Hugging Face's logo

LiconStudio
/
LTX-2.5-Multiple-Subject-Reference

Diffusers
video-generation
multi-reference
LTX-2.5
Model card Files
xet
 Community
7

identities not kept

#7
by ilkeraktuna - opened
Discussion
ilkeraktuna

I used the workflow: LTX2.5-MSR-sample-workflow-V2.json
and the following prompt:

Image 1: A man in jedi clothes, jedi robe and boots holding an orange lightsaber in one hand.
Image 2: A woman in jedi clothes, jedi robe and boots holding a green lightsaber in one hand.
Image 5: Scene, Tatooine planet from Star Wars. Binary sunset view with 2 suns.

The camera axis remains fixed throughout the sequence. Image 1 is established on the left side of the frame and Image 2 on the right side. Preserve their faces, bodies, identities, clothing, screen positions across every hard cut.

full body shot of two person on the Tatooine planet matching Image 5. Image 1 is on the left side with his orange lightsaber. Image 2 is on the right side with her green lightsaber.
They approach eachother. Then the woman (image 2) laughs and says, "Okay, let's start" The man (image 1) grins, and says, "Alright, if that's what you want". They both raise and clash their lightsabers in the air and the sound of clashing lightsabers fill the area.

But the result is not even close:

  1. 2 characters are totally different from the characters in image 1 and image 2 (secene is correct from image 5)
  2. characters are not doing the correct act
  3. characters are talking in a language that I don't understand (Chinese ?)
  4. lightsabers are weird

Everything in the workflow is same as the original workflow from here.
I just changed the diffusion model weight from default to fp8_e5m2 because although I have 128GB unified ram, workflow stucks at "Requested to load LTXAV" if I use "default" diffusion model weight.

So what is wrong ?
LiconStudio
 Owner

I used the workflow: LTX2.5-MSR-sample-workflow-V2.json
and the following prompt:

Image 1: A man in jedi clothes, jedi robe and boots holding an orange lightsaber in one hand.
Image 2: A woman in jedi clothes, jedi robe and boots holding a green lightsaber in one hand.
Image 5: Scene, Tatooine planet from Star Wars. Binary sunset view with 2 suns.

The camera axis remains fixed throughout the sequence. Image 1 is established on the left side of the frame and Image 2 on the right side. Preserve their faces, bodies, identities, clothing, screen positions across every hard cut.

full body shot of two person on the Tatooine planet matching Image 5. Image 1 is on the left side with his orange lightsaber. Image 2 is on the right side with her green lightsaber.
They approach eachother. Then the woman (image 2) laughs and says, "Okay, let's start" The man (image 1) grins, and says, "Alright, if that's what you want". They both raise and clash their lightsabers in the air and the sound of clashing lightsabers fill the area.

But the result is not even close:

  1. 2 characters are totally different from the characters in image 1 and image 2 (secene is correct from image 5)
  2. characters are not doing the correct act
  3. characters are talking in a language that I don't understand (Chinese ?)
  4. lightsabers are weird

Everything in the workflow is same as the original workflow from here.
I just changed the diffusion model weight from default to fp8_e5m2 because although I have 128GB unified ram, workflow stucks at "Requested to load LTXAV" if I use "default" diffusion model weight.

So what is wrong ?

Hi,

Thanks for sharing your prompt, results, and hardware details.

We tested the revised action prompt below and confirmed that it resolved the issue with the two characters not approaching each other in our test. The revised wording makes both characters' walking movements and the decreasing distance between them explicit.

Please also remove this sentence from the original prompt:

"Preserve their faces, bodies, identities, clothing, screen positions across every hard cut."

The instruction to preserve screen positions may conflict with the requested movement. We have not isolated it as the sole cause, but removing it avoids that potential conflict.

Please try this action prompt:

Full body shot of two people on the Tatooine planet matching Image 5. Image 1 is on the left side with his orange lightsaber. Image 2 is on the right side with her green lightsaber.
Both characters actively walk toward each other at the same time. The man walks from the left toward the woman, while the woman walks from the right toward the man. They each take several clear, visible steps forward, steadily closing the distance between them until they stop face-to-face within lightsaber striking distance. Keep both characters' full bodies and walking movements visible throughout their approach.
Then the woman (Image 2) laughs and says, "Okay, let's start." The man (Image 1) grins and says, "Alright, if that's what you want." They both raise their lightsabers and clash the blades in the air, and the sound of clashing lightsabers fills the area.

Regarding the other issues you reported:

1.Character appearance: You mentioned that the scene matches Image 5, but the characters do not match Images 1 and 2. We did not observe significant character consistency issues in our tests. Please try again with the revised prompt.

2.Dialogue language: You reported speech in an unexpected language. Please paste the prompt into the appropriate prompt field again and retry.

3.Lightsabers: Please share the generated video so we can check whether the issue involves blade shape, color, handling, or the clash itself. In our tests, we noticed some occasional visual artifacts, but nothing unusual beyond those.

4.Model loading: You mentioned that loading stalled at “Requested to load LTXAV” despite having 128GB of unified memory, so you changed the diffusion model weight setting from default to fp8_e5m2. We cannot yet determine whether this change affected the generated results or what caused the loading stall.

If any of these issues remain after trying the revised prompt, please send your exported workflow JSON, the generated video, your hardware and operating system details, and the loading log. These will help us compare your setup with ours and investigate the remaining issues separately.

Thanks again for the detailed feedback.
ilkeraktuna
•
edited 1 day ago

Thanks for your response. I changed the prompt as you advised. The movement has changed. But issues remain:

  1. people are different than the input
  2. spoken language is still not english
  3. there are more than 2 people by the end of video
  4. ending action is not as requested (no raise and clash of sabers)
  5. lightsabers are weird (they seem like toy lightsabers , ending points are not lit and one of the lightsabers is half colored, one is 2-colored)

I am also sending you input images here:

binary-sunset-2

Untitled-2
Untitled-1

This interface did not allow me to share json file so I pasted it here:
https://pastes.io/LUKA26gZ

OS is Ubuntu 24.04.4 LTS
I use Rocm 7.2 with my Radeon 8060S (strix halo) gpu and 128GB unified ram.

When I run the workflow with diffusion weight dtype default, it gets OOM at:

[INFO] got prompt
[INFO] VAE load device: cuda:0, offload device: cpu, dtype: torch.float32
[INFO] model weight dtype torch.bfloat16, manual cast: None
[INFO] model_type FLUX
[INFO] [LTX MSR-AVref] Loaded LTX-2.5-Licon-MSR-V2.safetensors with image/audio slot embeddings (5/5 tensors)
[INFO] [LTX MSR-AVref] image_dim=128 audio_dim=128 downscale=1 audio_layout=absolute_image_slot_windows duration=5.000s margin=0.040s max_audio_tokens=125
[INFO] VAE load device: cuda:0, offload device: cpu, dtype: torch.bfloat16
[INFO] CLIP/text encoder model load device: cuda:0, offload device: cpu, current: cpu, dtype: torch.float16
[INFO] Requested to load LTXAVTEModel_
[W1008 22:20:11.345933425 HIPCachingAllocator.cpp:4072] memory allocation failed with OOM on device 0 while trying to allocate 119537664 bytes (free: 27656192, total: 128849018880).
[W1008 22:20:11.373326205 HIPCachingAllocator.cpp:4072] memory allocation failed with OOM on device 0 while trying to allocate 119537664 bytes (free: 27656192, total: 128849018880).

So I run the workflow with diffusion weight dtype fp8_e5m2
May be this is the reason of all the mistakes.
But then why do I get OOM with lots of memory ? How can we prevent that ?

With the working dtype , here is the full log of Comfy from start to the end of workflow:

[INFO] setup plugin alembic.autogenerate.schemas
[INFO] setup plugin alembic.autogenerate.tables
[INFO] setup plugin alembic.autogenerate.types
[INFO] setup plugin alembic.autogenerate.constraints
[INFO] setup plugin alembic.autogenerate.defaults
[INFO] setup plugin alembic.autogenerate.comments
[INFO] [ComfyUI-Manager] Using uv as Python module for pip operations.
Using Python 3.12.3 environment at: venvnew
[START] Security scan
[DONE] Security scan

ComfyUI-Manager: installing dependencies done.

** ComfyUI startup time: 2026-10-08 23:56:30.693
** Platform: Linux
** Python version: 3.12.3 (main, Aug 31 2026, 10:18:26) [GCC 13.3.0]
** Python executable: /home/aadmin/ComfyUI/venvnew/bin/python3
** ComfyUI Path: /home/aadmin/ComfyUI
** ComfyUI Base Folder Path: /home/aadmin/ComfyUI
** User directory: /home/aadmin/ComfyUI/user
** ComfyUI-Manager config path: /home/aadmin/ComfyUI/user/manager/config.ini
** Log path: /home/aadmin/ComfyUI/user/comfyui.log
Using Python 3.12.3 environment at: venvnew
Using Python 3.12.3 environment at: venvnew
[INFO]
Prestartup times for custom nodes:
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/rgthree-comfy
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/comfyui-easy-use
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/comfyui-deno-custom-nodes
[INFO] 0.5 seconds: /home/aadmin/ComfyUI/custom_nodes/ComfyUI-Manager
[INFO]
[INFO] Found comfy_kitchen backend eager: {'available': True, 'disabled': False, 'unavailable_reason': None, 'capabilities': ['adaln', 'apply_rope', 'apply_rope1', 'apply_rope1', 'apply_rope', 'apply_rope_split_half', 'apply_rope_split_half1', 'apply_rope_split_half1_', 'apply_rope_split_half_', 'convrot_w4a4_linear', 'dequantize_convrot_w4a4_weight', 'dequantize_int8_convrot_weight', 'dequantize_int8_convrot_weight_dtype', 'dequantize_int8_embedding', 'dequantize_int8_simple', 'dequantize_int8_simple_dtype', 'dequantize_mxfp8', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'dequantize_w4a8_int8_weight', 'gemv_awq_w4a16', 'int8_linear', 'na3d', 'prepare_int4_weight_for_int8_linear', 'quantize_and_rotate_rowwise', 'quantize_convrot_w4a4_weight', 'quantize_int8_convrot_weight', 'quantize_int8_rowwise', 'quantize_int8_tensorwise', 'quantize_mxfp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8', 'quantize_svdquant_w4a4', 'quantize_w4a8_int8_weight', 'rms_adaln', 'rms_rope', 'rms_rope1', 'rms_rope1_', 'rms_rope_', 'rms_rope_split_half', 'rms_rope_split_half1', 'rms_rope_split_half1_', 'rms_rope_split_half_', 'rotate_int8_convrot_weight', 'scaled_mm_mxfp8', 'scaled_mm_nvfp4', 'scaled_mm_svdquant_w4a4', 'stochastic_rounding_fp8', 'w4a8_int8_linear']}
[INFO] Found comfy_kitchen backend cuda: {'available': True, 'disabled': True, 'unavailable_reason': None, 'capabilities': ['adaln', 'apply_rope', 'apply_rope1', 'apply_rope1_', 'apply_rope_', 'apply_rope_split_half', 'apply_rope_split_half1', 'apply_rope_split_half1_', 'apply_rope_split_half_', 'convrot_w4a4_linear', 'dequantize_convrot_w4a4_weight', 'dequantize_int8_convrot_weight', 'dequantize_int8_convrot_weight_dtype', 'dequantize_int8_simple', 'dequantize_int8_simple_dtype', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'dequantize_w4a8_int8_weight', 'gemv_awq_w4a16', 'int8_linear', 'na3d', 'prepare_int4_weight_for_int8_linear', 'quantize_and_rotate_rowwise', 'quantize_convrot_w4a4_weight', 'quantize_int8_convrot_weight', 'quantize_int8_rowwise', 'quantize_int8_tensorwise', 'quantize_mxfp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8', 'quantize_svdquant_w4a4', 'quantize_w4a8_int8_weight', 'rms_adaln', 'rms_rope', 'rms_rope1', 'rms_rope1_', 'rms_rope_', 'rms_rope_split_half', 'rms_rope_split_half1', 'rms_rope_split_half1_', 'rms_rope_split_half_', 'rotate_int8_convrot_weight', 'scaled_mm_nvfp4', 'scaled_mm_svdquant_w4a4', 'stochastic_rounding_fp8', 'w4a8_int8_linear']}
[INFO] Found comfy_kitchen backend triton: {'available': True, 'disabled': True, 'unavailable_reason': None, 'capabilities': ['adaln', 'apply_rope', 'apply_rope1', 'apply_rope1_', 'apply_rope_', 'apply_rope_split_half', 'apply_rope_split_half1', 'apply_rope_split_half1_', 'apply_rope_split_half_', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'int8_linear', 'na3d', 'quantize_and_rotate_rowwise', 'quantize_int8_rowwise', 'quantize_mxfp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8', 'rms_adaln', 'rms_rope', 'rms_rope1', 'rms_rope1_', 'rms_rope_', 'rms_rope_split_half', 'rms_rope_split_half1', 'rms_rope_split_half1_', 'rms_rope_split_half_', 'w4a8_int8_linear']}
[INFO] Found comfy_kitchen backend hip: {'available': True, 'disabled': False, 'unavailable_reason': None, 'capabilities': ['adaln', 'apply_rope', 'apply_rope1', 'apply_rope1_', 'apply_rope_', 'apply_rope_split_half', 'apply_rope_split_half1', 'apply_rope_split_half1_', 'apply_rope_split_half_', 'convrot_w4a4_linear', 'dequantize_convrot_w4a4_weight', 'dequantize_int8_convrot_weight_dtype', 'dequantize_int8_simple_dtype', 'dequantize_per_tensor_fp8', 'dequantize_w4a8_int8_weight', 'gemv_awq_w4a16', 'int8_linear', 'na3d', 'quantize_and_rotate_rowwise', 'quantize_convrot_w4a4_weight', 'quantize_int8_convrot_weight', 'quantize_int8_rowwise', 'quantize_int8_tensorwise', 'quantize_per_tensor_fp8', 'quantize_svdquant_w4a4', 'quantize_w4a8_int8_weight', 'rms_adaln', 'rms_rope', 'rms_rope1', 'rms_rope1_', 'rms_rope_', 'rms_rope_split_half', 'rms_rope_split_half1', 'rms_rope_split_half1_', 'rms_rope_split_half_', 'scaled_mm_svdquant_w4a4', 'stochastic_rounding_fp8', 'w4a8_int8_linear']}
[INFO] Checkpoint files will always be loaded safely.
[INFO] Total VRAM 122880 MB, total RAM 126429 MB
[INFO] pytorch version: 2.15.0.dev20260827+rocm7.2
[WARNING] WARNING[XFORMERS]: xFormers can't load C++/CUDA extensions. xFormers was built for:
PyTorch 2.7.1+rocm7.2.3.git1dab218d with CUDA None (you have 2.15.0.dev20260827+rocm7.2)
Python 3.12.13 (you have 3.12.3)
Please reinstall xformers (see https://github.com/facebookresearch/xformers#installing-xformers)
Memory-efficient attention, SwiGLU, sparse and more won't be available.
Set XFORMERS_MORE_DETAILS=1 for more details
[aiter] import [module_aiter_core] under /home/aadmin/ComfyUI/venvnew/lib/python3.12/site-packages/aiter/jit/module_aiter_core.so
/home/aadmin/ComfyUI/venvnew/lib/python3.12/site-packages/aiter/ops/deepgemm.py:22: RuntimeWarning: aiter.ops.opus (a16w16) is gfx950-only; detected arch='gfx1100'. opus_gemm_* calls will raise RuntimeError at invocation. opus_gemm uses gfx950-only intrinsics (MFMA, ds_read_b64_tr) and the 160 KiB LDS budget. Set GPU_ARCHS=gfx950 (or run on a gfx950 device) to use this module.
from .opus.gemm_op_a16w16 import opus_gemm_a16w16_tune as _opus_tune
[INFO] Set: torch.backends.cudnn.enabled = False for better AMD performance.
[INFO] AMD arch: gfx1100
[INFO] ROCm version: (7, 2)
[INFO] Set vram state to: HIGH_VRAM
[INFO] Device: cuda:0 AMD Radeon 8060S : native
[INFO] Using async weight offloading with 2 streams
[INFO] Enabled pinned memory 113785.0
[INFO] Using Flash Attention
[INFO] Python version: 3.12.3 (main, Aug 31 2026, 10:18:26) [GCC 13.3.0]
[INFO] ComfyUI version: 0.34.2
[INFO] comfy-aimdo version: 0.4.15
[INFO] comfy-kitchen version: 0.2.31
[INFO] comfyui-frontend-package version: 1.49.6
[INFO] comfyui-workflow-templates version: 0.11.50
[INFO] comfyui-embedded-docs version: 0.5.10
[INFO] comfy-kitchen version: 0.2.31
[INFO] comfy-aimdo version: 0.4.15
[INFO] [Prompt Server] web root: /home/aadmin/ComfyUI/venvnew/lib/python3.12/site-packages/comfyui_frontend_package/static
[INFO] Asset seeder disabled
[INFO] No OpenGL_accelerate module loaded: No module named 'OpenGL_accelerate'
Adding /home/aadmin/ComfyUI/custom_nodes to sys.path
Could not find efficiency nodes
Could not find comfyui_controlnet_aux nodes, AV_ControlNetPreprocessor will not work. Please install comfyui_controlnet_aux first
Could not find AdvancedControlNet nodes
Could not find AnimateDiff nodes
Loaded IPAdapter nodes from /home/aadmin/ComfyUI/custom_nodes/comfyui_ipadapter_plus
Could not find VideoHelperSuite nodes
Could not load ImpactPack nodes Could not find ImpactPack nodes
[INFO] ### Loading: ComfyUI-Manager (V3.41)
[INFO] [ComfyUI-Manager] network_mode: public
[INFO] [ComfyUI-Manager] ComfyUI per-queue preview override detected (PR #11261). Manager's preview method feature is disabled. Use ComfyUI's --preview-method CLI option or 'Settings > Execution > Live preview method'.
[INFO] ### ComfyUI Revision: 5830 [169fcf35] *DETACHED | Released on '2026-08-27'

  ＡＦ  －  ＣｏｍｆｙＵＩ  Ｎｏｄｅｓ
                                 
   🚀 AF - Prompt Nodes Pack Loaded!

[krea2edit] nodes v1.2.5 loaded
[INFO] [ComfyUI-Manager] default cache updated: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/model-list.json
[INFO] [ComfyUI-Manager] default cache updated: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/alter-list.json
[INFO] [ComfyUI-Manager] default cache updated: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/github-stats.json
[INFO] [ComfyUI-Manager] default cache updated: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/extension-node-map.json
[INFO] [ComfyUI-Manager] default cache updated: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/custom-node-list.json
WAS Node Suite: OpenCV Python FFMPEG support is enabled
WAS Node Suite Warning: ffmpeg_bin_path is not set in /home/aadmin/ComfyUI/custom_nodes/was-ns/was_suite_config.json config file. Will attempt to use system ffmpeg binaries if available.
WAS Node Suite: Finished. Loaded 220 nodes successfully.

    "Believe you deserve it and the universe will serve it." - Unknown

AMD GPU Monitor thread started
Using AMD SMI tool: /opt/rocm/bin/rocm-smiAMD GPU Monitor: Web directory set to /home/aadmin/ComfyUI/custom_nodes/amdgpumonitor/web

Adaptive LoRA Scheduler Node: Loaded

  ＡＦ  －  ＣｏｍｆｙＵＩ  Ｎｏｄｅｓ
                                 
 🔍 AF - Find Nodes Extension Loaded !

Use Ctrl+Shift+F to open the search panel

[INFO] ComfyUI-GGUF: Allowing full torch compile
[INFO] Using Flash Attention
(RES4LYF) Init
(RES4LYF) Importing beta samplers.
(RES4LYF) Importing legacy samplers.
Module 'diffusers' load failed. If you don't have it installed, do it:
pip install diffusers
[ComfyUI-Easy-Use] server: v1.3.6 Loaded
[ComfyUI-Easy-Use] web root: /home/aadmin/ComfyUI/custom_nodes/comfyui-easy-use/web_version/v2 Loaded
[PainterNodes] Failed to import PainterVRAM: NVML Shared Library Not Found
[PainterNodes] Loaded 35 nodes successfully!

😺dzNodes: LayerStyle -> Cannot import name 'guidedFilter' from 'cv2.ximgproc'

A few nodes cannot works properly, while most nodes are not affected. Please REINSTALL package 'opencv-contrib-python'.
For detail refer to https://github.com/chflame163/ComfyUI_LayerStyle/issues/5

[rgthree-comfy] Loaded 48 extraordinary nodes. 🎉

[rgthree-comfy] ComfyUI's new Node 2.0 rendering may be incompatible with some rgthree-comfy nodes and features, breaking some rendering as well as losing the ability to access a node's properties (a vital part of many nodes). It also appears to run MUCH more slowly spiking CPU usage and causing jankiness and unresponsiveness, especially with large workflows. Personally I am not planning to use the new Nodes 2.0 and, unfortunately, am not able to invest the time to investigate and overhaul rgthree-comfy where needed. If you have issues when Nodes 2.0 is enabled, I'd urge you to switch it off as well and join me in hoping ComfyUI is not planning to deprecate the existing, stable canvas rendering all together.

[INFO]
Import times for custom nodes:
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/websocket_image_save.py
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/text_adapter
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/comfyui-af-find-nodes
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/krea2-twostage-sampler
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/comfyui-qwenmultiangle
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/comfyui-krea2edit
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/comfyui-af-pack-prompt-nodes
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/comfyui-mxtoolkit
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/ltx2.5-msr
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/comfyui_ipadapter_plus
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/ComfyUI-Krea2T-Enhancer
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/ComfyUI-PromptRelay
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/amdgpumonitor
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/ComfyUI-GGUF
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/ComfyUI-Dynamic-Lora-Scheduler
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/ComfyMath
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/comfyui-various
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/vnccs
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/vnccs-utils
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/10s-comfy-nodes
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/rgthree-comfy
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/comfyui-painternodes
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/ComfyUI-PromptEnhancer
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/comfyui-animatediff-evolved
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/ComfyUI-LTXVideo
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/comfyui-art-venture
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/comfyui-kjnodes
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/comfyui-logicutils
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/comfyui-easy-use
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/comfyui-deno-custom-nodes
[INFO] 0.1 seconds: /home/aadmin/ComfyUI/custom_nodes/comfyui_layerstyle
[INFO] 0.1 seconds: /home/aadmin/ComfyUI/custom_nodes/ComfyUI-Manager
[INFO] 0.1 seconds: /home/aadmin/ComfyUI/custom_nodes/ComfyUI-WanVideoWrapper
[INFO] 0.2 seconds: /home/aadmin/ComfyUI/custom_nodes/comfyui-videohelpersuite
[INFO] 0.2 seconds: /home/aadmin/ComfyUI/custom_nodes/RES4LYF
[INFO] 0.5 seconds: /home/aadmin/ComfyUI/custom_nodes/was-ns
[INFO] 3.3 seconds: /home/aadmin/ComfyUI/custom_nodes/ltx2_sm
[INFO]
[INFO] Context impl SQLiteImpl.
[INFO] Will assume non-transactional DDL.
[INFO] Using RAM pressure cache.
[INFO] Starting server

[INFO] To see the GUI go to: http://0.0.0.0:8188
[INFO] To see the GUI go to: http://[::]:8188
[INFO] got prompt
[WARNING] [DENO] Resource Monitor could not initialize NVML: NVML Shared Library Not Found
[INFO] VAE load device: cuda:0, offload device: cpu, dtype: torch.float32
[INFO] model weight dtype torch.float8_e5m2, manual cast: torch.bfloat16
[INFO] model_type FLUX
[INFO] [LTX MSR-AVref] Loaded LTX-2.5-Licon-MSR-V2.safetensors with image/audio slot embeddings (5/5 tensors)
[INFO] [LTX MSR-AVref] image_dim=128 audio_dim=128 downscale=1 audio_layout=absolute_image_slot_windows duration=5.000s margin=0.040s max_audio_tokens=125
[INFO] VAE load device: cuda:0, offload device: cpu, dtype: torch.bfloat16
[INFO] Requested to load LTXAVTEModel_
[INFO] loaded completely; 24999.98 MB loaded, full load: True
[INFO] CLIP/text encoder model load device: cuda:0, offload device: cpu, current: cuda:0, dtype: torch.float16
[INFO] [PromptRelay] Global: tokens [0:74] (74 tokens)
[INFO] [PromptRelay] Segment 0: tokens [74:268] (194 tokens)
[INFO] [PromptRelay] Latent: 47 frames, 390 tokens/frame, segments: [47]
[INFO] [LTX MSR] Guide start: mode=MSR, references=3, reference_frames=33, downscale=1, temporal_scale=1, target_latent=(1, 128, 47, 15, 26)
[INFO] Requested to load CausalDiffusionVAE
[INFO] loaded completely; 1403.92 MB loaded, full load: True
[WARNING] [LTX MSR][VERIFY] slot_embedding=APPLIED label=pic1 slot_id=1 operation=guide_latent_plus_broadcast_embedding embedding_dim=128 embedding_norm=0.739934 embedding_preview=[0.010132, -0.057821, -0.10936, -0.010906, -0.044901, 0.034269, 0.023298, -0.006341]
[INFO] [LTX MSR] pic1: slot_embedding=applied, slot_id=1, embedding_dim=128, embedding_norm=0.739934, time_offset=-3, guide_latent=(1, 128, 5, 15, 26)
[WARNING] [LTX MSR][VERIFY] negative_time_offset=APPLIED label=pic1 slot_id=1 operation=append_keyframe frame_offset=-3 temporal_scale=1 compatibility_default=APPLIED temporal_position=negative_3 append_keyframe=SUCCESS
[WARNING] [LTX MSR][VERIFY] slot_embedding=APPLIED label=pic2 slot_id=2 operation=guide_latent_plus_broadcast_embedding embedding_dim=128 embedding_norm=0.767755 embedding_preview=[-0.015987, -0.037434, -0.089852, -0.035059, -0.07984, 0.021686, 0.01271, 0.018696]
[INFO] [LTX MSR] pic2: slot_embedding=applied, slot_id=2, embedding_dim=128, embedding_norm=0.767755, time_offset=-2, guide_latent=(1, 128, 5, 15, 26)
[WARNING] [LTX MSR][VERIFY] negative_time_offset=APPLIED label=pic2 slot_id=2 operation=append_keyframe frame_offset=-2 temporal_scale=1 compatibility_default=APPLIED temporal_position=negative_2 append_keyframe=SUCCESS
[WARNING] [LTX MSR][VERIFY] slot_embedding=APPLIED label=background slot_id=3 operation=guide_latent_plus_broadcast_embedding embedding_dim=128 embedding_norm=0.847053 embedding_preview=[-0.027266, -0.034561, -0.077652, -0.075744, -0.083365, 0.021973, 0.021777, 0.014609]
[INFO] [LTX MSR] background: slot_embedding=applied, slot_id=3, embedding_dim=128, embedding_norm=0.847053, time_offset=-1, guide_latent=(1, 128, 5, 15, 26)
[WARNING] [LTX MSR][VERIFY] negative_time_offset=APPLIED label=background slot_id=3 operation=append_keyframe frame_offset=-1 temporal_scale=1 compatibility_default=APPLIED temporal_position=negative_1 append_keyframe=SUCCESS
[INFO] [LTX MSR] Guide complete: added=3, order=pic1..pic3, frames_each=33, mode=MSR, output_latent=(1, 128, 62, 15, 26)
[INFO] Requested to load LTXAV
[INFO] loaded completely; 20025.73 MB loaded, full load: True
0%| | 0/8 [00:00<?, ?it/s][INFO] [PromptRelay] Built penalty matrix (scaled): Lq=24180, Lk=1024, nonzero=498968/24760320
[INFO] [PromptRelay] Built penalty matrix (scaled): Lq=376, Lk=1024, nonzero=7566/385024
12%|█▎ | 1/8 [01:36<11:17, 96.84s/it]FETCH ComfyRegistry Data [DONE]
[INFO] [ComfyUI-Manager] default cache updated: https://api.comfy.org/nodes
FETCH DATA from: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/custom-node-list.json [DONE]
[INFO] [ComfyUI-Manager] All startup tasks have been completed.
100%|██████████| 8/8 [12:31<00:00, 93.95s/it]
[INFO] Requested to load AudioVAE
[INFO] loaded completely; 693.46 MB loaded, full load: True
[INFO] Requested to load LatentUpsampler
[INFO] loaded completely; 949.61 MB loaded, full load: True
[INFO] [PromptRelay] Global: tokens [0:74] (74 tokens)
[INFO] [PromptRelay] Segment 0: tokens [74:268] (194 tokens)
[INFO] [PromptRelay] Latent: 47 frames, 1560 tokens/frame, segments: [47]
[INFO] [LTX MSR] Guide start: mode=MSR, references=3, reference_frames=33, downscale=1, temporal_scale=1, target_latent=(1, 128, 47, 30, 52)
[W1009 00:13:12.132929361 HIPCachingAllocator.cpp:4072] memory allocation failed with OOM on device 0 while trying to allocate 5521801216 bytes (free: 4967206912, total: 128849018880).
[WARNING] [LTX MSR][VERIFY] slot_embedding=APPLIED label=pic1 slot_id=1 operation=guide_latent_plus_broadcast_embedding embedding_dim=128 embedding_norm=0.739934 embedding_preview=[0.010132, -0.057821, -0.10936, -0.010906, -0.044901, 0.034269, 0.023298, -0.006341]
[INFO] [LTX MSR] pic1: slot_embedding=applied, slot_id=1, embedding_dim=128, embedding_norm=0.739934, time_offset=-3, guide_latent=(1, 128, 5, 30, 52)
[WARNING] [LTX MSR][VERIFY] negative_time_offset=APPLIED label=pic1 slot_id=1 operation=append_keyframe frame_offset=-3 temporal_scale=1 compatibility_default=APPLIED temporal_position=negative_3 append_keyframe=SUCCESS
[WARNING] [LTX MSR][VERIFY] slot_embedding=APPLIED label=pic2 slot_id=2 operation=guide_latent_plus_broadcast_embedding embedding_dim=128 embedding_norm=0.767755 embedding_preview=[-0.015987, -0.037434, -0.089852, -0.035059, -0.07984, 0.021686, 0.01271, 0.018696]
[INFO] [LTX MSR] pic2: slot_embedding=applied, slot_id=2, embedding_dim=128, embedding_norm=0.767755, time_offset=-2, guide_latent=(1, 128, 5, 30, 52)
[WARNING] [LTX MSR][VERIFY] negative_time_offset=APPLIED label=pic2 slot_id=2 operation=append_keyframe frame_offset=-2 temporal_scale=1 compatibility_default=APPLIED temporal_position=negative_2 append_keyframe=SUCCESS
[WARNING] [LTX MSR][VERIFY] slot_embedding=APPLIED label=background slot_id=3 operation=guide_latent_plus_broadcast_embedding embedding_dim=128 embedding_norm=0.847053 embedding_preview=[-0.027266, -0.034561, -0.077652, -0.075744, -0.083365, 0.021973, 0.021777, 0.014609]
[INFO] [LTX MSR] background: slot_embedding=applied, slot_id=3, embedding_dim=128, embedding_norm=0.847053, time_offset=-1, guide_latent=(1, 128, 5, 30, 52)
[WARNING] [LTX MSR][VERIFY] negative_time_offset=APPLIED label=background slot_id=3 operation=append_keyframe frame_offset=-1 temporal_scale=1 compatibility_default=APPLIED temporal_position=negative_1 append_keyframe=SUCCESS
[INFO] [LTX MSR] Guide complete: added=3, order=pic1..pic3, frames_each=33, mode=MSR, output_latent=(1, 128, 62, 30, 52)
[INFO] Requested to load LTXAV
0%| | 0/3 [00:00<?, ?it/s][INFO] [PromptRelay] Built penalty matrix (scaled): Lq=96720, Lk=1024, nonzero=1996066/99041280
[INFO] [PromptRelay] Built penalty matrix (scaled): Lq=376, Lk=1024, nonzero=7566/385024
100%|██████████| 3/3 [34:15<00:00, 685.25s/it]
[INFO] Prompt executed in 00:57:11

I just tried the "default" type again and it filled the GTT up to 92284MB and got stuck:
92284M / 131054M GTT 70.42% ³

log:
[INFO] got prompt
[INFO] VAE load device: cuda:0, offload device: cpu, dtype: torch.float32

[INFO] model weight dtype torch.bfloat16, manual cast: None
[INFO] model_type FLUX
[INFO] [LTX MSR-AVref] Loaded LTX-2.5-Licon-MSR-V2.safetensors with image/audio slot embeddings (5/5 tensors)
[INFO] [LTX MSR-AVref] image_dim=128 audio_dim=128 downscale=1 audio_layout=absolute_image_slot_windows duration=5.000s margin=0.040s max_audio_tokens=125
[INFO] VAE load device: cuda:0, offload device: cpu, dtype: torch.bfloat16
[INFO] Requested to load LTXAVTEModel_
[INFO] loaded completely; 24999.98 MB loaded, full load: True
[INFO] CLIP/text encoder model load device: cuda:0, offload device: cpu, current: cuda:0, dtype: torch.float16
[INFO] [PromptRelay] Global: tokens [0:74] (74 tokens)
[INFO] [PromptRelay] Segment 0: tokens [74:268] (194 tokens)
[INFO] [PromptRelay] Latent: 47 frames, 390 tokens/frame, segments: [47]
[INFO] [LTX MSR] Guide start: mode=MSR, references=3, reference_frames=33, downscale=1, temporal_scale=1, target_latent=(1, 128, 47, 15, 26)
[INFO] Requested to load CausalDiffusionVAE
[INFO] loaded completely; 1403.92 MB loaded, full load: True
[WARNING] [LTX MSR][VERIFY] slot_embedding=APPLIED label=pic1 slot_id=1 operation=guide_latent_plus_broadcast_embedding embedding_dim=128 embedding_norm=0.739934 embedding_preview=[0.010132, -0.057821, -0.10936, -0.010906, -0.044901, 0.034269, 0.023298, -0.006341]
[INFO] [LTX MSR] pic1: slot_embedding=applied, slot_id=1, embedding_dim=128, embedding_norm=0.739934, time_offset=-3, guide_latent=(1, 128, 5, 15, 26)
[WARNING] [LTX MSR][VERIFY] negative_time_offset=APPLIED label=pic1 slot_id=1 operation=append_keyframe frame_offset=-3 temporal_scale=1 compatibility_default=APPLIED temporal_position=negative_3 append_keyframe=SUCCESS
[WARNING] [LTX MSR][VERIFY] slot_embedding=APPLIED label=pic2 slot_id=2 operation=guide_latent_plus_broadcast_embedding embedding_dim=128 embedding_norm=0.767755 embedding_preview=[-0.015987, -0.037434, -0.089852, -0.035059, -0.07984, 0.021686, 0.01271, 0.018696]
[INFO] [LTX MSR] pic2: slot_embedding=applied, slot_id=2, embedding_dim=128, embedding_norm=0.767755, time_offset=-2, guide_latent=(1, 128, 5, 15, 26)
[WARNING] [LTX MSR][VERIFY] negative_time_offset=APPLIED label=pic2 slot_id=2 operation=append_keyframe frame_offset=-2 temporal_scale=1 compatibility_default=APPLIED temporal_position=negative_2 append_keyframe=SUCCESS
[WARNING] [LTX MSR][VERIFY] slot_embedding=APPLIED label=background slot_id=3 operation=guide_latent_plus_broadcast_embedding embedding_dim=128 embedding_norm=0.847053 embedding_preview=[-0.027266, -0.034561, -0.077652, -0.075744, -0.083365, 0.021973, 0.021777, 0.014609]
[INFO] [LTX MSR] background: slot_embedding=applied, slot_id=3, embedding_dim=128, embedding_norm=0.847053, time_offset=-1, guide_latent=(1, 128, 5, 15, 26)
[WARNING] [LTX MSR][VERIFY] negative_time_offset=APPLIED label=background slot_id=3 operation=append_keyframe frame_offset=-1 temporal_scale=1 compatibility_default=APPLIED temporal_position=negative_1 append_keyframe=SUCCESS
[INFO] [LTX MSR] Guide complete: added=3, order=pic1..pic3, frames_each=33, mode=MSR, output_latent=(1, 128, 62, 15, 26)
[INFO] Requested to load LTXAV
[INFO] Unloaded partially: 13674.97 MB freed, 11341.42 MB remains loaded, 562.50 MB buffer reserved, lowvram patches: 0
FETCH ComfyRegistry Data [DONE]
[INFO] [ComfyUI-Manager] default cache updated: https://api.comfy.org/nodes
FETCH DATA from: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/custom-node-list.json [DONE]
[INFO] [ComfyUI-Manager] All startup tasks have been completed.
[INFO] loaded completely; 40051.47 MB loaded, full load: True
0%| | 0/8 [00:00<?, ?it/s][INFO] [PromptRelay] Built penalty matrix (scaled): Lq=24180, Lk=1024, nonzero=498968/24760320
[INFO] [PromptRelay] Built penalty matrix (scaled): Lq=376, Lk=1024, nonzero=7566/385024
100%|██████████| 8/8 [08:35<00:00, 64.39s/it]
[INFO] Requested to load AudioVAE
[INFO] loaded completely; 693.46 MB loaded, full load: True
[INFO] Requested to load CausalDiffusionVAE
LiconStudio
 Owner

Thanks for your response. I changed the prompt as you advised. The movement has changed. But issues remain:

  1. people are different than the input
  2. spoken language is still not english
  3. there are more than 2 people by the end of video
  4. ending action is not as requested (no raise and clash of sabers)
  5. lightsabers are weird (they seem like toy lightsabers , ending points are not lit and one of the lightsabers is half colored, one is 2-colored)

I am also sending you input images here:

binary-sunset-2

Untitled-2
Untitled-1

This interface did not allow me to share json file so I pasted it here:
https://pastes.io/LUKA26gZ

OS is Ubuntu 24.04.4 LTS
I use Rocm 7.2 with my Radeon 8060S (strix halo) gpu and 128GB unified ram.

When I run the workflow with diffusion weight dtype default, it gets OOM at:

[INFO] got prompt
[INFO] VAE load device: cuda:0, offload device: cpu, dtype: torch.float32
[INFO] model weight dtype torch.bfloat16, manual cast: None
[INFO] model_type FLUX
[INFO] [LTX MSR-AVref] Loaded LTX-2.5-Licon-MSR-V2.safetensors with image/audio slot embeddings (5/5 tensors)
[INFO] [LTX MSR-AVref] image_dim=128 audio_dim=128 downscale=1 audio_layout=absolute_image_slot_windows duration=5.000s margin=0.040s max_audio_tokens=125
[INFO] VAE load device: cuda:0, offload device: cpu, dtype: torch.bfloat16
[INFO] CLIP/text encoder model load device: cuda:0, offload device: cpu, current: cpu, dtype: torch.float16
[INFO] Requested to load LTXAVTEModel_
[W1008 22:20:11.345933425 HIPCachingAllocator.cpp:4072] memory allocation failed with OOM on device 0 while trying to allocate 119537664 bytes (free: 27656192, total: 128849018880).
[W1008 22:20:11.373326205 HIPCachingAllocator.cpp:4072] memory allocation failed with OOM on device 0 while trying to allocate 119537664 bytes (free: 27656192, total: 128849018880).

So I run the workflow with diffusion weight dtype fp8_e5m2
May be this is the reason of all the mistakes.
But then why do I get OOM with lots of memory ? How can we prevent that ?

With the working dtype , here is the full log of Comfy from start to the end of workflow:

[INFO] setup plugin alembic.autogenerate.schemas
[INFO] setup plugin alembic.autogenerate.tables
[INFO] setup plugin alembic.autogenerate.types
[INFO] setup plugin alembic.autogenerate.constraints
[INFO] setup plugin alembic.autogenerate.defaults
[INFO] setup plugin alembic.autogenerate.comments
[INFO] [ComfyUI-Manager] Using uv as Python module for pip operations.
Using Python 3.12.3 environment at: venvnew
[START] Security scan
[DONE] Security scan

ComfyUI-Manager: installing dependencies done.

** ComfyUI startup time: 2026-10-08 23:56:30.693
** Platform: Linux
** Python version: 3.12.3 (main, Aug 31 2026, 10:18:26) [GCC 13.3.0]
** Python executable: /home/aadmin/ComfyUI/venvnew/bin/python3
** ComfyUI Path: /home/aadmin/ComfyUI
** ComfyUI Base Folder Path: /home/aadmin/ComfyUI
** User directory: /home/aadmin/ComfyUI/user
** ComfyUI-Manager config path: /home/aadmin/ComfyUI/user/manager/config.ini
** Log path: /home/aadmin/ComfyUI/user/comfyui.log
Using Python 3.12.3 environment at: venvnew
Using Python 3.12.3 environment at: venvnew
[INFO]
Prestartup times for custom nodes:
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/rgthree-comfy
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/comfyui-easy-use
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/comfyui-deno-custom-nodes
[INFO] 0.5 seconds: /home/aadmin/ComfyUI/custom_nodes/ComfyUI-Manager
[INFO]
[INFO] Found comfy_kitchen backend eager: {'available': True, 'disabled': False, 'unavailable_reason': None, 'capabilities': ['adaln', 'apply_rope', 'apply_rope1', 'apply_rope1', 'apply_rope', 'apply_rope_split_half', 'apply_rope_split_half1', 'apply_rope_split_half1_', 'apply_rope_split_half_', 'convrot_w4a4_linear', 'dequantize_convrot_w4a4_weight', 'dequantize_int8_convrot_weight', 'dequantize_int8_convrot_weight_dtype', 'dequantize_int8_embedding', 'dequantize_int8_simple', 'dequantize_int8_simple_dtype', 'dequantize_mxfp8', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'dequantize_w4a8_int8_weight', 'gemv_awq_w4a16', 'int8_linear', 'na3d', 'prepare_int4_weight_for_int8_linear', 'quantize_and_rotate_rowwise', 'quantize_convrot_w4a4_weight', 'quantize_int8_convrot_weight', 'quantize_int8_rowwise', 'quantize_int8_tensorwise', 'quantize_mxfp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8', 'quantize_svdquant_w4a4', 'quantize_w4a8_int8_weight', 'rms_adaln', 'rms_rope', 'rms_rope1', 'rms_rope1_', 'rms_rope_', 'rms_rope_split_half', 'rms_rope_split_half1', 'rms_rope_split_half1_', 'rms_rope_split_half_', 'rotate_int8_convrot_weight', 'scaled_mm_mxfp8', 'scaled_mm_nvfp4', 'scaled_mm_svdquant_w4a4', 'stochastic_rounding_fp8', 'w4a8_int8_linear']}
[INFO] Found comfy_kitchen backend cuda: {'available': True, 'disabled': True, 'unavailable_reason': None, 'capabilities': ['adaln', 'apply_rope', 'apply_rope1', 'apply_rope1_', 'apply_rope_', 'apply_rope_split_half', 'apply_rope_split_half1', 'apply_rope_split_half1_', 'apply_rope_split_half_', 'convrot_w4a4_linear', 'dequantize_convrot_w4a4_weight', 'dequantize_int8_convrot_weight', 'dequantize_int8_convrot_weight_dtype', 'dequantize_int8_simple', 'dequantize_int8_simple_dtype', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'dequantize_w4a8_int8_weight', 'gemv_awq_w4a16', 'int8_linear', 'na3d', 'prepare_int4_weight_for_int8_linear', 'quantize_and_rotate_rowwise', 'quantize_convrot_w4a4_weight', 'quantize_int8_convrot_weight', 'quantize_int8_rowwise', 'quantize_int8_tensorwise', 'quantize_mxfp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8', 'quantize_svdquant_w4a4', 'quantize_w4a8_int8_weight', 'rms_adaln', 'rms_rope', 'rms_rope1', 'rms_rope1_', 'rms_rope_', 'rms_rope_split_half', 'rms_rope_split_half1', 'rms_rope_split_half1_', 'rms_rope_split_half_', 'rotate_int8_convrot_weight', 'scaled_mm_nvfp4', 'scaled_mm_svdquant_w4a4', 'stochastic_rounding_fp8', 'w4a8_int8_linear']}
[INFO] Found comfy_kitchen backend triton: {'available': True, 'disabled': True, 'unavailable_reason': None, 'capabilities': ['adaln', 'apply_rope', 'apply_rope1', 'apply_rope1_', 'apply_rope_', 'apply_rope_split_half', 'apply_rope_split_half1', 'apply_rope_split_half1_', 'apply_rope_split_half_', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'int8_linear', 'na3d', 'quantize_and_rotate_rowwise', 'quantize_int8_rowwise', 'quantize_mxfp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8', 'rms_adaln', 'rms_rope', 'rms_rope1', 'rms_rope1_', 'rms_rope_', 'rms_rope_split_half', 'rms_rope_split_half1', 'rms_rope_split_half1_', 'rms_rope_split_half_', 'w4a8_int8_linear']}
[INFO] Found comfy_kitchen backend hip: {'available': True, 'disabled': False, 'unavailable_reason': None, 'capabilities': ['adaln', 'apply_rope', 'apply_rope1', 'apply_rope1_', 'apply_rope_', 'apply_rope_split_half', 'apply_rope_split_half1', 'apply_rope_split_half1_', 'apply_rope_split_half_', 'convrot_w4a4_linear', 'dequantize_convrot_w4a4_weight', 'dequantize_int8_convrot_weight_dtype', 'dequantize_int8_simple_dtype', 'dequantize_per_tensor_fp8', 'dequantize_w4a8_int8_weight', 'gemv_awq_w4a16', 'int8_linear', 'na3d', 'quantize_and_rotate_rowwise', 'quantize_convrot_w4a4_weight', 'quantize_int8_convrot_weight', 'quantize_int8_rowwise', 'quantize_int8_tensorwise', 'quantize_per_tensor_fp8', 'quantize_svdquant_w4a4', 'quantize_w4a8_int8_weight', 'rms_adaln', 'rms_rope', 'rms_rope1', 'rms_rope1_', 'rms_rope_', 'rms_rope_split_half', 'rms_rope_split_half1', 'rms_rope_split_half1_', 'rms_rope_split_half_', 'scaled_mm_svdquant_w4a4', 'stochastic_rounding_fp8', 'w4a8_int8_linear']}
[INFO] Checkpoint files will always be loaded safely.
[INFO] Total VRAM 122880 MB, total RAM 126429 MB
[INFO] pytorch version: 2.15.0.dev20260827+rocm7.2
[WARNING] WARNING[XFORMERS]: xFormers can't load C++/CUDA extensions. xFormers was built for:
PyTorch 2.7.1+rocm7.2.3.git1dab218d with CUDA None (you have 2.15.0.dev20260827+rocm7.2)
Python 3.12.13 (you have 3.12.3)
Please reinstall xformers (see https://github.com/facebookresearch/xformers#installing-xformers)
Memory-efficient attention, SwiGLU, sparse and more won't be available.
Set XFORMERS_MORE_DETAILS=1 for more details
[aiter] import [module_aiter_core] under /home/aadmin/ComfyUI/venvnew/lib/python3.12/site-packages/aiter/jit/module_aiter_core.so
/home/aadmin/ComfyUI/venvnew/lib/python3.12/site-packages/aiter/ops/deepgemm.py:22: RuntimeWarning: aiter.ops.opus (a16w16) is gfx950-only; detected arch='gfx1100'. opus_gemm_* calls will raise RuntimeError at invocation. opus_gemm uses gfx950-only intrinsics (MFMA, ds_read_b64_tr) and the 160 KiB LDS budget. Set GPU_ARCHS=gfx950 (or run on a gfx950 device) to use this module.
from .opus.gemm_op_a16w16 import opus_gemm_a16w16_tune as _opus_tune
[INFO] Set: torch.backends.cudnn.enabled = False for better AMD performance.
[INFO] AMD arch: gfx1100
[INFO] ROCm version: (7, 2)
[INFO] Set vram state to: HIGH_VRAM
[INFO] Device: cuda:0 AMD Radeon 8060S : native
[INFO] Using async weight offloading with 2 streams
[INFO] Enabled pinned memory 113785.0
[INFO] Using Flash Attention
[INFO] Python version: 3.12.3 (main, Aug 31 2026, 10:18:26) [GCC 13.3.0]
[INFO] ComfyUI version: 0.34.2
[INFO] comfy-aimdo version: 0.4.15
[INFO] comfy-kitchen version: 0.2.31
[INFO] comfyui-frontend-package version: 1.49.6
[INFO] comfyui-workflow-templates version: 0.11.50
[INFO] comfyui-embedded-docs version: 0.5.10
[INFO] comfy-kitchen version: 0.2.31
[INFO] comfy-aimdo version: 0.4.15
[INFO] [Prompt Server] web root: /home/aadmin/ComfyUI/venvnew/lib/python3.12/site-packages/comfyui_frontend_package/static
[INFO] Asset seeder disabled
[INFO] No OpenGL_accelerate module loaded: No module named 'OpenGL_accelerate'
Adding /home/aadmin/ComfyUI/custom_nodes to sys.path
Could not find efficiency nodes
Could not find comfyui_controlnet_aux nodes, AV_ControlNetPreprocessor will not work. Please install comfyui_controlnet_aux first
Could not find AdvancedControlNet nodes
Could not find AnimateDiff nodes
Loaded IPAdapter nodes from /home/aadmin/ComfyUI/custom_nodes/comfyui_ipadapter_plus
Could not find VideoHelperSuite nodes
Could not load ImpactPack nodes Could not find ImpactPack nodes
[INFO] ### Loading: ComfyUI-Manager (V3.41)
[INFO] [ComfyUI-Manager] network_mode: public
[INFO] [ComfyUI-Manager] ComfyUI per-queue preview override detected (PR #11261). Manager's preview method feature is disabled. Use ComfyUI's --preview-method CLI option or 'Settings > Execution > Live preview method'.
[INFO] ### ComfyUI Revision: 5830 [169fcf35] *DETACHED | Released on '2026-08-27'

  ＡＦ  －  ＣｏｍｆｙＵＩ  Ｎｏｄｅｓ
                                 
   🚀 AF - Prompt Nodes Pack Loaded!

[krea2edit] nodes v1.2.5 loaded
[INFO] [ComfyUI-Manager] default cache updated: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/model-list.json
[INFO] [ComfyUI-Manager] default cache updated: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/alter-list.json
[INFO] [ComfyUI-Manager] default cache updated: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/github-stats.json
[INFO] [ComfyUI-Manager] default cache updated: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/extension-node-map.json
[INFO] [ComfyUI-Manager] default cache updated: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/custom-node-list.json
WAS Node Suite: OpenCV Python FFMPEG support is enabled
WAS Node Suite Warning: ffmpeg_bin_path is not set in /home/aadmin/ComfyUI/custom_nodes/was-ns/was_suite_config.json config file. Will attempt to use system ffmpeg binaries if available.
WAS Node Suite: Finished. Loaded 220 nodes successfully.

    "Believe you deserve it and the universe will serve it." - Unknown

AMD GPU Monitor thread started
Using AMD SMI tool: /opt/rocm/bin/rocm-smiAMD GPU Monitor: Web directory set to /home/aadmin/ComfyUI/custom_nodes/amdgpumonitor/web

Adaptive LoRA Scheduler Node: Loaded

  ＡＦ  －  ＣｏｍｆｙＵＩ  Ｎｏｄｅｓ
                                 
 🔍 AF - Find Nodes Extension Loaded !

Use Ctrl+Shift+F to open the search panel

[INFO] ComfyUI-GGUF: Allowing full torch compile
[INFO] Using Flash Attention
(RES4LYF) Init
(RES4LYF) Importing beta samplers.
(RES4LYF) Importing legacy samplers.
Module 'diffusers' load failed. If you don't have it installed, do it:
pip install diffusers
[ComfyUI-Easy-Use] server: v1.3.6 Loaded
[ComfyUI-Easy-Use] web root: /home/aadmin/ComfyUI/custom_nodes/comfyui-easy-use/web_version/v2 Loaded
[PainterNodes] Failed to import PainterVRAM: NVML Shared Library Not Found
[PainterNodes] Loaded 35 nodes successfully!

😺dzNodes: LayerStyle -> Cannot import name 'guidedFilter' from 'cv2.ximgproc'

A few nodes cannot works properly, while most nodes are not affected. Please REINSTALL package 'opencv-contrib-python'.
For detail refer to https://github.com/chflame163/ComfyUI_LayerStyle/issues/5

[rgthree-comfy] Loaded 48 extraordinary nodes. 🎉

[rgthree-comfy] ComfyUI's new Node 2.0 rendering may be incompatible with some rgthree-comfy nodes and features, breaking some rendering as well as losing the ability to access a node's properties (a vital part of many nodes). It also appears to run MUCH more slowly spiking CPU usage and causing jankiness and unresponsiveness, especially with large workflows. Personally I am not planning to use the new Nodes 2.0 and, unfortunately, am not able to invest the time to investigate and overhaul rgthree-comfy where needed. If you have issues when Nodes 2.0 is enabled, I'd urge you to switch it off as well and join me in hoping ComfyUI is not planning to deprecate the existing, stable canvas rendering all together.

[INFO]
Import times for custom nodes:
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/websocket_image_save.py
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/text_adapter
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/comfyui-af-find-nodes
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/krea2-twostage-sampler
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/comfyui-qwenmultiangle
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/comfyui-krea2edit
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/comfyui-af-pack-prompt-nodes
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/comfyui-mxtoolkit
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/ltx2.5-msr
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/comfyui_ipadapter_plus
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/ComfyUI-Krea2T-Enhancer
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/ComfyUI-PromptRelay
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/amdgpumonitor
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/ComfyUI-GGUF
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/ComfyUI-Dynamic-Lora-Scheduler
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/ComfyMath
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/comfyui-various
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/vnccs
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/vnccs-utils
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/10s-comfy-nodes
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/rgthree-comfy
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/comfyui-painternodes
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/ComfyUI-PromptEnhancer
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/comfyui-animatediff-evolved
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/ComfyUI-LTXVideo
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/comfyui-art-venture
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/comfyui-kjnodes
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/comfyui-logicutils
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/comfyui-easy-use
[INFO] 0.0 seconds: /home/aadmin/ComfyUI/custom_nodes/comfyui-deno-custom-nodes
[INFO] 0.1 seconds: /home/aadmin/ComfyUI/custom_nodes/comfyui_layerstyle
[INFO] 0.1 seconds: /home/aadmin/ComfyUI/custom_nodes/ComfyUI-Manager
[INFO] 0.1 seconds: /home/aadmin/ComfyUI/custom_nodes/ComfyUI-WanVideoWrapper
[INFO] 0.2 seconds: /home/aadmin/ComfyUI/custom_nodes/comfyui-videohelpersuite
[INFO] 0.2 seconds: /home/aadmin/ComfyUI/custom_nodes/RES4LYF
[INFO] 0.5 seconds: /home/aadmin/ComfyUI/custom_nodes/was-ns
[INFO] 3.3 seconds: /home/aadmin/ComfyUI/custom_nodes/ltx2_sm
[INFO]
[INFO] Context impl SQLiteImpl.
[INFO] Will assume non-transactional DDL.
[INFO] Using RAM pressure cache.
[INFO] Starting server

[INFO] To see the GUI go to: http://0.0.0.0:8188
[INFO] To see the GUI go to: http://[::]:8188
[INFO] got prompt
[WARNING] [DENO] Resource Monitor could not initialize NVML: NVML Shared Library Not Found
[INFO] VAE load device: cuda:0, offload device: cpu, dtype: torch.float32
[INFO] model weight dtype torch.float8_e5m2, manual cast: torch.bfloat16
[INFO] model_type FLUX
[INFO] [LTX MSR-AVref] Loaded LTX-2.5-Licon-MSR-V2.safetensors with image/audio slot embeddings (5/5 tensors)
[INFO] [LTX MSR-AVref] image_dim=128 audio_dim=128 downscale=1 audio_layout=absolute_image_slot_windows duration=5.000s margin=0.040s max_audio_tokens=125
[INFO] VAE load device: cuda:0, offload device: cpu, dtype: torch.bfloat16
[INFO] Requested to load LTXAVTEModel_
[INFO] loaded completely; 24999.98 MB loaded, full load: True
[INFO] CLIP/text encoder model load device: cuda:0, offload device: cpu, current: cuda:0, dtype: torch.float16
[INFO] [PromptRelay] Global: tokens [0:74] (74 tokens)
[INFO] [PromptRelay] Segment 0: tokens [74:268] (194 tokens)
[INFO] [PromptRelay] Latent: 47 frames, 390 tokens/frame, segments: [47]
[INFO] [LTX MSR] Guide start: mode=MSR, references=3, reference_frames=33, downscale=1, temporal_scale=1, target_latent=(1, 128, 47, 15, 26)
[INFO] Requested to load CausalDiffusionVAE
[INFO] loaded completely; 1403.92 MB loaded, full load: True
[WARNING] [LTX MSR][VERIFY] slot_embedding=APPLIED label=pic1 slot_id=1 operation=guide_latent_plus_broadcast_embedding embedding_dim=128 embedding_norm=0.739934 embedding_preview=[0.010132, -0.057821, -0.10936, -0.010906, -0.044901, 0.034269, 0.023298, -0.006341]
[INFO] [LTX MSR] pic1: slot_embedding=applied, slot_id=1, embedding_dim=128, embedding_norm=0.739934, time_offset=-3, guide_latent=(1, 128, 5, 15, 26)
[WARNING] [LTX MSR][VERIFY] negative_time_offset=APPLIED label=pic1 slot_id=1 operation=append_keyframe frame_offset=-3 temporal_scale=1 compatibility_default=APPLIED temporal_position=negative_3 append_keyframe=SUCCESS
[WARNING] [LTX MSR][VERIFY] slot_embedding=APPLIED label=pic2 slot_id=2 operation=guide_latent_plus_broadcast_embedding embedding_dim=128 embedding_norm=0.767755 embedding_preview=[-0.015987, -0.037434, -0.089852, -0.035059, -0.07984, 0.021686, 0.01271, 0.018696]
[INFO] [LTX MSR] pic2: slot_embedding=applied, slot_id=2, embedding_dim=128, embedding_norm=0.767755, time_offset=-2, guide_latent=(1, 128, 5, 15, 26)
[WARNING] [LTX MSR][VERIFY] negative_time_offset=APPLIED label=pic2 slot_id=2 operation=append_keyframe frame_offset=-2 temporal_scale=1 compatibility_default=APPLIED temporal_position=negative_2 append_keyframe=SUCCESS
[WARNING] [LTX MSR][VERIFY] slot_embedding=APPLIED label=background slot_id=3 operation=guide_latent_plus_broadcast_embedding embedding_dim=128 embedding_norm=0.847053 embedding_preview=[-0.027266, -0.034561, -0.077652, -0.075744, -0.083365, 0.021973, 0.021777, 0.014609]
[INFO] [LTX MSR] background: slot_embedding=applied, slot_id=3, embedding_dim=128, embedding_norm=0.847053, time_offset=-1, guide_latent=(1, 128, 5, 15, 26)
[WARNING] [LTX MSR][VERIFY] negative_time_offset=APPLIED label=background slot_id=3 operation=append_keyframe frame_offset=-1 temporal_scale=1 compatibility_default=APPLIED temporal_position=negative_1 append_keyframe=SUCCESS
[INFO] [LTX MSR] Guide complete: added=3, order=pic1..pic3, frames_each=33, mode=MSR, output_latent=(1, 128, 62, 15, 26)
[INFO] Requested to load LTXAV
[INFO] loaded completely; 20025.73 MB loaded, full load: True
0%| | 0/8 [00:00<?, ?it/s][INFO] [PromptRelay] Built penalty matrix (scaled): Lq=24180, Lk=1024, nonzero=498968/24760320
[INFO] [PromptRelay] Built penalty matrix (scaled): Lq=376, Lk=1024, nonzero=7566/385024
12%|█▎ | 1/8 [01:36<11:17, 96.84s/it]FETCH ComfyRegistry Data [DONE]
[INFO] [ComfyUI-Manager] default cache updated: https://api.comfy.org/nodes
FETCH DATA from: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/custom-node-list.json [DONE]
[INFO] [ComfyUI-Manager] All startup tasks have been completed.
100%|██████████| 8/8 [12:31<00:00, 93.95s/it]
[INFO] Requested to load AudioVAE
[INFO] loaded completely; 693.46 MB loaded, full load: True
[INFO] Requested to load LatentUpsampler
[INFO] loaded completely; 949.61 MB loaded, full load: True
[INFO] [PromptRelay] Global: tokens [0:74] (74 tokens)
[INFO] [PromptRelay] Segment 0: tokens [74:268] (194 tokens)
[INFO] [PromptRelay] Latent: 47 frames, 1560 tokens/frame, segments: [47]
[INFO] [LTX MSR] Guide start: mode=MSR, references=3, reference_frames=33, downscale=1, temporal_scale=1, target_latent=(1, 128, 47, 30, 52)
[W1009 00:13:12.132929361 HIPCachingAllocator.cpp:4072] memory allocation failed with OOM on device 0 while trying to allocate 5521801216 bytes (free: 4967206912, total: 128849018880).
[WARNING] [LTX MSR][VERIFY] slot_embedding=APPLIED label=pic1 slot_id=1 operation=guide_latent_plus_broadcast_embedding embedding_dim=128 embedding_norm=0.739934 embedding_preview=[0.010132, -0.057821, -0.10936, -0.010906, -0.044901, 0.034269, 0.023298, -0.006341]
[INFO] [LTX MSR] pic1: slot_embedding=applied, slot_id=1, embedding_dim=128, embedding_norm=0.739934, time_offset=-3, guide_latent=(1, 128, 5, 30, 52)
[WARNING] [LTX MSR][VERIFY] negative_time_offset=APPLIED label=pic1 slot_id=1 operation=append_keyframe frame_offset=-3 temporal_scale=1 compatibility_default=APPLIED temporal_position=negative_3 append_keyframe=SUCCESS
[WARNING] [LTX MSR][VERIFY] slot_embedding=APPLIED label=pic2 slot_id=2 operation=guide_latent_plus_broadcast_embedding embedding_dim=128 embedding_norm=0.767755 embedding_preview=[-0.015987, -0.037434, -0.089852, -0.035059, -0.07984, 0.021686, 0.01271, 0.018696]
[INFO] [LTX MSR] pic2: slot_embedding=applied, slot_id=2, embedding_dim=128, embedding_norm=0.767755, time_offset=-2, guide_latent=(1, 128, 5, 30, 52)
[WARNING] [LTX MSR][VERIFY] negative_time_offset=APPLIED label=pic2 slot_id=2 operation=append_keyframe frame_offset=-2 temporal_scale=1 compatibility_default=APPLIED temporal_position=negative_2 append_keyframe=SUCCESS
[WARNING] [LTX MSR][VERIFY] slot_embedding=APPLIED label=background slot_id=3 operation=guide_latent_plus_broadcast_embedding embedding_dim=128 embedding_norm=0.847053 embedding_preview=[-0.027266, -0.034561, -0.077652, -0.075744, -0.083365, 0.021973, 0.021777, 0.014609]
[INFO] [LTX MSR] background: slot_embedding=applied, slot_id=3, embedding_dim=128, embedding_norm=0.847053, time_offset=-1, guide_latent=(1, 128, 5, 30, 52)
[WARNING] [LTX MSR][VERIFY] negative_time_offset=APPLIED label=background slot_id=3 operation=append_keyframe frame_offset=-1 temporal_scale=1 compatibility_default=APPLIED temporal_position=negative_1 append_keyframe=SUCCESS
[INFO] [LTX MSR] Guide complete: added=3, order=pic1..pic3, frames_each=33, mode=MSR, output_latent=(1, 128, 62, 30, 52)
[INFO] Requested to load LTXAV
0%| | 0/3 [00:00<?, ?it/s][INFO] [PromptRelay] Built penalty matrix (scaled): Lq=96720, Lk=1024, nonzero=1996066/99041280
[INFO] [PromptRelay] Built penalty matrix (scaled): Lq=376, Lk=1024, nonzero=7566/385024
100%|██████████| 3/3 [34:15<00:00, 685.25s/it]
[INFO] Prompt executed in 00:57:11

I just tried the "default" type again and it filled the GTT up to 92284MB and got stuck:
92284M / 131054M GTT 70.42% ³

log:
[INFO] got prompt
[INFO] VAE load device: cuda:0, offload device: cpu, dtype: torch.float32

[INFO] model weight dtype torch.bfloat16, manual cast: None
[INFO] model_type FLUX
[INFO] [LTX MSR-AVref] Loaded LTX-2.5-Licon-MSR-V2.safetensors with image/audio slot embeddings (5/5 tensors)
[INFO] [LTX MSR-AVref] image_dim=128 audio_dim=128 downscale=1 audio_layout=absolute_image_slot_windows duration=5.000s margin=0.040s max_audio_tokens=125
[INFO] VAE load device: cuda:0, offload device: cpu, dtype: torch.bfloat16
[INFO] Requested to load LTXAVTEModel_
[INFO] loaded completely; 24999.98 MB loaded, full load: True
[INFO] CLIP/text encoder model load device: cuda:0, offload device: cpu, current: cuda:0, dtype: torch.float16
[INFO] [PromptRelay] Global: tokens [0:74] (74 tokens)
[INFO] [PromptRelay] Segment 0: tokens [74:268] (194 tokens)
[INFO] [PromptRelay] Latent: 47 frames, 390 tokens/frame, segments: [47]
[INFO] [LTX MSR] Guide start: mode=MSR, references=3, reference_frames=33, downscale=1, temporal_scale=1, target_latent=(1, 128, 47, 15, 26)
[INFO] Requested to load CausalDiffusionVAE
[INFO] loaded completely; 1403.92 MB loaded, full load: True
[WARNING] [LTX MSR][VERIFY] slot_embedding=APPLIED label=pic1 slot_id=1 operation=guide_latent_plus_broadcast_embedding embedding_dim=128 embedding_norm=0.739934 embedding_preview=[0.010132, -0.057821, -0.10936, -0.010906, -0.044901, 0.034269, 0.023298, -0.006341]
[INFO] [LTX MSR] pic1: slot_embedding=applied, slot_id=1, embedding_dim=128, embedding_norm=0.739934, time_offset=-3, guide_latent=(1, 128, 5, 15, 26)
[WARNING] [LTX MSR][VERIFY] negative_time_offset=APPLIED label=pic1 slot_id=1 operation=append_keyframe frame_offset=-3 temporal_scale=1 compatibility_default=APPLIED temporal_position=negative_3 append_keyframe=SUCCESS
[WARNING] [LTX MSR][VERIFY] slot_embedding=APPLIED label=pic2 slot_id=2 operation=guide_latent_plus_broadcast_embedding embedding_dim=128 embedding_norm=0.767755 embedding_preview=[-0.015987, -0.037434, -0.089852, -0.035059, -0.07984, 0.021686, 0.01271, 0.018696]
[INFO] [LTX MSR] pic2: slot_embedding=applied, slot_id=2, embedding_dim=128, embedding_norm=0.767755, time_offset=-2, guide_latent=(1, 128, 5, 15, 26)
[WARNING] [LTX MSR][VERIFY] negative_time_offset=APPLIED label=pic2 slot_id=2 operation=append_keyframe frame_offset=-2 temporal_scale=1 compatibility_default=APPLIED temporal_position=negative_2 append_keyframe=SUCCESS
[WARNING] [LTX MSR][VERIFY] slot_embedding=APPLIED label=background slot_id=3 operation=guide_latent_plus_broadcast_embedding embedding_dim=128 embedding_norm=0.847053 embedding_preview=[-0.027266, -0.034561, -0.077652, -0.075744, -0.083365, 0.021973, 0.021777, 0.014609]
[INFO] [LTX MSR] background: slot_embedding=applied, slot_id=3, embedding_dim=128, embedding_norm=0.847053, time_offset=-1, guide_latent=(1, 128, 5, 15, 26)
[WARNING] [LTX MSR][VERIFY] negative_time_offset=APPLIED label=background slot_id=3 operation=append_keyframe frame_offset=-1 temporal_scale=1 compatibility_default=APPLIED temporal_position=negative_1 append_keyframe=SUCCESS
[INFO] [LTX MSR] Guide complete: added=3, order=pic1..pic3, frames_each=33, mode=MSR, output_latent=(1, 128, 62, 15, 26)
[INFO] Requested to load LTXAV
[INFO] Unloaded partially: 13674.97 MB freed, 11341.42 MB remains loaded, 562.50 MB buffer reserved, lowvram patches: 0
FETCH ComfyRegistry Data [DONE]
[INFO] [ComfyUI-Manager] default cache updated: https://api.comfy.org/nodes
FETCH DATA from: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/custom-node-list.json [DONE]
[INFO] [ComfyUI-Manager] All startup tasks have been completed.
[INFO] loaded completely; 40051.47 MB loaded, full load: True
0%| | 0/8 [00:00<?, ?it/s][INFO] [PromptRelay] Built penalty matrix (scaled): Lq=24180, Lk=1024, nonzero=498968/24760320
[INFO] [PromptRelay] Built penalty matrix (scaled): Lq=376, Lk=1024, nonzero=7566/385024
100%|██████████| 8/8 [08:35<00:00, 64.39s/it]
[INFO] Requested to load AudioVAE
[INFO] loaded completely; 693.46 MB loaded, full load: True
[INFO] Requested to load CausalDiffusionVAE

I tested using the same inputs you provided and didn’t encounter the issues you mentioned. Could your reference images be in the wrong order?
ilkeraktuna

no , I am sure about that.
and even if the images were in wrong order , the characters would be reversed. Now I have totally different identities and with additional characters.
Also I attached the workflow I used. If images were in wrong order, you could say that.

Did you try with the diffusion dtype I wrote above or did you use the default dtype ?
ilkeraktuna

btw, I updated my rocm setup to rocm 10 now and tried the "default" dtype again.

now it hangs at:

100%|██████████| 8/8 [07:41<00:00, 57.63s/it]
[INFO] Requested to load AudioVAE
[INFO] loaded completely; 693.46 MB loaded, full load: True
[INFO] Requested to load CausalDiffusionVAE

Sign up or log in to comment