Based on dealignai/DeepSeek-V4.1-Flash-UNCENSORED-FP8 The UNCENSORED / abliterated weights in this checkpoint come from dealignai. All credit for that work goes to them.
Not For All Audiences. This is an unofficial community-built checkpoint with reduced/removed refusal behavior inherited from an uncensored/abliterated weight source. It can generate unsafe, offensive, explicit, illegal, inaccurate, misleading, or otherwise objectionable content. Use appropriate access controls, moderation, sandboxing, and human review for your use case.
This repository contains an unofficial hybrid checkpoint derived from:
deepseek-ai/DeepSeek-V4.1-Flashnvidia/DeepSeek-V4.1-Flash-NVFP4dealignai/DeepSeek-V4.1-Flash-UNCENSORED-FP8The goal is to preserve NVIDIA's Blackwell-oriented NVFP4 representation for the routed experts while incorporating the weight-level UNCENSORED modifications from the dealignai checkpoint.
This repository is not an official release from DeepSeek, NVIDIA, dealignai, Hugging Face, Modal, or vLLM.
The build process uses the NVIDIA NVFP4 checkpoint as the base and replaces only the 54 weight tensors identified as modified in the UNCENSORED checkpoint:
layers.10..36.attn.wo_b.weightlayers.10..36.ffn.shared_experts.w2.weightThat is:
The routed MoE experts are not replaced by this merge and remain from the NVIDIA NVFP4 checkpoint.
The NVIDIA quantization configuration excludes the attention and shared-expert paths above from NVFP4 quantization, which is why the merge is performed directly on those source-precision weights rather than requantizing the entire model.
A build/patch manifest and checksums should be kept with the repository whenever possible so that users can verify provenance and integrity.
This is an unofficial merge. Do not assume that matching tensor shapes or successful loading proves behavioral equivalence or long-context stability.
Before relying on this checkpoint, independently validate at least:
Until those tests have been completed and documented, no claim is made that this merge has zero quality loss, zero degeneration, or identical behavior to either upstream checkpoint.
Potential uses include:
The "UNCENSORED" label describes the source checkpoint's reduced-refusal / abliterated behavior. It does not mean that outputs are correct, safe, lawful, or appropriate.
This checkpoint may produce content that other models would normally refuse or filter, including potentially dangerous or unlawful instructions.
It may also produce:
Do not treat model output as authoritative professional advice. High-impact or safety-critical uses require independent expert review and appropriate safeguards.
Operators are responsible for deciding what controls are appropriate for their deployment, including authentication, rate limits, logging, moderation, sandboxing, network isolation, and human review.
The checkpoint, weights, configuration, scripts, documentation, and related materials are provided "AS IS", without warranties of any kind, express or implied, to the maximum extent permitted by applicable law.
No representation or guarantee is made regarding:
Use and deployment are at your own risk.
To the maximum extent permitted by applicable law, the maintainer and contributors disclaim liability for claims, losses, damages, costs, or consequences arising from the use, misuse, modification, redistribution, hosting, or inability to use this repository or its contents.
Nothing in this notice excludes or limits liability that cannot lawfully be excluded or limited.
This section is a general project disclaimer, not legal advice.
Downloading or using this checkpoint does not transfer responsibility for a deployment to the maintainer.
Users and operators should independently ensure that their use:
Do not commit API keys, Hugging Face tokens, Modal credentials, or other secrets to this repository.
The Hugging Face repositories listed above currently identify the relevant model weights as MIT licensed. This repository does not grant any rights beyond those provided by the applicable upstream licenses.
Preserve all copyright, license, and attribution notices required by the upstream projects.
Credit: this checkpoint is based on dealignai/DeepSeek-V4.1-Flash-UNCENSORED-FP8. Any derivative of this repository (quantization, fine-tune, redistribution, benchmark, write-up) should also credit dealignai as the source of the UNCENSORED weights and link back to that repository.
In particular, consult the current upstream repositories before redistribution or commercial deployment:
Upstream licenses and terms control where applicable. This README does not replace them.
This project is independently assembled by a community maintainer.
It is not affiliated with, endorsed by, sponsored by, or officially supported by DeepSeek, NVIDIA, dealignai, Hugging Face, Modal, or vLLM.
Names and trademarks are used only to identify upstream projects and technical provenance.
If you redistribute this checkpoint or create a derivative:
For public-facing services, consider adding safeguards outside the model itself. Weight-level removal of refusal behavior means the model should not be assumed to provide a reliable safety boundary.
Use of this repository is subject to applicable law and the licenses/terms of the upstream components. This model card is informational and is not legal advice.
model-00048-of-00048.safetensors 102 GBmodel-00047-of-00048.safetensors 102 GBmodel-00017-of-00048.safetensors 7.8 GBmodel-00005-of-00048.safetensors 7.8 GBmodel-00011-of-00048.safetensors 7.8 GBmodel-00023-of-00048.safetensors 7.8 GBmodel-00027-of-00048.safetensors 7.8 GBmodel-00031-of-00048.safetensors 7.8 GBmodel-00035-of-00048.safetensors 7.8 GBmodel-00039-of-00048.safetensors 7.8 GBmodel-00013-of-00048.safetensors 7.8 GBmodel-00014-of-00048.safetensors 7.8 GBmodel-00015-of-00048.safetensors 7.8 GBmodel-00016-of-00048.safetensors 7.8 GBmodel-00018-of-00048.safetensors 7.8 GBmodel-00019-of-00048.safetensors 7.8 GBmodel-00020-of-00048.safetensors 7.8 GBmodel-00021-of-00048.safetensors 7.8 GBmodel-00022-of-00048.safetensors 7.8 GBmodel-00024-of-00048.safetensors 7.8 GBmodel-00025-of-00048.safetensors 7.8 GBmodel-00026-of-00048.safetensors 7.8 GBmodel-00028-of-00048.safetensors 7.8 GBmodel-00029-of-00048.safetensors 7.8 GBmodel-00030-of-00048.safetensors 7.8 GBmodel-00032-of-00048.safetensors 7.8 GBmodel-00033-of-00048.safetensors 7.8 GBmodel-00034-of-00048.safetensors 7.8 GBmodel-00036-of-00048.safetensors 7.8 GBmodel-00037-of-00048.safetensors 7.8 GBmodel-00038-of-00048.safetensors 7.8 GBmodel-00040-of-00048.safetensors 7.8 GBmodel-00041-of-00048.safetensors 7.8 GBmodel-00042-of-00048.safetensors 7.8 GBmodel-00003-of-00048.safetensors 7.8 GBmodel-00004-of-00048.safetensors 7.8 GBmodel-00006-of-00048.safetensors 7.8 GBmodel-00007-of-00048.safetensors 7.8 GBmodel-00008-of-00048.safetensors 7.8 GBmodel-00009-of-00048.safetensors 7.8 GBmodel-00010-of-00048.safetensors 7.8 GBmodel-00012-of-00048.safetensors 7.8 GBmodel-00046-of-00048.safetensors 2.7 GBmodel-00044-of-00048.safetensors 2.7 GBmodel-00045-of-00048.safetensors 2.6 GBmodel-00002-of-00048.safetensors 1.3 GBmodel-00043-of-00048.safetensors 1.3 GBmodel-00001-of-00048.safetensors 971 MBdealignai/DeepSeek-V4.1-Flash-UNCENSORED-FP8Have the files? Submit a torrent.
Submit a torrentHave the complete files and permission to share them? Replace the example folder path. The script checks the files, makes a torrent and submits its magnet and checksums. Keep your torrent client seeding. Get your community key from your settings, and read the script before running it.
curl -fsSL https://pirateface.co/package.sh | bash -s -- --repo MomonV/DeepSeek-V4.1-Flash-UNCENSORED-NVFP4 ./model-folder