Instructions to use rizla/trrapi-16b with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries
PEFT
How to use rizla/trrapi-16b with PEFT:
```
Task type is invalid.
```

How to use rizla/trrapi-16b with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="rizla/trrapi-16b",
	filename="trrapi-q5km.gguf",
)

output = llm(
	"Once upon a time,",
	max_tokens=512,
	echo=True
)
print(output)

Notebooks
Google Colab
Kaggle
Local Apps

llama.cpp

How to use rizla/trrapi-16b with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf rizla/trrapi-16b
# Run inference directly in the terminal:
llama-cli -hf rizla/trrapi-16b

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf rizla/trrapi-16b
# Run inference directly in the terminal:
llama-cli -hf rizla/trrapi-16b

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf rizla/trrapi-16b
# Run inference directly in the terminal:
./llama-cli -hf rizla/trrapi-16b

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf rizla/trrapi-16b
# Run inference directly in the terminal:
./build/bin/llama-cli -hf rizla/trrapi-16b

Use Docker

docker model run hf.co/rizla/trrapi-16b

LM Studio
Jan
Ollama
How to use rizla/trrapi-16b with Ollama:
```
ollama run hf.co/rizla/trrapi-16b
```

Unsloth Studio new

How to use rizla/trrapi-16b with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for rizla/trrapi-16b to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for rizla/trrapi-16b to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for rizla/trrapi-16b to start chatting

Docker Model Runner
How to use rizla/trrapi-16b with Docker Model Runner:
```
docker model run hf.co/rizla/trrapi-16b
```

Lemonade

How to use rizla/trrapi-16b with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull rizla/trrapi-16b

Run and chat with the model

lemonade run user.trrapi-16b-{{QUANT_TAG}}

List all available models

lemonade list

qlora finetune of frankensteined rizla/rizla-17 model

The original Rizla models, already displaying promising multilingual capabilities, underwent multiple rounds of customization and optimizations to further enhance their versatility across languages. The process involves not only fine-tuned adjustments for better language comprehension but also strategic modifications to the underlying framework itself.

This continual refinement in response to specific requirements exemplifies a dynamic approach towards tackling natural language understanding tasks, where adaptability and flexibility are key factors contributing to performance improvements. In essence, these iterative advancements strive to bridge the gap between generalized pre-trained models and highly specialized applications.

*To run a localhost 127.0.0.1:8080 server with llama.cpp do


wget https://huggingface.co/rizla/trrapi-16b/resolve/main/trrapi-q5km.gguf

git clone https://github.com/ggerganov/llama.cpp && cd llama.cpp && make -j

./server -m ../trrapi-q5km.gguf --port 8080 -c 2000 -cb -t 8 -ngl 80