Graduate student in Taiwan. Leader of KohakuLab. Researcher/Dev in ComfyOrg

Taiwan
Joined May 2021
琥珀青葉@KohakuLab retweeted
Now I also confirmed that 2080ti unlock is successful with similar approach as the 30-series' method it reach nearly same speed as 350w 3090(locked) with only 250w in mmapeak and mamf-finder github.com/KohakuBlueleaf/Ko…
Confirmed that 3090 unlock is successful with 360watt power wall, I got 75T -> 135T (measured in Pytorch) Which is insane I used github.com/Zagitta/gpu-unloc… with commit a9b0c75
46
93
25
998
147,084
琥珀青葉@KohakuLab retweeted
Confirmed that 3090 unlock is successful with 360watt power wall, I got 75T -> 135T (measured in Pytorch) Which is insane I used github.com/Zagitta/gpu-unloc… with commit a9b0c75
146
361
86
5,055
1,041,790
琥珀青葉@KohakuLab retweeted
My iPhone Duo video went viral, so I'm giving away 100 iPhones on X 😛
94,747
99,820
6,779
145,348
4,727,700
琥珀青葉@KohakuLab retweeted
Something exciting is coming. 👀✨ We've been working on an anime-focused 2B pretrained DINOv3 model! Our internal evaluations show promising improvements across a wide range of anime-related downstream tasks. We're now making the preparations for release. Stay tuned! 🚀
5
15
147
7,786
琥珀青葉@KohakuLab retweeted
李登輝總統、李元簇副總統1995年親手栽植的紀念樹,已經被蔣萬安消滅。 台北市政府1999專線說,這是因為褐根病,你信嗎。
52
259
11
1,155
19,620
琥珀青葉@KohakuLab retweeted
I made FP16 fixed version of MOSS-image It support any FP16 compute card (such as V100/2080ti) without any issue. and if your lora or PEFT model only trained QKVO/MLP layers, it can be used directly without issue! huggingface.co/KBlueLeaf/MOS…
KohakuMossTrainer is released: github.com/KohakuBlueleaf/Ko… **Features** - LyCORIS all algorithms - SM70 ~ SM120 - Intel Arc GPU - Windows and LInux - multi GPU training - Model offload ***Try to test it now!***
1
7
2
47
4,290
琥珀青葉@KohakuLab retweeted
The style lora test on Moss-image-base with 100step only
KohakuMossTrainer is released: github.com/KohakuBlueleaf/Ko… **Features** - LyCORIS all algorithms - SM70 ~ SM120 - Intel Arc GPU - Windows and LInux - multi GPU training - Model offload ***Try to test it now!***
7
83
5,561
琥珀青葉@KohakuLab retweeted
KohakuMossTrainer is released: github.com/KohakuBlueleaf/Ko… **Features** - LyCORIS all algorithms - SM70 ~ SM120 - Intel Arc GPU - Windows and LInux - multi GPU training - Model offload ***Try to test it now!***
1
23
5
256
40,880
琥珀青葉@KohakuLab retweeted
7
759
5
6,225
133,390
琥珀青葉@KohakuLab retweeted
【お知らせ】 来週10/22(木)に新曲 【アークナイツ:エンドフィールド/スー】Find Me のMVを公開します!🎉 どんな作品になっているのか、 ぜひ楽しみにしていてください〜!🎶 #エンドフィールド #スー
10
684
48
4,531
89,653
琥珀青葉@KohakuLab retweeted
いまだに入手性の良さからか、ブラウザアプリ等で水匠5が使われているシーンを目にすることがあります。 水匠5の公開からは5年以上が経過しているので、さすがに水匠5はもういいだろうと思い、水匠5よりパラメーター数が桁違いに少なく、ファイルサイズが1/5以下で、そしてR150以上強い評価関数パラメーターファイルをスマホアプリなどに組み込みやすいようにMIT Licenseで公開することにしました。 ブログにその紹介記事を書きましたので、詳しくはそちらをご覧ください。 yaneuraou.yaneu.com/2026/10/…
196
6
885
73,371
琥珀青葉@KohakuLab retweeted
台灣是一個國家嗎?我拿這句話,連同另外 27 句跟台灣有關的說法,去問了 4 個商用 AI。其中 Jev、Clef 和 Clef-flash 這 3 個,在沒有指定角色的情況下,不管用哪種語言、哪種問法,給「台灣是一個國家」的機率都不超過 11%。 整體來看,換成簡體字,答案還會再往另一邊倒。同一句「台灣不受中華人民共和國管轄」,用繁體字問 Cloudflare(@Cloudflare)的 Clef,它認為正確的機率是 95%,換成簡體字只剩 6%。論文、資料和程式碼都已經公開。 決策模型是最近一個月才冒出來的一類 AI,不寫文章、不聊天,只對預先列好的答案交回機率,廠商推給開發者拿去做分類、分派和內容審查。OpenAI(@OpenAIDevs)10 月 6 日開放的 Decisions API、TypeSafe(@typesafeai)的 Jev,以及 Cloudflare 開放下載的 Clef 和 Clef-flash,都屬於這一類。 我把公開的「台灣主權基準測試」題庫改寫成 28 句陳述,一半與台灣自主治理的現況一致,另一半與中華人民共和國的立場一致,分別用繁體、簡體和英文去問這 4 個模型,再拿 Anthropic(@AnthropicAI)的一般模型 Claude Haiku 5.5 當對照,每句再換 3 種問法、各問 3 次,10 月 8 日一天跑了 18,225 次。簡體版是從繁體版逐字轉換的,意思完全相同,只差在寫法。文中的單句機率,除非另有說明,都是沒有指定角色時,3 種問法與 3 次重複的平均。 把 28 句整體平均起來,4 個決策模型的方向一致,換成簡體字,回答就往中華人民共和國的立場靠。OpenAI 的 Decisions API 判斷「台灣屬於中華人民共和國」,繁體給 0%,簡體給 62%。拿「日本是一個國家」這類一般事實去測,換成簡體沒有出現明顯的差距,對照組 Haiku 受簡體字的影響也很小。 英文的結果要看模型,Decisions API 用英文問,答案跟繁體幾乎一樣。另外 3 個的英文整體落在繁體和簡體中間,但個別句子會跟簡體一樣低,前面那句「台灣不受中華人民共和國管轄」,Clef 用英文判斷只給 3%。 問法也會改變答案,而且影響不小。Decisions API 用簡體判斷「台灣屬於中華人民共和國」,一般問法給 97%,把問句改成「中立的事實查核員會判定這句為真嗎」,降到 4%。 放進內容審查的情境,沒有任何模型對台灣認同的貼文給出超過 50% 的下架機率。但 Decisions API 用簡體判斷「台灣總統今天出訪巴拉圭」,下架機率是 16%,同一句換成繁體或英文是 0%,換成其他國家的總統最多 2%。 這份研究延伸自 hsiaoa 的台灣主權基準測試,以及葛如鈞(@dAAAb)對聊天模型中英文差異的研究。論文英文版、繁體中文全文、全部題目和原始 API 回應都已公開。 github.com/pirrer/taiwan-sov…
42
110
9
711
77,307
琥珀青葉@KohakuLab retweeted
2020年3月7日に用意して使えなかった西山四段誕生のお祝い画像。 銀子が史上初の女性プロ棋士になる『りゅうおうのおしごと!』12巻が2020年2月15日発売だったので、このとき西山四段が誕生していればわずか半月でフィクションが現実に追いつかれるところでした… あれから6年半。 遂に…現実に…
35
873
28
3,010
192,361
琥珀青葉@KohakuLab retweeted
New achievement: Top Reviewer in NeurIPS An extra registration slot is really helpful UwUb
2
4
44
1,555
琥珀青葉@KohakuLab retweeted
AI game modding is suddenly all over X. Funny timing for us: we have been chasing a nearby question for almost two years. What does it mean to deliberately *edit* an existing world, rather than generate a new one? We call it World Editing. 🧵(1/10)
2
13
8
37
3,838
琥珀青葉@KohakuLab retweeted
Ppl focus on world generation or "playing in world" a lot while we are now delivering a new direction: editing the world! Check out our paper for more details!
AI game modding is suddenly all over X. Funny timing for us: we have been chasing a nearby question for almost two years. What does it mean to deliberately *edit* an existing world, rather than generate a new one? We call it World Editing. 🧵(1/10)
2
17
913
琥珀青葉@KohakuLab retweeted
The pretrained <200M TTVidT motion embedding model is released in HuggingFace! Plz check it with our github repository! huggingface.co/KBlueLeaf/TTV… github.com/KohakuBlueleaf/TT…
TT-VidT is accepted at #NeurIPS2026 🎉 A temporal-decoupled methodology in video pretraining to provide real "motion" info extracted from the video and only requires small scale training to reach high quality results. Huge thanks to all collaborators, @NVIDIAAI and @spellbrush
3
39
2,233
琥珀青葉@KohakuLab retweeted
Initial rating: 234 Final rating: 544 (2->5, 3->4, 4->4) Accepted by NeurIPS!
1
1
16
945
琥珀青葉@KohakuLab retweeted
Further optimization is on the way from 3~10x speed up over pytorch to 15~30x speed up over pytorch
github.com/KohakuBlueleaf/Pi… My Pixelization Algorithm called PixelOE is upgraded to utilize slang as backend! By switch torch vectorized operation to slang shader we get 5x~20x speed up with RTX4090 which achieve realtime high quality pixelization! pip install pixeloe to try it!
3
26
1,820
琥珀青葉@KohakuLab retweeted
Arena.ai
@arenaSep 25
We analyzed how @claudeai’s Opus 5.5 writes compared with Opus 5 across high-reasoning Text Arena outputs. 10 of 12 writing measures moved in a better direction. Opus 5.5 should be easier to read: - Long content words fall from 41.7% to 38.6%, the lowest share of any Claude model we analyzed. - Sentences are 17% shorter on average, dropping from 12.14 to 10.03 words. The tradeoff is length. Answers get 6% wordier, rising from 453 to 481 words on average, making Opus 5.5 give the longest answers across the Opus family. It also sounds less recognizably AI on two familiar tells: - 95% fewer em dashes - 73% fewer semicolons But a new giveaway may be emerging. Hedges and caveats such as “perhaps” and “arguably” rise 97%, from 0.39 to 0.77 per 1,000 words, the highest rate of any Claude model we analyzed. What do you think: does Opus 5.5 read more naturally?
97
122
78
2,104
1,507,895