metadata
library_name: transformers
inference:
parameters:
temperature: 1
top_p: 0.95
top_k: 40
repetition_penalty: 1.2
license: apache-2.0
language:
- en
pipeline_tag: text-generation
tags:
- TensorBlock
- GGUF
base_model: ministral/Ministral-3b-instruct
Feedback and support: TensorBlock's Twitter/X, Telegram Group and Discord server
ministral/Ministral-3b-instruct - GGUF
This repo contains GGUF format model files for ministral/Ministral-3b-instruct.
The files were quantized using machines provided by TensorBlock, and they are compatible with llama.cpp as of commit b4011.
Prompt template
<s>system
{system_prompt}</s>
<s>user
{prompt}</s>
<s>assistant
Model file specification
| Filename | Quant type | File Size | Description |
|---|---|---|---|
| Ministral-3b-instruct-Q2_K.gguf | Q2_K | 1.187 GB | smallest, significant quality loss - not recommended for most purposes |
| Ministral-3b-instruct-Q3_K_S.gguf | Q3_K_S | 1.376 GB | very small, high quality loss |
| Ministral-3b-instruct-Q3_K_M.gguf | Q3_K_M | 1.515 GB | very small, high quality loss |
| Ministral-3b-instruct-Q3_K_L.gguf | Q3_K_L | 1.644 GB | small, substantial quality loss |
| Ministral-3b-instruct-Q4_0.gguf | Q4_0 | 1.770 GB | legacy; small, very high quality loss - prefer using Q3_K_M |
| Ministral-3b-instruct-Q4_K_S.gguf | Q4_K_S | 1.778 GB | small, greater quality loss |
| Ministral-3b-instruct-Q4_K_M.gguf | Q4_K_M | 1.860 GB | medium, balanced quality - recommended |
| Ministral-3b-instruct-Q5_0.gguf | Q5_0 | 2.140 GB | legacy; medium, balanced quality - prefer using Q4_K_M |
| Ministral-3b-instruct-Q5_K_S.gguf | Q5_K_S | 2.140 GB | large, low quality loss - recommended |
| Ministral-3b-instruct-Q5_K_M.gguf | Q5_K_M | 2.187 GB | large, very low quality loss - recommended |
| Ministral-3b-instruct-Q6_K.gguf | Q6_K | 2.534 GB | very large, extremely low quality loss |
| Ministral-3b-instruct-Q8_0.gguf | Q8_0 | 3.282 GB | very large, extremely low quality loss - not recommended |
Downloading instruction
Command line
Firstly, install Huggingface Client
pip install -U "huggingface_hub[cli]"
Then, downoad the individual model file the a local directory
huggingface-cli download tensorblock/Ministral-3b-instruct-GGUF --include "Ministral-3b-instruct-Q2_K.gguf" --local-dir MY_LOCAL_DIR
If you wanna download multiple model files with a pattern (e.g., *Q4_K*gguf), you can try:
huggingface-cli download tensorblock/Ministral-3b-instruct-GGUF --local-dir MY_LOCAL_DIR --local-dir-use-symlinks False --include='*Q4_K*gguf'