Open-Assistant
Open-Assistant copied to clipboard
OpenAssistant is a chat-based assistant that understands tasks, can interact with third-party systems, and retrieve information dynamically to do so.
How much VRAM do the models approx. consume? OA_SFT_Llama_30B_6 = my estimate about 68GB vram 12B model = 40GB vram ... Are those assumptions right? Should i be able to...
For the Issue #2827. I have made changes to JokeExplaniation Class. This PR implements the DatasetEntry class in the JokeExplaination class to generalize the data. The DatasetEntry class provides a...
I would like to ask if it would be somehow possible AND feasible to implement this paper into openassistant: https://arxiv.org/pdf/2304.11062.pdf Even it shows the examples for a BERT like model,...
Update instruction datasets to return the `DatasetEntry` class. Also checked with `python check_dataset_appearances.py -d humaneval_mbpp_codegen_qa --cache_dir .cache --mode sft` but did not find anything.
The OA servers are sponsored by Evobits https://evobitsit.com/ .. they need to be added to our list of sponsors. Icon: https://evobitsit.com/img/favicon/favicon.ico
I apologize in advance but I just cannot seem to find a definitive answer to this, even though I'm sure it must be somewhere. How do we effectively prompt the...
Bumps [msw](https://github.com/mswjs/msw) from 0.49.3 to 1.2.1. Release notes Sourced from msw's releases. v1.2.1 (2023-03-24) Bug Fixes update dependencies to support ESM (#1575) (e7a60081526bd6f00ac17a62c424abf59976555b) @kettanaito update "headers-polyfill" to 3.1.2 to fix...
I feel that the current prompt/answer rating system can be too subjective and unclear at times, which may be affecting RLHF quality. So, from current experiences, here's my proposal on...
The monospace font used by the web client cannot render Morse code properly. The dots are sometimes rendered level with the dashes (in the middle, like a multiply operator), and...
This **PR** is **_Serbian language_** translate for **OA** .