plawanrath/minstral-7b-rust-fine-tuned

Model Card for Model ID

Minstral 7B parameter model fine-tuned on Rust code dataset.

Model Details

This is majorly fine-tuned on axum and async-graphql code.

Model Description

This took Minstral 7b parameter base model and fine-tuned it on Rust code majorly containing axum and async graphql code. It should be useful to assist in writing Rust backend code.

• Developed by: Plawan Rath
• Finetuned from model: Minstral-7B

Downstream Use [optional]

This can be locally used using GUI like LM Studio. To use it with LM Studio you can follow the following steps:

Converting to GGUF for LMStudio

1. Install llama.cpp

git clone https://github.com/ggerganov/llama.cpp
cd llama.cpp
pip install -r requirements.txt

2. Run convert-hf-to-gguf.py

python ./convert-hf-to-gguf.py \
   <path-to-merged-model> \
   --outfile ./mistral-lora-f16.gguf \
   --outtype f16

3. Put it in LM Studio Folder structure LM Studio doesn’t just scan every file directly inside ~/.lmstudio/models/. For each model it expects two nested folders:

~/.lmstudio/models/
└── <publisher-name>/
    └── <model-name>/
        └── model-file.gguf

4. Now refresh "My Models" in Lm Studio to see this model.

You should now be able to use this model to chat for codegen.