穏やかな一年になりますように
いつもブログを読んでくださり、ありがとうございます。 私事で恐縮ですが、喪中のため新年のご挨拶を失礼させていただきます。
毎年この時期は気合を入れた記事を書いてまいりましたが、今年は少し心を休める時間に充てたいと考えています。 皆様にとって、2026年が健康で、穏やかな光に満ちた一年となりますようお祈り申し上げます。
Wishing you a peaceful year ahead
Thank you so much for always reading my blog and for your continued support.
I would like to share a personal update. Having lost my mother last year, I am currently in a period of mourning. Therefore, I will be refraining from the usual celebratory New Year’s greetings this year.
In previous years, I always put a lot of energy into my special New Year’s posts. However, for this year, I have decided to take some time to rest my heart and quietly honor my mother’s memory.
I sincerely hope that 2026 will be a healthy and peaceful year for all of you, filled with gentle light.
今年の抱負
1.過集中して主たる業務に影響が出ないよう、今後もこのスタンスを維持できるよう努力する
2.約1530ある記事を全て見直し、ラベルの付け直しを行なう
3.過去記事で内容に不整合がある場合 修正を行なう
4.リンク切れ箇所を修正 もしくはリンク切れを明示する
5.陳腐化した過去投稿記事の更新（再訪を含む）をする
6.ローカルLLMを構築して約1530ある記事を読み込ませて、今後のブログ記事作 成に生かす。
機材
ML350 Gen9
Memory:256GB OS：M.2-2280 NVMe 1TB Data:M.2-2280 NVMe 2TB HDD6TB(RAID1)
GPU: RTX4000Ada 20GB, Quadro RTX4000 8GB, QuadroK620 2G(画像出力）の3枚差し
CPU Intel(R) Xeon(R) CPU E5-2698 v4 @ 2.20GHz (2 プロセッサで20Core,40Thread）
Windows11 Pro (refusで導入25H2）上で環境を構築
LLM「AnythingLLM」と「LM Studio」を連携 Windows上にPython導入
次の段階OllamaとOpen WebUI
早速動かしてみた。
1536. スタンドアローンLLMを構築して約1530ある記事を読み込ませて、今後のブログ記事作 成に生かす。とりあえずLM Studioが動いた
My New Year’s Resolutions for 2026
-
Maintain a Healthy Balance: I will strive to manage my tendency to hyperfocus, ensuring that my deep-dive projects do not interfere with my primary responsibilities.
-
Blog Archive Audit: I plan to review all 1,530+ posts and systematically update/re-organize the labels.
-
Content Consistency: I will identify and correct any inconsistencies in the content of past articles.
-
Link Maintenance: I will fix broken links or clearly mark them as "dead links" where necessary.
-
Post Refresh: I will update outdated content (including "revisiting" old topics) to ensure the information remains relevant.
Leveraging Local LLMs I have built a Local LLM environment to ingest and analyze my 1,530+ blog posts. My goal is to utilize this "personal knowledge base" to enhance my future writing process.
Hardware Specifications
-
Model: HP ProLiant ML350 Gen9
-
CPU: Dual Intel Xeon E5-2698 v4 (Total: 40 Cores / 80 Threads)
-
Memory: 256GB
-
Storage: * OS: 1TB M.2 NVMe SSD
Data: 2TB M.2 NVMe SSD
Archive: 6TB HDD (RAID 1)
-
GPU Configuration (Triple GPU Setup):
-
NVIDIA RTX 4000 Ada (20GB) — Primary Inference
-
NVIDIA Quadro RTX 4000 (8GB)
-
NVIDIA Quadro K620 (2GB) — Display Output
-
-
OS: Windows 11 Pro (Version 25H2, installed via Rufus)
Software & AI Stack I have built the environment on Windows with Python installed. Currently, I am integrating AnythingLLM with LM Studio. My next step is to transition to a more robust setup using Ollama and Open WebUI.
