← back to catalog · registered 2026-10-10 12:58

NidAll/Qwen3VL-8B-Abliterated-INT8-ConvRot

NidAll Qwen 8B GGUF image-gen second-order
Your rig guess connected
? Why do I need an app?
Reading your rig…

This is a rough estimate. Install the free app - we'll show exact numbers.

Reading real hardware from your app right now. Numbers below are exact.

Below is the per-quantization compatibility for this model.

curl -H "Authorization: Bearer $ABL_KEY" \
     "https://abliteration.org/api/v1/models/NidAll%2FQwen3VL-8B-Abliterated-INT8-ConvRot"
Response includes
  • classification unknown
  • files 4
  • author_summary 2 models
  • readme_text full
10 credits · hourly refresh · ~4 KB payload Get an API key →
Abliteration classifier · v1.0.0
?
Primary method

Unclassified

No clear signals of an abliteration technique in this model.
Confidence
UNKNOWN
Why this label 1 signal
No classification signals present. This may not be an abliterated model at all - it could be a repackaging, a merge with unrelated goals, or unrelated content that mentions the term.
  • no classification signals present (no abliterated, uncensored, or known producer/method markers)
Refusal direction extraction

No specific extraction method could be identified for this model. The producer either did not document it or used a proprietary pipeline.

What is a refusal direction? →
Downloads · 30-day
0
Likes
0
Model age
today
created 2026-10-10

Genealogy 0 direct forks

Full fork graph →

This model's place in the market. Above: what it was derived from. Below: the tree of everything derived from it.

Metadata

License
apache-2.0
Quantizations
BF16
Tags
diffusion-single-file comfyui qwen3-vl qwen-image-2.1 text-encoder bf16 int8 convrot abliteration not-for-all-audiences text-to-image base_model:AtomicChat/Qwen-Image-2.1-Turbo-Uncensored-GGUF

Related

Total size
25.0 GB
Files
4
Quantizations
2
Registered
2026-10-10 12:58
Last updated on HF
2026-10-10 12:52

Files by quantization

BF16 1 file 16.3 GB
qwen3vl_8b_abliterated_bf16.safetensors 16.3 GB e9e4e7f0 download
Auxiliary files 3 files 8.71 GB
qwen3vl_8b_abliterated_int8_convrot.safetensors 8.71 GB 34153e12 download
README.md 21.6 KB 46cb271d download
.gitattributes 3.35 KB b44889b3 download

README current version from Hugging Face


license: apache-2.0
base_model:

  • AtomicChat/Qwen-Image-2.1-Turbo-Uncensored-GGUF
  • Comfy-Org/Qwen3-VL
  • Qwen/Qwen3-VL-8B-Instruct
    base_model_relation: quantized
    library_name: diffusion-single-file
    pipeline_tag: text-to-image
    tags:
  • comfyui
  • qwen3-vl
  • qwen-image-2.1
  • text-encoder
  • bf16
  • int8
  • convrot
  • abliteration
  • not-for-all-audiences

Qwen3-VL 8B Abliterated · INT8 ConvRot & BF16

A complete multimodal Qwen3-VL text encoder for Qwen Image 2.1 Turbo in ComfyUI, reconstructed from AtomicChat's BF16 GGUF release. Choose the 9.35 GB INT8 ConvRot checkpoint for reduced storage, or 17.53 GB BF16 for the original reconstructed precision.

Both include the entire BF16 vision tower and image-to-language mergers needed for image conditioning. No separate mmproj is required. The source modification affects the language encoder; the diffusion transformer and VAE are unchanged.

Choose your checkpoint

Checkpoint Download Exact size Precision
INT8 ConvRot qwen3vl_8b_abliterated_int8_convrot.safetensors 9,350,798,360 bytes · 9.35 GB INT8 language matrices; BF16 vision and norms
BF16 qwen3vl_8b_abliterated_bf16.safetensors 17,534,334,976 bytes · 17.53 GB Complete BF16 reconstruction

The INT8 checkpoint matches the official Comfy-Org encoder's exact file size and tensor precision policy. It is 46.67% smaller than BF16. Each file is a complete encoder; download one for ordinary use.

Quick start in ComfyUI

  1. Download your chosen checkpoint into ComfyUI/models/text_encoders/.
  2. In CLIPLoader, select the checkpoint and set type to qwen_image.
  3. Connect the CLIP output to TextEncodeQwenImage21 in a Qwen Image 2.1 Turbo workflow.
  4. Use the matching diffusion model and VAE below. For image editing, also connect your reference image and VAE to TextEncodeQwenImage21, then use its latent output as the sampler's input.

CLI download, from the directory containing your ComfyUI installation:

hf download NidAll/Qwen3VL-8B-Abliterated-INT8-ConvRot \
  qwen3vl_8b_abliterated_int8_convrot.safetensors \
  --local-dir ComfyUI/models/text_encoders

To download BF16 instead, substitute qwen3vl_8b_abliterated_bf16.safetensors.

Required companion models

Download these separately from Comfy-Org/Qwen-Image-2.1:

Component Tested file ComfyUI folder
Diffusion model qwen_image_2.1_turbo_bf16.safetensors models/diffusion_models/
VAE qwen_image_2.1_vae_bf16.safetensors models/vae/

For the tested Turbo setup, use Euler, CFG 1, eight steps, with these exact sigmas:

1.0, 0.978453, 0.95418, 0.926626, 0.89508,
0.845148, 0.704534, 0.414568, 0.0

Image-editing tests used the same reference image across encoder variants. Use the reference image's dimensions and preserve the reference conditioning produced by the encoder node.

The tested ComfyUI revision is 0df64eb242b7c5759c3e86afd5d1846d923b1033, with comfy-kitchen==0.2.37. An older ComfyUI build may lack Qwen Image 2.1 support or the required inline INT8 ConvRot/embedding loader.

What is included

Language: all 36 transformer blocks, token embeddings, normalization tensors and the language output head. Exactly 73 tensors differ from the original BF16 reference: 36 attention-output projections, 36 MLP down-projections and the token embedding.

Vision: the complete 27-block tower, patch and positional embeddings, visual attention/MLPs, normalization, the main merger and three DeepStack mergers. All 351 final vision tensors remain BF16 and were verified numerically against the BF16 mmproj. The GGUF's two temporal patch-convolution slices were correctly reassembled.

The reconstructed checkpoint contains 750 BF16 tensors. GGUF BF16 bits were extracted directly, without an FP16 intermediate. GGUF FP32 tensors were reused in BF16 only after exact equality with the widened reference was established. All 677 unchanged reference tensors are preserved bit for bit.

INT8 ConvRot policy

Setting Value
Native storage format int8_tensorwise
Quantized matrices 254: 252 attention/MLP projections, token embeddings and lm_head
Transformer coverage All language blocks 0–35
Rotation Regular orthogonal H4 ConvRot, group 256
Scaling FP32 per output row
Conversion arithmetic FP32, bounded butterfly transform
BF16 preservation Complete vision/mergers, norms, biases and remaining official BF16 tensors
Weight error gate Relative L2 ≤ 0.012 for every selected matrix

Conversion started from the validated BF16 checkpoint, using the ComfyUI Native Quantizer codec and a pinned official tensor-layout template. No activation calibration, learned rounding, rank reduction or double quantization was applied. Every selected matrix passed the gate; there are no BF16 boundary-block exclusions.

The final file has 1,258 stored tensors: 254 INT8 weights, 496 BF16 tensors, 254 FP32 scales and 254 inline quantization metadata tensors. 93.42% of model parameter elements are INT8.

Storage and execution

Stock ComfyUI CLIP encoding can retain INT8 storage while performing floating-point matrix multiplication after inverse rotation. The filename does not imply automatic INT8 GEMM acceleration. The quantized embeddings use native token-row dequantization with inverse ConvRot.

A separate test explicitly enabled native quantized operations and disabled forced floating casts in runtime test objects. It observed 504 CUDA INT8 linear calls per generation, plus native embedding dequantization and INT8 tensor-core GEMM kernels. That experimental compute mode is not automatically enabled by ordinary CLIPLoader use.

Validation results

Reconstruction passed complete tensor coverage, serialized identity, tokenizer and NaN/Inf checks. Source reconstruction errors were zero; all 151,669 token IDs and 151,387 BPE merges matched. Independent comfy-kitchen decoding of all INT8 matrices produced:

Weight metric Result
Mean per-matrix relative L2 error 0.00887948
Maximum per-matrix relative L2 error 0.01020596
Minimum per-matrix cosine similarity 0.99994792

16 fresh INT8 generations—eight stock-compute and eight native-INT8-compute runs—were compared against eight unchanged BF16 baseline images. All used 1024-pixel output, seed 42, identical prompts/reference image, diffusion model, VAE and sampling settings. Every image edit exercised the vision tower and three DeepStack mergers; all conditioning, latent and decoded-image tensors were finite.

Matched image similarity results

SSIM measures image similarity, not prompt adherence or semantic correctness.

Case Stock INT8 SSIM Native INT8 compute SSIM
Still life 0.9900 0.9839
Long spatial prompt 0.7676 0.9087
English typography 0.9770 0.9914
Chinese typography 0.9825 0.9867
Ordered objects 0.9554 0.9569
Object replacement 0.9972 0.9967
Style edit 0.9983 0.9984
Lighting/structure edit 0.9981 0.9980

Manual inspection confirmed readable English/Chinese text, the requested object order and successful teapot-to-watering-can replacement in these examples. The library scene showed placement/detail differences. Watercolor and dusk changes were weak in BF16 and both INT8 modes: those cases passed execution but did not pass semantic transformation acceptance.

These measurements cover eight prompts at one seed. They do not establish universal quality equivalence, robust typography, broad style editing, multi-reference behavior, chat generation quality or refusal reduction. Weight error is not activation error. No portable latency or memory-capacity claim is made.

Sources and attribution

Source Pinned revision Role
AtomicChat/Qwen-Image-2.1-Turbo-Uncensored-GGUF e3b01e9bfacc373148bdb93824496e0c72643250 Modified BF16 language GGUF and BF16 mmproj
Comfy-Org/Qwen3-VL 02f0d3eefb59528799653d16735818801177ef1f text_encoders/qwen3vl_8b_bf16.safetensors structural reference
Comfy-Org/Qwen-Image-2.1 df94239739eef7205973a145ffa6f441f7b64e84 Official INT8 policy/layout and tested diffusion/VAE

AtomicChat sources: Qwen-Image-2.1-Turbo-Abliterated-Uncensored-BF16.gguf and mmproj-Qwen-Image-2.1-Turbo-Abliterated-Uncensored-BF16.gguf. The original architecture and weights originate from Qwen/Qwen3-VL-8B-Instruct.

Credit to Qwen, AtomicChat, Comfy-Org, the ComfyUI Native Quantizer and comfy-kitchen projects for the original weights, modifications, compatible layout and conversion/runtime implementations. The silveroxides/convert_to_quant project was researched for format compatibility; its learned-rounding converter did not produce these files.

The source's Not For All Audiences classification is retained. “Abliterated” identifies the inherited language-weight modification; no image uncensoring effect or new safety/refusal behavior has been established by this release.

SHA-256 checksums

34153e12be478cac8debc0619c6c1d4f4dd53cd42c333346585e268d5ca7dde3  qwen3vl_8b_abliterated_int8_convrot.safetensors
e9e4e7f0c046ac95749628a50234faf1fd3334a62a7a34273e6c578464095cbc  qwen3vl_8b_abliterated_bf16.safetensors

License and notices

The Qwen3-VL weights and AtomicChat-derived encoder are distributed under Apache-2.0. This distribution reconstructs the 73 modified language tensors in the Comfy-Org checkpoint layout and supplies an INT8 ConvRot derivative. Both include the complete vision tower. The diffusion model and VAE are separate dependencies and are not redistributed here.

ComfyUI and comfy-kitchen retain their own licenses. Conversion used the Apache-2.0 ComfyUI Native Quantizer backend; conversion software is not bundled in this model-only repository.

Apache License 2.0 — full text
Apache License
                           Version 2.0, January 2004
                        http://www.apache.org/licenses/

   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION

   1. Definitions.

      "License" shall mean the terms and conditions for use, reproduction,
      and distribution as defined by Sections 1 through 9 of this document.

      "Licensor" shall mean the copyright owner or entity authorized by
      the copyright owner that is granting the License.

      "Legal Entity" shall mean the union of the acting entity and all
      other entities that control, are controlled by, or are under common
      control with that entity. For the purposes of this definition,
      "control" means (i) the power, direct or indirect, to cause the
      direction or management of such entity, whether by contract or
      otherwise, or (ii) ownership of fifty percent (50%) or more of the
      outstanding shares, or (iii) beneficial ownership of such entity.

      "You" (or "Your") shall mean an individual or Legal Entity
      exercising permissions granted by this License.

      "Source" form shall mean the preferred form for making modifications,
      including but not limited to software source code, documentation
      source, and configuration files.

      "Object" form shall mean any form resulting from mechanical
      transformation or translation of a Source form, including but
      not limited to compiled object code, generated documentation,
      and conversions to other media types.

      "Work" shall mean the work of authorship, whether in Source or
      Object form, made available under the License, as indicated by a
      copyright notice that is included in or attached to the work
      (an example is provided in the Appendix below).

      "Derivative Works" shall mean any work, whether in Source or Object
      form, that is based on (or derived from) the Work and for which the
      editorial revisions, annotations, elaborations, or other modifications
      represent, as a whole, an original work of authorship. For the purposes
      of this License, Derivative Works shall not include works that remain
      separable from, or merely link (or bind by name) to the interfaces of,
      the Work and Derivative Works thereof.

      "Contribution" shall mean any work of authorship, including
      the original version of the Work and any modifications or additions
      to that Work or Derivative Works thereof, that is intentionally
      submitted to Licensor for inclusion in the Work by the copyright owner
      or by an individual or Legal Entity authorized to submit on behalf of
      the copyright owner. For the purposes of this definition, "submitted"
      means any form of electronic, verbal, or written communication sent
      to the Licensor or its representatives, including but not limited to
      communication on electronic mailing lists, source code control systems,
      and issue tracking systems that are managed by, or on behalf of, the
      Licensor for the purpose of discussing and improving the Work, but
      excluding communication that is conspicuously marked or otherwise
      designated in writing by the copyright owner as "Not a Contribution."

      "Contributor" shall mean Licensor and any individual or Legal Entity
      on behalf of whom a Contribution has been received by Licensor and
      subsequently incorporated within the Work.

   2. Grant of Copyright License. Subject to the terms and conditions of
      this License, each Contributor hereby grants to You a perpetual,
      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
      copyright license to reproduce, prepare Derivative Works of,
      publicly display, publicly perform, sublicense, and distribute the
      Work and such Derivative Works in Source or Object form.

   3. Grant of Patent License. Subject to the terms and conditions of
      this License, each Contributor hereby grants to You a perpetual,
      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
      (except as stated in this section) patent license to make, have made,
      use, offer to sell, sell, import, and otherwise transfer the Work,
      where such license applies only to those patent claims licensable
      by such Contributor that are necessarily infringed by their
      Contribution(s) alone or by combination of their Contribution(s)
      with the Work to which such Contribution(s) was submitted. If You
      institute patent litigation against any entity (including a
      cross-claim or counterclaim in a lawsuit) alleging that the Work
      or a Contribution incorporated within the Work constitutes direct
      or contributory patent infringement, then any patent licenses
      granted to You under this License for that Work shall terminate
      as of the date such litigation is filed.

   4. Redistribution. You may reproduce and distribute copies of the
      Work or Derivative Works thereof in any medium, with or without
      modifications, and in Source or Object form, provided that You
      meet the following conditions:

      (a) You must give any other recipients of the Work or
          Derivative Works a copy of this License; and

      (b) You must cause any modified files to carry prominent notices
          stating that You changed the files; and

      (c) You must retain, in the Source form of any Derivative Works
          that You distribute, all copyright, patent, trademark, and
          attribution notices from the Source form of the Work,
          excluding those notices that do not pertain to any part of
          the Derivative Works; and

      (d) If the Work includes a "NOTICE" text file as part of its
          distribution, then any Derivative Works that You distribute must
          include a readable copy of the attribution notices contained
          within such NOTICE file, excluding those notices that do not
          pertain to any part of the Derivative Works, in at least one
          of the following places: within a NOTICE text file distributed
          as part of the Derivative Works; within the Source form or
          documentation, if provided along with the Derivative Works; or,
          within a display generated by the Derivative Works, if and
          wherever such third-party notices normally appear. The contents
          of the NOTICE file are for informational purposes only and
          do not modify the License. You may add Your own attribution
          notices within Derivative Works that You distribute, alongside
          or as an addendum to the NOTICE text from the Work, provided
          that such additional attribution notices cannot be construed
          as modifying the License.

      You may add Your own copyright statement to Your modifications and
      may provide additional or different license terms and conditions
      for use, reproduction, or distribution of Your modifications, or
      for any such Derivative Works as a whole, provided Your use,
      reproduction, and distribution of the Work otherwise complies with
      the conditions stated in this License.

   5. Submission of Contributions. Unless You explicitly state otherwise,
      any Contribution intentionally submitted for inclusion in the Work
      by You to the Licensor shall be under the terms and conditions of
      this License, without any additional terms or conditions.
      Notwithstanding the above, nothing herein shall supersede or modify
      the terms of any separate license agreement you may have executed
      with Licensor regarding such Contributions.

   6. Trademarks. This License does not grant permission to use the trade
      names, trademarks, service marks, or product names of the Licensor,
      except as required for reasonable and customary use in describing the
      origin of the Work and reproducing the content of the NOTICE file.

   7. Disclaimer of Warranty. Unless required by applicable law or
      agreed to in writing, Licensor provides the Work (and each
      Contributor provides its Contributions) on an "AS IS" BASIS,
      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
      implied, including, without limitation, any warranties or conditions
      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
      PARTICULAR PURPOSE. You are solely responsible for determining the
      appropriateness of using or redistributing the Work and assume any
      risks associated with Your exercise of permissions under this License.

   8. Limitation of Liability. In no event and under no legal theory,
      whether in tort (including negligence), contract, or otherwise,
      unless required by applicable law (such as deliberate and grossly
      negligent acts) or agreed to in writing, shall any Contributor be
      liable to You for damages, including any direct, indirect, special,
      incidental, or consequential damages of any character arising as a
      result of this License or out of the use or inability to use the
      Work (including but not limited to damages for loss of goodwill,
      work stoppage, computer failure or malfunction, or any and all
      other commercial damages or losses), even if such Contributor
      has been advised of the possibility of such damages.

   9. Accepting Warranty or Additional Liability. While redistributing
      the Work or Derivative Works thereof, You may choose to offer,
      and charge a fee for, acceptance of support, warranty, indemnity,
      or other liability obligations and/or rights consistent with this
      License. However, in accepting such obligations, You may act only
      on Your own behalf and on Your sole responsibility, not on behalf
      of any other Contributor, and only if You agree to indemnify,
      defend, and hold each Contributor harmless for any liability
      incurred by, or claims asserted against, such Contributor by reason
      of your accepting any such warranty or additional liability.

   END OF TERMS AND CONDITIONS

   APPENDIX: How to apply the Apache License to your work.

      To apply the Apache License to your work, attach the following
      boilerplate notice, with the fields enclosed by brackets "[]"
      replaced with your own identifying information. (Don't include
      the brackets!)  The text should be enclosed in the appropriate
      comment syntax for the file format. We also recommend that a
      file or class name and description of purpose be included on the
      same "printed page" as the copyright notice for easier
      identification within third-party archives.

   Copyright [yyyy] [name of copyright owner]

   Licensed under the Apache License, Version 2.0 (the "License");
   you may not use this file except in compliance with the License.
   You may obtain a copy of the License at

       http://www.apache.org/licenses/LICENSE-2.0

   Unless required by applicable law or agreed to in writing, software
   distributed under the License is distributed on an "AS IS" BASIS,
   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
   See the License for the specific language governing permissions and
   limitations under the License.
Catalog is the map. Apps are the tools.

Run models on your own machine, not in the cloud.

Every model page has an "Open in Abliteration" button that hands the model directly to the first-party desktop client, at the quantization your rig can actually run. No API keys, no subscription, no prompt leakage.

Open in Abliteration