Triangle104/VersatiLlama-Llama-3.2-3B-Instruct-Abliterated-Q8_0-GGUF

This model was converted to GGUF format from Devarui379/VersatiLlama-Llama-3.2-3B-Instruct-Abliterated using llama.cpp via the ggml.ai's GGUF-my-repo space. Refer to the original model card for more details on the model.

---

Model details:

Small but Smart

Fine-Tuned on Vast dataset of Conversations

Able to Generate Human like text with high performance within its size.

It is Very Versatile when compared for it's size and Parameters and offers capability almost as good as Llama 3.1 8B Instruct

Feel free to Check it out!!

[This model was trained for 5hrs on GPU T4 15gb vram]

Developed by: Meta AI
Fine-Tuned by: Devarui379
Model type: Transformers
Language(s) (NLP): English
License: cc-by-4.0

Model Sources [optional]

base model:meta-llama/Llama-3.2-3B-Instruct

Repository: Devarui379/VersatiLlama-Llama-3.2-3B-Instruct-Abliterated
Demo: Use LM Studio with the Quantized version

Uses

Use desired System prompt when using in LM Studio The optimal chat template seems to be Jinja but feel free to test it out as you want!

Technical Specifications

Model Architecture and Objective

Llama 3.2

Hardware

NVIDIA TESLA T4 GPU 15GB VRAM

---

Use with llama.cpp

Install llama.cpp through brew (works on Mac and Linux)

brew install llama.cpp

Invoke the llama.cpp server or the CLI.

CLI:

llama-cli --hf-repo Triangle104/VersatiLlama-Llama-3.2-3B-Instruct-Abliterated-Q8_0-GGUF --hf-file versatillama-llama-3.2-3b-instruct-abliterated-q8_0.gguf -p "The meaning to life and the universe is"

Server:

llama-server --hf-repo Triangle104/VersatiLlama-Llama-3.2-3B-Instruct-Abliterated-Q8_0-GGUF --hf-file versatillama-llama-3.2-3b-instruct-abliterated-q8_0.gguf -c 2048

Note: You can also use this checkpoint directly through the usage steps listed in the Llama.cpp repo as well.

Step 1: Clone llama.cpp from GitHub.

git clone https://github.com/ggerganov/llama.cpp

Step 2: Move into the llama.cpp folder and build it with LLAMA_CURL=1 flag along with other hardware-specific flags (for ex: LLAMA_CUDA=1 for Nvidia GPUs on Linux).

cd llama.cpp && LLAMA_CURL=1 make

Step 3: Run inference through the main binary.

./llama-cli --hf-repo Triangle104/VersatiLlama-Llama-3.2-3B-Instruct-Abliterated-Q8_0-GGUF --hf-file versatillama-llama-3.2-3b-instruct-abliterated-q8_0.gguf -p "The meaning to life and the universe is"

or

./llama-server --hf-repo Triangle104/VersatiLlama-Llama-3.2-3B-Instruct-Abliterated-Q8_0-GGUF --hf-file versatillama-llama-3.2-3b-instruct-abliterated-q8_0.gguf -c 2048