GLM-4-32B-EleutherAI

NOTE: this is not an official THUDM model

This is a model finetuned on the #off-topic channel of the EleutherAI Discord server. It is an updated version of THUDM/Mistral-Small-24B-EleutherAI.

Base model: arcee-ai/GLM-4-32B-Base-32K

Data sourced from: https://atlas.nomic.ai/data/hivemind/last-year-of-ai-ml-discord-servers-2024-04-10

Usage

Text completions with this format:

_fleetwood: oh its free in SG

main.ai: ????
bro i paid $2k for that shit

_fleetwood: well isn't it like trivial cost compared to actual procedure

main.ai: it might be free for people who wait 6 months and get a whole bunch of fake diagnoses first

_fleetwood: he said the whole thing about it in the book

main.ai: i did not want to do that

Notes

Knowledge cutoff: 2024-10

Compared to the previous version, the extra punctuation issue should be less prevalent, however it may still occasionally appear.

Downloads last month
-
Safetensors
Model size
1.51M params
Tensor type
BF16
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for not-THUDM/GLM-4-32B-EleutherAI

Quantizations
2 models