---
license: mit
datasets:
- ZeroAgency/ru-big-russian-dataset
language:
- ru
- en
tags:
- mistral
- chat
- conversational
- transformers
inference:
  parameters:
    temperature: 0
pipeline_tag: text-generation
base_model:
- ZeroAgency/Zero-Mistral-24B
library_name: llama.cpp
base_model_relation: quantized
quantized_by: bethrezen
---

# Model Card for Zero-Mistral

This is a GGUF version of [ZeroAgency/Zero-Mistral-24B](https://huggingface.co/ZeroAgency/Zero-Mistral-24B).

All quants made with [llama.cpp](https://github.com/ggerganov/llama.cpp) version b5083.

Quants available:
- BF16
- F16
- IQ4_NL
- IQ4_NL_L - same as above but with `--leave-output-tensors`
- IQ4_XS
- IQ4_XS_L - same as above but with `--leave-output-tensors`
- Q4_K_M
- Q4_K_M_L - same as above but with `--leave-output-tensors`
- Q6_K
- Q6_K_L - same as above but with `--leave-output-tensors`
- Q8_0 - quantized from bf16 gguf
- Q8_0-direct - direct convertation from hf
- Q8_0_L - quantized from bf16 but with `--leave-output-tensors`

![image/png](https://huggingface.co/ZeroAgency/Zero-Mistral-24B/resolve/main/zero-mistral-500.png)