license: apache-2.0
tags:
- mistral
- conversational
- text-generation-inference
base_model: mistralai/Mistral-Nemo-Instruct-2407
library_name: transformers
Sampling:
Mistral-Nemo-12B-Instruct-2407 is very sensitive to the temperature sampler, try values near 0.3 or else you will get some weird results. This is mentioned by MistralAI at their Transformers section
Changes:
This model is the original Mistral-Nemo-Instruct-2407 converted to GGUF and quantized using llama.cpp.
How to Use:
As of July 19, 2024, llama.cpp does not support Mistral-Nemo-Instruct-2407. However, you can use it by building from source using iamlemec's branch mistral-nemo at llama.cpp GitHub repository.
git clone -b mistral-nemo https://github.com/iamlemec/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build --config Release
Recommended to use cmake -B build -DGGML_CUDA=ON if you're using a CUDA compatible GPU.
If the build takes too long use cmake -B build --config Release -j 4, which uses 4 threads to build. Adjust the number to the amount of physical cores on your CPU.
Use:
llama-server.exe -m .\models\Mistral-Nemo-12B-Instruct-2407-Q8_0.gguf -b 512 -ub 512 -c 4096 -ngl 100
Set -b to batch size
Set -ub to physical batch size
Set -c to context size
Set -ngl to amount of layers to load onto GPU
Change the path to where the model is actually stored.
If you need more clarification on parameters check out the llama.cpp Server Docs
License:
Apache 2.0
Original Model:
Mistral-Nemo-Instruct-2407
Note: MistralAI does not have any affiliation with starble-dev.
Quants
| Name | Quant Type | Size |
|---|---|---|
| Mistral-Nemo-12B-Instruct-2407-Q2_K.gguf | Q2_K | 4.79 GB |
| Mistral-Nemo-12B-Instruct-2407-Q3_K.gguf | Q3_K | 6.08 GB |
| Mistral-Nemo-12B-Instruct-2407-Q4_K_S.gguf | Q4_K_S | 7.12 GB |
| Mistral-Nemo-12B-Instruct-2407-Q4_K_M.gguf | Q4_K_M | 7.48 GB |
| Mistral-Nemo-12B-Instruct-2407-Q5_K_M.gguf | Q5_K_M | 8.73 GB |
| Mistral-Nemo-12B-Instruct-2407-Q6_K.gguf | Q6_K | 10.1 GB |
| Mistral-Nemo-12B-Instruct-2407-Q8_0.gguf | Q8_0 | 13 GB |
| Mistral-Nemo-12B-Instruct-2407-Q8_0_L.gguf | Q8_0 | 13.7 GB |
Note: Q8_0_L
Leaves output.weight unquantized. Increases model size but may also increase quality.
The MistralAI Team:
Albert Jiang, Alexandre Sablayrolles, Alexis Tacnet, Alok Kothari, Antoine Roux, Arthur Mensch, Audrey Herblin-Stoop, Augustin Garreau, Austin Birky, Bam4d, Baptiste Bout, Baudouin de Monicault, Blanche Savary, Carole Rambaud, Caroline Feldman, Devendra Singh Chaplot, Diego de las Casas, Eleonore Arcelin, Emma Bou Hanna, Etienne Metzger, Gaspard Blanchet, Gianna Lengyel, Guillaume Bour, Guillaume Lample, Harizo Rajaona, Henri Roussez, Hichem Sattouf, Ian Mack, Jean-Malo Delignon, Jessica Chudnovsky, Justus Murke, Kartik Khandelwal, Lawrence Stewart, Louis Martin, Louis Ternon, Lucile Saulnier, Lélio Renard Lavaud, Margaret Jennings, Marie Pellat, Marie Torelli, Marie-Anne Lachaux, Marjorie Janiewicz, Mickaël Seznec, Nicolas Schuhl, Niklas Muhs, Olivier de Garrigues, Patrick von Platen, Paul Jacob, Pauline Buche, Pavan Kumar Reddy, Perry Savas, Pierre Stock, Romain Sauvestre, Sagar Vaze, Sandeep Subramanian, Saurabh Garg, Sophia Yang, Szymon Antoniak, Teven Le Scao, Thibault Schueller, Thibaut Lavril, Thomas Wang, Théophile Gervet, Timothée Lacroix, Valera Nemychnikova, Wendy Shang, William El Sayed, William Marshall