AuteDev (@autedev_com) header
AuteDev (@autedev_com) profile picture

2026年8月ごろからローカルLLMを触り始めた素人です。ローカルLLMだけでアプリを自動開発させる実験をしています。

Joined August 2026
Photo by AuteDev (@autedev_com)Photo by AuteDev (@autedev_com)Photo by AuteDev (@autedev_com)Photo by AuteDev (@autedev_com)Photo by AuteDev (@autedev_com)Photo by AuteDev (@autedev_com)Photo by AuteDev (@autedev_com)Photo by AuteDev (@autedev_com)Photo by AuteDev (@autedev_com)Photo by AuteDev (@autedev_com)
The top 5 most downloaded uncensored models on Hugging Face right now: - DeepSeek V4.1 Flash 1017 Uncensored - Qwen 3.8 27B Abliterated - GLM 5.3 Flash Abliterated - Huihui Qwen 3.8 27B Abliterated - GLM 5.3 Uncensored - Heretic Qwen3.8-27B the aggressive Download them now, just in case.
24
97
7
1,135
55,676
Strata環境を整えるために、ようやく P100を触ったので計測結果と現行マシンとの比較レポ 3行でサマると - 生成は思ったより速いかも? - 長文のinputは要注意 - 熱管理だいじ 画像が20枚も貼れるようになったらしいので、Claude Codeが生成したスライドをそのまま貼っておきますね #地元のLLM
Strata環境を整理中 IQ3_S 262K想定 既存では RTX 3080 20GB * 2 で動かしてたけど、情報集めてみたら RTX 4090 より快適なことに今更気づいた 引越しを検討してたけどメイン維持で、サブにP100 16GB * 2 を始動させて、 RTX 5060 Ti 16GBはLLM以外のメディア生成で活発化させよかな #地元のLLM
2
1
10
515
Maya v1.1.58 is out 🚀 Full notes 👇 github.com/mw00/project-maya…
12
25
2
155
9,779
Which model fits your PC? Add your GPU memory and your system RAM. That total is what counts. The Recommended column is the fast setup: the whole model sitting in VRAM + RAM. You don't need that to start. Maya keeps the most-used parts on the GPU, the next in RAM, and reads the rest from your NVMe. A supported GPU, about 32GB of RAM, and a fast SSD is the floor. More memory is faster. Save this. github.com/mw00/project-maya
9
15
1
115
9,263
Project Maya v1.1.29 is here, and it's a big one: MiniMax-M3.1-Flash now runs on Maya. That makes three frontier open models you can run at home on your own GPUs: GLM-5.3-Flash, DeepSeek-V4.1-Flash and now MiniMax-M3.1-Flash, a 428B model with 23B active per token and up to a 1M context. MiniMax gets its own engine, built by @needmorevram. We brought it up to the same standard as GLM and DeepSeek before shipping it: - Runs on NVIDIA from V100 up, on Linux and Windows - Fits itself to your PC: the setup picks the context your memory can hold, and the engine plans VRAM and RAM instead of failing - A compact cache by default, so long contexts take half the memory with no loss in quality - Pictures, tool calls and every one of MiniMax's thinking levels, with its own recommended settings out of the box - The vision tower stays off your GPU until you send a picture, so the VRAM goes to the model Huge thanks to @needmorevram, and to everyone who tests, reports and contributes to Maya. This is what community open source looks like. Next up (WiP): our own Maya GGUF quants of MiniMax-M3.1-Flash, small enough for a 128 GB PC. Link in the reply 👇
18
34
6
176
11,205
DeepSeek v4.1 Flashが持て囃されてる理由が全然わからない 安いので単純作業代行という意味では超優秀だけど、コーディングさせるとQwen3.8Flashとさほど変わらん MiniMax 3.1 Flashがとにかく優秀
32
13
307
66,729
Maya models have surpassed 410K+ downloads on @HuggingFace 🚀🔥 And that’s without even counting the EXL3 models! 👀 Thank you to everyone who's been trying Maya, downloading the models and supporting the project. This wouldn't happen without you ❤️
7
8
1
72
3,707
AuteDev retweeted
How much reposts and tags for @Alibaba_Qwen @QwenDevs is needed to get me in? 👀
Qwen employees are teasing some big reveals and product announcements at the upcoming Apsara Conference in Amsterdam on October 29 The Qwen 4 models have been rumored to launch later this month
11
22
4
242
14,674
今日分かったことがあります。 リポストって本当に素晴らしい。 自分だけじゃなく、周りの人にも良い影響を与えられるってすごくないですか?😳 いつも沢山リポストしていただいているので、僕も積極的にお返ししていきます🙇 皆さん、いつも本当にありがとう!!🫶
15
26
59
1,067
皆様。吉報中の吉報です Strata開発者の、NIKOさんがっ。 「Macでも早くなるようにトライしてやんぜ」って言ってくれました! 僕の方でもStudioを払い出せるようにNW周り調整中ですが、手持ちの環境でまずは実施してくれるとのこと。 勝ちで、ローカルLLMの最適化、何かが大きく変わってからじゃ遅いから自分たちの手段のためにも何ができるのかトライしようぜ Thank you so much, Niko-san!
Replying to @hokazuya
I already got it Hodachi! So no more worries
2
8
2
87
7,886