embedding results lost

[honahlee]

Bug Report

dell rtx5090 24G.
2T harddisk
32G memory

I downloaded the latest GPT4All, installed some local offline large model(gguf llm), and started analyzing, embedding the slices of the local documents (500M), which took me 12 hours to complete. and chatting with ai is doing well.

However, after exiting the program and entering again, the completed embedding were gone! They are lost!!!

I tried several times; sometimes when I re-enter the program, the results are still there; other times they are gone.

Steps to Reproduce

how to find the results of the embedding work back?

Expected Behavior

Your Environment

• GPT4All version: new
• Operating System: win11 pro
• Chat model used (if applicable): qwen

[honahlee]

net error, pls del other two same issues...

[honahlee]

anyone could help me solve this problem? or I'll never dare to run the program again! terrible!

[honahlee]

I guess it's because running 2 or more embedding tasks at the same time?

If so, it's should be fixed in order to ensure users' data security!

[djeada]

Could you describe in detail the steps you initially took when it worked? How did you embed your document the first time?

[honahlee]

Hi, I'm in Chinese computer, inputting the files for embedding. Opening some more embedding tasks at the same time. After exiting and come back, all embedded files are lost, dispeared!

…

------------------ 原始邮件 ------------------ 发件人: "Adam ***@***.***>; 发送时间: 2025年8月29日(星期五) 凌晨3:13 收件人: ***@***.***>; 抄送: " ***@***.***>; ***@***.***>; 主题: Re: [nomic-ai/gpt4all] embedding results lost (Issue #3616) djeada left a comment (nomic-ai/gpt4all#3616) Could you describe in detail the steps you initially took when it worked? How did you embed your document the first time? — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: ***@***.***>

[honahlee]

I have a 24G memory graphics card, Win-11 system, and a local offline AI notebook, without internet.  I have downloaded many GGUF models. Could you please tell me how to set the parameters for GPT4all to achieve AI functionalities?  I've mainly three needs: 1, lengthy local documents analysis; 2, writing new document; 3, doc  translation. The models I have downloaded are as follows, could please let me know how to configure them in GPT4all？ Qwen3-1.7B-MT-en2zh-Gungeon.GGUF Madlad400-3B-MT.gguf(q2k,q3k,q4k) codeqwen-1_5-7b-chat-q4_k_m.gguf DeepSeek-Coder-V2-Lite-Instruct.Q4_K_M.gguf DeepSeek-R1-Distill-Llama-8B-Q5_K_M.gguf gte-qwen1.5-7b-instruct.Q4_K_M.gguf Meta-Llama-3-8B-Instruct-Q5_K_M.gguf Ministral-8B-Instruct-2410-Q8_0.gguf

…

------------------ 原始邮件 ------------------ 发件人: "nomic-ai/gpt4all" ***@***.***>; 发送时间: 2025年8月29日(星期五) 凌晨3:13 ***@***.***>; 抄送: " ***@***.******@***.***>; 主题: Re: [nomic-ai/gpt4all] embedding results lost (Issue #3616) djeada left a comment (nomic-ai/gpt4all#3616) Could you describe in detail the steps you initially took when it worked? How did you embed your document the first time? — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: ***@***.***>

[honahlee]

Sorry, it was my mistake. Because my local document has several hundred gigabytes, it opens very slowly. Nothing is lost, but it takes a long time to display.

…

------------------ 原始邮件 ------------------ 发件人: "nomic-ai/gpt4all" ***@***.***>; 发送时间: 2025年8月29日(星期五) 凌晨3:13 ***@***.***>; 抄送: " ***@***.******@***.***>; 主题: Re: [nomic-ai/gpt4all] embedding results lost (Issue #3616) djeada left a comment (nomic-ai/gpt4all#3616) Could you describe in detail the steps you initially took when it worked? How did you embed your document the first time? — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: ***@***.***>