ukisai/Swift-1.5-Qwen3.8-27b
Image-Text-to-Text β’ 28B β’ Updated β’ 6.92k β’ 180
Voronov Igor
VoronBrothers
AI & ML interests
None yet
Recent Activity
liked a model 2 days ago
ukisai/Swift-1.5-Qwen3.8-27b liked a model 2 days ago
unsloth/Qwen3.8-27B-GGUF liked a model 2 days ago
orcarouter/Mistral-Large-4-1T-A52B-GGUFOrganizations
None yet
liked 3
models 2 days ago
liked a
model 3 days ago
reacted to Parveshiiii's post with π 5 days ago
Post
2,766
Most AI image detectors work fine in the lab and fail on social media.
The reason is simple. Real photos get compressed, resized and re-uploaded again and again, and detectors trained on clean images read that damage as a sign of fakery. A real photo that has passed through a few platforms often gets flagged as AI-generated.
AIRealNet-2 is built to fix that. It is the upgrade to our original Modotte/AIRealNet, trained on images collected from platforms like Instagram and X, so it has seen the kinds of quality loss and compression real images go through online.
The training set has 98k images, split 50/50 between AI-generated and real, so the model has no class bias to lean on. Personal photos were excluded by design, including AI-edited ones.
The architecture is unchanged: a SwinV2 Tiny backbone with a binary classification head, about 0.2B parameters, working at 256x256 resolution. It stays fast and easy to deploy for research or production.
Model: Modotte/AIRealNet-2
The reason is simple. Real photos get compressed, resized and re-uploaded again and again, and detectors trained on clean images read that damage as a sign of fakery. A real photo that has passed through a few platforms often gets flagged as AI-generated.
AIRealNet-2 is built to fix that. It is the upgrade to our original Modotte/AIRealNet, trained on images collected from platforms like Instagram and X, so it has seen the kinds of quality loss and compression real images go through online.
The training set has 98k images, split 50/50 between AI-generated and real, so the model has no class bias to lean on. Personal photos were excluded by design, including AI-edited ones.
The architecture is unchanged: a SwinV2 Tiny backbone with a binary classification head, about 0.2B parameters, working at 256x256 resolution. It stays fast and easy to deploy for research or production.
Model: Modotte/AIRealNet-2
- 4 replies
liked 2
models 5 days ago
liked a
model 6 days ago
liked a
model 12 days ago
liked a
model 14 days ago
liked a
model 15 days ago
liked 2
models 16 days ago
reacted to AtAndDev's post with ππ 23 days ago
Post
5,295
@Banaxi-Tech stop hiding my comments. AND STOP STEALING PAPERS AND SPREADING MISINFORMATION.
your BGA blog is a copy of NSA (deepseek, 2025) branded under your name. literally the same top16 selected blocks, 512 local window, router over block summaries, all you did was change block size from 64 to 128.
you didnt cite NSA once but you put a βplease cite BGAβ bibtex at the bottom.
i commented under your post and said that there is no way that you can support claims like: βThe Accuracy Should BE WAy better than DSA but untested yet.β you didnt run a single experiment. and the 256x isnt from BGA, its just n/2k with k=2048 so the exact same k DSA uses. if opus wrote this for you, at least read it before posting.
i commented again after you hid my comment despite it having constructive and correct feedback and you hid that too. and again.
you can hide the truth and just try to get hf post likes..... but is it really the thing that needs to be done? do you really want to take papers and make them yours while barely even changing the params?
admitting your mistakes and doing something about them needs humbleness, intelligence, humanness.
i encourage you to admit your mistakes and try to do better next time (at least read what blog your ai wrote or do proper experiments to back your stuff up).
your BGA blog is a copy of NSA (deepseek, 2025) branded under your name. literally the same top16 selected blocks, 512 local window, router over block summaries, all you did was change block size from 64 to 128.
you didnt cite NSA once but you put a βplease cite BGAβ bibtex at the bottom.
i commented under your post and said that there is no way that you can support claims like: βThe Accuracy Should BE WAy better than DSA but untested yet.β you didnt run a single experiment. and the 256x isnt from BGA, its just n/2k with k=2048 so the exact same k DSA uses. if opus wrote this for you, at least read it before posting.
i commented again after you hid my comment despite it having constructive and correct feedback and you hid that too. and again.
you can hide the truth and just try to get hf post likes..... but is it really the thing that needs to be done? do you really want to take papers and make them yours while barely even changing the params?
admitting your mistakes and doing something about them needs humbleness, intelligence, humanness.
i encourage you to admit your mistakes and try to do better next time (at least read what blog your ai wrote or do proper experiments to back your stuff up).
- 49 replies
liked a
model 27 days ago
liked a
model 29 days ago
reacted to SeaWolf-AI's post with π about 1 month ago
Post
6,287
π» Data-center AI, now on a laptop: POCKET-Darwin-180B
We're releasing a 4-bit GGUF build of Darwin-180B-RSI, #1 on seven official Hugging Face leaderboards (self-reported), that runs without a GPU.
π¦ 360 GB β 111 GB (4-bit GGUF, 4 files)
π₯οΈ No GPU: one server CPU (16 threads) at 18.4β21.0 tokens/s
π» RTX 5060 laptop (8 GB VRAM) + 32 GB RAM: 4.17 tokens/s
π§ 128 GB mini PC: whole model in memory, no GPU needed
π― MMLU-Pro, 2,000 questions, paired: original 87.65% = 4-bit 87.65%
How?
Β· Only ~3B of 180B parameters are active per token (10 of 512 experts)
Β· llama.cpp streams just the needed experts from SSD, so 32 GB RAM is enough
Β· Graft quantization: we took the proven Unsloth UD-Q4_K_XL base build and swapped in only the 300 tensors our RSI training changed (300/300 verified)
Under the hood is Model-level Recursive Self-Improvement. The model solves verifiable problems, keeps only its own solutions that check out as correct, and trains on them. No human-written solutions or reasoning traces.
Built for teams that can't send data to an external cloud (defense, finance, public sector) to run a top-tier model fully offline.
π Article: https://huggingface.co/blog/FINAL-Bench/data-center-ai-now-on-a-laptop-pocket-darwin-180b
π€ Model: FINAL-Bench/POCKET-Darwin-180B-GGUF
𧬠Original: FINAL-Bench/Darwin-180B-RSI
#Darwin #RSI #GGUF #llamacpp #OnDevice #MoE
We're releasing a 4-bit GGUF build of Darwin-180B-RSI, #1 on seven official Hugging Face leaderboards (self-reported), that runs without a GPU.
π¦ 360 GB β 111 GB (4-bit GGUF, 4 files)
π₯οΈ No GPU: one server CPU (16 threads) at 18.4β21.0 tokens/s
π» RTX 5060 laptop (8 GB VRAM) + 32 GB RAM: 4.17 tokens/s
π§ 128 GB mini PC: whole model in memory, no GPU needed
π― MMLU-Pro, 2,000 questions, paired: original 87.65% = 4-bit 87.65%
How?
Β· Only ~3B of 180B parameters are active per token (10 of 512 experts)
Β· llama.cpp streams just the needed experts from SSD, so 32 GB RAM is enough
Β· Graft quantization: we took the proven Unsloth UD-Q4_K_XL base build and swapped in only the 300 tensors our RSI training changed (300/300 verified)
Under the hood is Model-level Recursive Self-Improvement. The model solves verifiable problems, keeps only its own solutions that check out as correct, and trains on them. No human-written solutions or reasoning traces.
Built for teams that can't send data to an external cloud (defense, finance, public sector) to run a top-tier model fully offline.
π Article: https://huggingface.co/blog/FINAL-Bench/data-center-ai-now-on-a-laptop-pocket-darwin-180b
π€ Model: FINAL-Bench/POCKET-Darwin-180B-GGUF
𧬠Original: FINAL-Bench/Darwin-180B-RSI
#Darwin #RSI #GGUF #llamacpp #OnDevice #MoE
- 30 replies
liked 2
models about 1 month ago
