and is liable to
be the VRAM peak. Instead, kick spatial padding back to the op and
consolidate the temporal padding with the cat for the cache.
* wan-vae: implement zero pad fast path
The WAN VAE is also QWEN where it is used single-image. These
convolutions are however zero padded 3d convolutions, which means the
VAE is actually just 2D down the last element of the conv weight in
the temporal dimension. Fast path this, to avoid adding zeros that
then just evaporate in convoluton math but cost computation.
* Disable timestep embed compression when inpainting
Spatial inpainting not compatible with the compression
* Reduce crossattn peak VRAM
* LTX2: Refactor forward function for better VRAM efficiency
* causal_video_ae: Remove attention ResNet
This attention_head_dim argument does not exist on this constructor so
this is dead code. Remove as generic attention mid VAE conflicts with
temporal roll.
* ltx-vae: consoldate causal/non-causal code paths
* ltx-vae: add cache rolling adder
* ltx-vae: use cached adder for resnet
* ltx-vae: Implement rolling VAE
Implement a temporal rolling VAE for the LTX2 VAE.
Usually when doing temporal rolling VAEs you can just chunk on time relying
on causality and cache behind you as you go. The LTX VAE is however
non-causal.
So go whole hog and implement per layer run ahead and backpressure between
the decoder layers using recursive state beween the layers.
Operations are ammended with temporal_cache_state{} which they can use to
hold any state then need for partial execution. Convolutions cache their
inputs behind the up to N-1 frames, and skip connections need to cache the
mismatch between convolution input and output that happens due to missing
future (non-causal) input.
Each call to run_up() processes a layer accross a range on input that
may or may not be complete. It goes depth first to process as much as
possible to try and digest frames to the final output ASAP. If layers run
out of input due to convolution losses, they simply return without action
effectively applying back-pressure to the earlier layers. As the earlier
layers do more work and caller deeper, the partial states are reconciled
and output continues to digest depth first as much as possible.
Chunking is done using a size quota rather than a fixed frame length and
any layer can initiate chunking, and multiple layers can chunk at different
granulatiries. This remove the old limitation of always having to process
1 latent frame to entirety and having to hold 8 full decoded frames as
the VRAM peak.
* re-init
* Update model_multitalk.py
* whitespace...
* Update model_multitalk.py
* remove print
* this is redundant
* remove import
* Restore preview functionality
* Move block_idx to transformer_options
* Remove LoopingSamplerCustomAdvanced
* Remove looping functionality, keep extension functionality
* Update model_multitalk.py
* Handle ref_attn_mask with separate patch to avoid having to always return q and k from self_attn
* Chunk attention map calculation for multiple speakers to reduce peak VRAM usage
* Update model_multitalk.py
* Add ModelPatch type back
* Fix for latest upstream
* Use DynamicCombo for cleaner node
Basically just so that single_speaker mode hides mask inputs and 2nd audio input
* Update nodes_wan.py
For LTX Audio VAE, remove normalization of audio during MEL spectrogram creation.
This aligs inference with training and prevents loud audio from being attenuated.
* Brought over minimal elements from PR 10045 to reproduce seed_assets and register_assets_system without adding anything to the DB or server routes yet, for now making everything sync (can introduce async once everything is cleaned up and brought over)
* Added db script to insert assets stuff, cleaned up some code; assets (models) now get added/rescanned
* Added support for 5 http endpoints for assets
* Replaced Optional with | None in schemas_in.py and schemas_out.py
* Remove two routes that will not be relevant yet in this PR: HEAD /api/assets/hash/<hash> and PUT /api/assets/<id>/preview
* Remove some functions the two deleted endpoints were using
* Don't show assets scan message upon calling /object_info endpoint
* removed unsued import to satisfy ruff
* Simplified hashing function tpye hint and _hash_file_obj
* Satisfied ruff
This logic was checking comfy_cast_weights, and going straight to
to the forward_comfy_cast_weights implementation without
attempting to downscale input to fp8 in the event comfy_cast_weights
is set.
The main reason comfy_cast_weights would be set would be for async
offload, which is not a good reason to nix FP8MM.
So instead, and together the underlying exclusions for FP8MM which
are:
* having a weight_function (usually LowVramPatch)
* force_cast_weights (compute dtype override)
* the weight is not Quantized
* the input is already quantized
* the model or layer has MM explictily disabled.
If you get past all of those exclusions, quantize the input tensor.
Then hand the new input, quantized or not off to
forward_comfy_cast_weights to handle it. If the weight is offloaded
but input is quantized you will get an offloaded MM8.
If the loader passes 1e32 as the usable memory size, it means force
the full load. This happens with CPU loads and a few other misc cases.
Removing the confusing number and just leave the other details.
* Add support for sage attention 3 in comfyui, enable via new cli arg
--use-sage-attiention3
* Fix some bugs found in PR review. The N dimension at which Sage
Attention 3 takes effect is reduced to 1024 (although the improvement is
not significant at this scale).
* Remove the Sage Attention3 switch, but retain the attention function
registration.
* Fix a ruff check issue in attention.py
Pretty much every error cudaHostRegister can throw also queues the same
error on the async GPU queue. This was fixed for repinning error case,
but there is the bad mmap and just enomem cases that are harder to
detect.
Do some dummy GPU work to clean the error state.
index_timestep_zero can be selected in the
FluxKontextMultiReferenceLatentMethod now with the display name set to the
more generic "Edit Model Reference Method" node.
The inpaint part is currently missing and will be implemented later.
I think they messed up this model pretty bad. They added some
control_noise_refiner blocks but don't actually use them. There is a typo
in their code so instead of doing control_noise_refiner -> control_layers
it runs the whole control_layers twice.
Unfortunately they trained with this typo so the model works but is kind
of slow and would probably perform a lot better if they corrected their
code and trained it again.
* make setattr safe for non existent attributes
Handle the case where the attribute doesnt exist by returning a static
sentinel (distinct from None). If the sentinel is passed in as the set
value, del the attr.
* Account for dequantization and type-casts in offload costs
When measuring the cost of offload, identify weights that need a type
change or dequantization and add the size of the conversion result
to the offload cost.
This is mutually exclusive with lowvram patches which already has
a large conservative estimate and wont overlap the dequant cost so\
dont double count.
* Set the compute type on CLIP MPs
So that the loader can know the size of weights for dequant accounting.
In the lowvram case, this now does its math in the model dtype in the
post de-quantization domain. Account for that. The patching was also
put back on the compute stream getting it off-peak so relax the
MATH_FACTOR to only x2 so get out of the worst-case assumption of
everything peaking at once.
slow down the CPU on model load to not run ahead. This fixes a VRAM on
flux 2 load.
I went to try and debug this with the memory trace pickles, which needs
--disable-cuda-malloc which made the bug go away. So I tried this
synchronize and it worked.
The has some very complex interactions with the cuda malloc async and
I dont have solid theory on this one yet.
Still debugging but this gets us over the OOM for the moment.
* Add Kandinsky5 model support
lite and pro T2V tested to work
* Update kandinsky5.py
* Fix fp8
* Fix fp8_scaled text encoder
* Add transformer_options for attention
* Code cleanup, optimizations, use fp32 for all layers originally at fp32
* ImageToVideo -node
* Fix I2V, add necessary latent post process nodes
* Support text to image model
* Support block replace patches (SLG mostly)
* Support official LoRAs
* Don't scale RoPE for lite model as that just doesn't work...
* Update supported_models.py
* Rever RoPE scaling to simpler one
* Fix typo
* Handle latent dim difference for image model in the VAE instead
* Add node to use different prompts for clip_l and qwen25_7b
* Reduce peak VRAM usage a bit
* Further reduce peak VRAM consumption by chunking ffn
* Update chunking
* Update memory_usage_factor
* Code cleanup, don't force the fp32 layers as it has minimal effect
* Allow for stronger changes with first frames normalization
Default values are too weak for any meaningful changes, these should probably be exposed as advanced node options when that's available.
* Add image model's own chat template, remove unused image2video template
* Remove hard error in ReplaceVideoLatentFrames -node
* Update kandinsky5.py
* Update supported_models.py
* Fix typos in prompt template
They were now fixed in the original repository as well
* Update ReplaceVideoLatentFrames
Add tooltips
Make source optional
Better handle negative index
* Rename NormalizeVideoLatentFrames -node
For bit better clarity what it does
* Fix NormalizeVideoLatentStart node out on non-op
* Apply cond slice fix
* Add FreeNoise
* Update context_windows.py
* Add option to retain condition by indexes for each window
This allows for example Wan/HunyuanVideo image to video to "work" by using the initial start frame for each window, otherwise windows beyond first will be pure T2V generations.
* Update context_windows.py
* Allow splitting multiple conds into different windows
* Add handling for audio_embed
* whitespace
* Allow freenoise to work on other dims, handle 4D batch timestep
Refactor Freenoise function. And fix batch handling as timesteps seem to be expanded to batch size now.
* Disable experimental options for now
So that the Freenoise and bugfixes can be merged first
---------
Co-authored-by: Jedrzej Kosinski <kosinkadink1@gmail.com>
Co-authored-by: ozbayb <17261091+ozbayb@users.noreply.github.com>
Im able to push vram above estimate on partial unload. Bump the
estimate. This is experimentally determined with a 720P and 480P
datapoint calibrating for 24GB VRAM total.
TIL that the WAN TE has a 2GB weight followed by 16MB as the next size
down. This means that team 8GB VRAM would fully offload the TE in async
offload mode as it just multiplied this giant size my the num streams.
Do the more complex logic of summing up the upcoming to-load weight
sizes to avoid triple counting this massive weight.
partial unload does the converse of recording the NS most recent
unloads as they go.
* mp: only count the offload cost of math once
This was previously bundling the combined weight storage and computation
cost
* ops: put all post async transfer compute on the main stream
Some models have massive weights that need either complex
dequantization or lora patching. Don't do these patchings on the offload
stream, instead do them on the main stream to syncrhonize the
potentially large vram spikes for these compute processes. This avoids
having to assume a worst case scenario of multiple offload streams
all spiking VRAM is parallel with whatever the main stream is doing.
* hunyuan upsampler: rework imports
Remove the transitive import of VideoConv3d and Resnet and takes these
from actual implementation source.
* model: remove unused give_pre_end
According to git grep, this is not used now, and was not used in the
initial commit that introduced it (see below).
This semantic is difficult to implement temporal roll VAE for (and would
defeat the purpose). Rather than implement the complex if, just delete
the unused feature.
(venv) rattus@rattus-box2:~/ComfyUI$ git log --oneline
220afe33 (HEAD) Initial commit.
(venv) rattus@rattus-box2:~/ComfyUI$ git grep give_pre
comfy/ldm/modules/diffusionmodules/model.py: resolution, z_channels, give_pre_end=False, tanh_out=False, use_linear_attn=False,
comfy/ldm/modules/diffusionmodules/model.py: self.give_pre_end = give_pre_end
comfy/ldm/modules/diffusionmodules/model.py: if self.give_pre_end:
(venv) rattus@rattus-box2:~/ComfyUI$ git co origin/master
Previous HEAD position was 220afe33 Initial commit.
HEAD is now at 9d8a8179 Enable async offloading by default on Nvidia. (#10953)
(venv) rattus@rattus-box2:~/ComfyUI$ git grep give_pre
comfy/ldm/modules/diffusionmodules/model.py: resolution, z_channels, give_pre_end=False, tanh_out=False, use_linear_attn=False,
comfy/ldm/modules/diffusionmodules/model.py: self.give_pre_end = give_pre_end
comfy/ldm/modules/diffusionmodules/model.py: if self.give_pre_end:
* move refiner VAE temporal roller to core
Move the carrying conv op to the common VAE code and give it a better
name. Roll the carry implementation logic for Resnet into the base
class and scrap the Hunyuan specific subclass.
* model: Add temporal roll to main VAE decoder
If there are no attention layers, its a standard resnet and VideoConv3d
is asked for, substitute in the temporal rolloing VAE algorithm. This
reduces VAE usage by the temporal dimension (can be huge VRAM savings).
* model: Add temporal roll to main VAE encoder
If there are no attention layers, its a standard resnet and VideoConv3d
is asked for, substitute in the temporal rolling VAE algorithm. This
reduces VAE usage by the temporal dimension (can be huge VRAM savings).
These are not actual controlnets so put it in the models/model_patches
folder and use the ModelPatchLoader + QwenImageDiffsynthControlnet node to
use it.
* Support video tiny VAEs
* lighttaew scaling fix
* Also support video taes in previews
Only first frame for now as live preview playback is currently only available through VHS custom nodes.
* Support Wan 2.1 lightVAE
* Relocate elif block and set Wan VAE dim directly without using pruning rate for lightvae
* mm: default to 0 for NUM_STREAMS
Dont count the compute stream as an offload stream. This makes async
offload accounting easier.
* mm: remove 128MB minimum
This is from a previous offloading system requirement. Remove it to
make behaviour of the loader and partial unloader consistent.
* mp: order the module list by offload expense
Calculate an approximate offloading temporary VRAM cost to offload a
weight and primary order the module load list by that. In the simple
case this is just the same as the module weight, but with Loras, a
weight with a lora consumes considerably more VRAM to do the Lora
application on-the-fly.
This will slightly prioritize lora weights, but is really for
proper VRAM offload accounting.
* mp: Account for the VRAM cost of weight offloading
when checking the VRAM headroom, assume that the weight needs to be
offloaded, and only load if it has space for both the load and offload
* the number of streams.
As the weights are ordered from largest to smallest by offload cost
this is guaranteed to fit in VRAM (tm), as all weights that follow
will be smaller.
Make the partial unload aware of this system as well by saving the
budget for offload VRAM to the model state and accounting accordingly.
Its possible that partial unload increases the size of the largest
offloaded weights, and thus needs to unload a little bit more than
asked to accomodate the bigger temp buffers.
Honor the existing codes floor on model weight loading of 128MB by
having the patcher honor this separately withough regard to offloading.
Otherwise when MM specifies its 128MB minimum, MP will see the biggest
weights, and budget that 128MB to only offload buffer and load nothing
which isnt the intent of these minimums. The same clamp applies in
case of partial offload of the currently loading model.
* init
* update
* Update model.py
* Update model.py
* remove print
* Fix text encoding
* Prevent empty negative prompt
Really doesn't work otherwise
* fp16 works
* I2V
* Update model_base.py
* Update nodes_hunyuan.py
* Better latent rgb factors
* Use the correct sigclip output...
* Support HunyuanVideo1.5 SR model
* whitespaces...
* Proper latent channel count
* SR model fixes
This also still needs timesteps scheduling based on the noise scale, can be used with two samplers too already
* vae_refiner: roll the convolution through temporal
Work in progress.
Roll the convolution through time using 2-latent-frame chunks and a
FIFO queue for the convolution seams.
* Support HunyuanVideo15 latent resampler
* fix
* Some cleanup
Co-Authored-By: comfyanonymous <121283862+comfyanonymous@users.noreply.github.com>
* Proper hyvid15 I2V channels
Co-Authored-By: comfyanonymous <121283862+comfyanonymous@users.noreply.github.com>
* Fix TokenRefiner for fp16
Otherwise x.sum has infs, just in case only casting if input is fp16, I don't know if necessary.
* Bugfix for the HunyuanVideo15 SR model
* vae_refiner: roll the convolution through temporal II
Roll the convolution through time using 2-latent-frame chunks and a
FIFO queue for the convolution seams.
Added support for encoder, lowered to 1 latent frame to save more
VRAM, made work for Hunyuan Image 3.0 (as code shared).
Fixed names, cleaned up code.
* Allow any number of input frames in VAE.
* Better VAE encode mem estimation.
* Lowvram fix.
* Fix hunyuan image 2.1 refiner.
* Fix mistake.
* Name changes.
* Rename.
* Whitespace.
* Fix.
* Fix.
---------
Co-authored-by: kijai <40791699+kijai@users.noreply.github.com>
Co-authored-by: Rattus <rattus128@gmail.com>
Clean up a bunch of stacked and no-longer-needed tensors on the QWEN
VRAM peak (currently FFN).
With this I go from OOMing at B=37x1328x1328 to being able to
succesfully run B=47 (RTX5090).
The partial unloader path in model re-use flow skips straight to the
actual unload without any check of the patching UUID. This means that
if you do an upscale flow with a model patch on an existing model, it
will not apply your patchings.
Fix by delaying the partial_unload until after the uuid checks. This
is done by making partial_unload a model of partial_load where extra_mem
is -ve.
* ops: dont take an offload stream if you dont need one
* ops: prioritize mem transfer
The async offload streams reason for existence is to transfer from
RAM to GPU. The post processing compute steps are a bonus on the side
stream, but if the compute stream is running a long kernel, it can
stall the side stream, as it wait to type-cast the bias before
transferring the weight. So do a pure xfer of the weight straight up,
then do everything bias, then go back to fix the weight type and do
weight patches.
vickytsang
Jukka Seppänen
fappaz
comfyanonymous
Tavi Halperin
rattus
Terry Jia
chaObserv
krigeta
askmyteapot
Kohaku-Blueleaf
blepping
tdrussell
asagi4
guill
ComfyUI Wiki
Omri Marom
Markury
Mylo
Ivan Zorin
rkfg
Silver
kelseyee
DELUXA
Jedrzej Kosinski
mengqin
drozbay
woctordho
BradPepersAMD
Haoming
Dr.Lt.Data
Farshore
Lodestone
dxqb
Urle Sistiana
contentis