|
|
--- |
|
|
library_name: llama.cpp |
|
|
pipeline_tag: text-generation |
|
|
tags: |
|
|
- gguf |
|
|
- conversational |
|
|
license: apache-2.0 |
|
|
base_model: Heralax/Augmentoolkit-DataSpecialist-v0.1 |
|
|
base_model_relation: quantized |
|
|
language: |
|
|
- en |
|
|
model-index: |
|
|
- name: Augmentoolkit-DataSpecialist-v0.1-GGUF |
|
|
results: [] |
|
|
--- |
|
|
|
|
|
# Augmentoolkit-DataSpecialist-v0.1-GGUF |
|
|
|
|
|
This repository contains GGUF quantizations of the model Heralax/Augmentoolkit-DataSpecialist-v0.1. |
|
|
Files are provided in GGUF format for llama.cpp-compatible runtimes (e.g., llama.cpp, LM Studio, KoboldCpp, etc.). |
|
|
|
|
|
## Source model |
|
|
|
|
|
- Base model: Heralax/Augmentoolkit-DataSpecialist-v0.1 (Apache-2.0). |
|
|
- Original base lineage (per upstream card): Heralax/datagen-pretrain-v1-7b-mistralv0.2 (Mistral 7B v0.2 derived). |
|
|
|
|
|
## Intended use |
|
|
|
|
|
Text generation and conversational assistance for data-specialist workflows. |
|
|
|
|
|
## Quantization notes |
|
|
|
|
|
- Filenames follow GGUF conventions including the quantization type suffix (e.g., Q4_K_M, Q5_K_M, Q8_0). |
|
|
- No additional fine-tuning was performed; weights are a direct quantization of the upstream safetensors. |
|
|
|
|
|
## Prompt template |
|
|
|
|
|
Use the chat template compatible with Mistral-style chat or your runtime’s default. |
|
|
|
|
|
## Hardware and runtimes |
|
|
|
|
|
Tested with llama.cpp and LM Studio on Apple Silicon. |
|
|
|
