Instructions to use OddTheGreat/Machina_24B.V2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use OddTheGreat/Machina_24B.V2 with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="OddTheGreat/Machina_24B.V2")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("OddTheGreat/Machina_24B.V2")
model = AutoModelForCausalLM.from_pretrained("OddTheGreat/Machina_24B.V2", device_map="auto")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Inference
Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use OddTheGreat/Machina_24B.V2 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "OddTheGreat/Machina_24B.V2"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "OddTheGreat/Machina_24B.V2",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/OddTheGreat/Machina_24B.V2

SGLang

How to use OddTheGreat/Machina_24B.V2 with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "OddTheGreat/Machina_24B.V2" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "OddTheGreat/Machina_24B.V2",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "OddTheGreat/Machina_24B.V2" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "OddTheGreat/Machina_24B.V2",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use OddTheGreat/Machina_24B.V2 with Docker Model Runner:
```
docker model run hf.co/OddTheGreat/Machina_24B.V2
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

merge

This is a merge of pre-trained language models.

This merge is better than v1. v1 machina will be deleted tomorrow.

Has neutral or negative bias, if prompt is good enough model can be evil and cruel. maybe even too cruel.

Main idea is to merge base mistral with something, and preserve base's ability to write on russian.

Secondary idea was to create model that will be not "friendship is magic" thing, to use in more "dark" scenarios.

Judging by the tests, i've succeeded with both ideas.

Model was tested on russian and on english with ~1000 responses, was stable and seems that it not lose original's "intellectual abilities", but maybe i'm just very lucky.

With char cards that have first message on russian, but whole description on english, model have some difficulties, but still able to answer on desired language.

With full ru cards model performs without issues.

Tested on T1.01 ChatML

I reccomend to use sphiratrioth666/SillyTavern-Presets-Sphiratrioth, for me it works very good with minor adjustments.