I'm using Webui 0.4.8 with the latest Ollama3.3 installed with Docker and currently running it in on an AWS EC2 instance.
I attach to the chat 4 files (not via the Knowledge functionality), let's call them FILE_1.md, FILE_2.md, FILE_3.md, FILE_4.md. I have setup my prompt as follows:
### Variables
**Filenames:** `FILES = ["FILE_1.md", "FILE_2.md", "FILE_3.md", "FILE_4.md"]`
### Task:
Evaluate each file with a filename listed in the list `FILES` against criterion X.
### Output Format for Batch Evaluations:
For all files with filenames in the list `PROPOSALS` do X & Y
Prompts like this always fail and the system complains that there are no files attached to work with, so it improvises. When I modify the prompt to work on a single file it works as expected. Is there any suggested format when writing prompts to declare lists with filenames and then instruct the model to do certain actions on each file recursively?
Originally created by @dimk1 on GitHub (Dec 11, 2024).
Original GitHub issue: https://github.com/open-webui/open-webui/issues/7784
Hello,
I'm using Webui 0.4.8 with the latest Ollama3.3 installed with Docker and currently running it in on an AWS EC2 instance.
I attach to the chat 4 files (not via the Knowledge functionality), let's call them `FILE_1.md, FILE_2.md, FILE_3.md, FILE_4.md`. I have setup my prompt as follows:
```
### Variables
**Filenames:** `FILES = ["FILE_1.md", "FILE_2.md", "FILE_3.md", "FILE_4.md"]`
### Task:
Evaluate each file with a filename listed in the list `FILES` against criterion X.
### Output Format for Batch Evaluations:
For all files with filenames in the list `PROPOSALS` do X & Y
```
Prompts like this always fail and the system complains that there are no files attached to work with, so it improvises. When I modify the prompt to work on a single file it works as expected. Is there any suggested format when writing prompts to declare lists with filenames and then instruct the model to do certain actions on each file recursively?
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Originally created by @dimk1 on GitHub (Dec 11, 2024).
Original GitHub issue: https://github.com/open-webui/open-webui/issues/7784
Hello,
I'm using Webui 0.4.8 with the latest Ollama3.3 installed with Docker and currently running it in on an AWS EC2 instance.
I attach to the chat 4 files (not via the Knowledge functionality), let's call them
FILE_1.md, FILE_2.md, FILE_3.md, FILE_4.md. I have setup my prompt as follows:Prompts like this always fail and the system complains that there are no files attached to work with, so it improvises. When I modify the prompt to work on a single file it works as expected. Is there any suggested format when writing prompts to declare lists with filenames and then instruct the model to do certain actions on each file recursively?