Instructions to use bralynn/omnim with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use bralynn/omnim with Transformers:

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

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

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("bralynn/omnim")
model = AutoModelForCausalLM.from_pretrained("bralynn/omnim", 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=256)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use bralynn/omnim with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/bralynn/omnim

SGLang

How to use bralynn/omnim 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 "bralynn/omnim" \
    --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": "bralynn/omnim",
		"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 "bralynn/omnim" \
        --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": "bralynn/omnim",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Unsloth Desktop
Docker Model Runner
How to use bralynn/omnim with Docker Model Runner:
```
docker model run hf.co/bralynn/omnim
```

Uploaded finetuned model

Developed by: bralynn
License: apache-2.0
Finetuned from model : huihui-ai/Llama-3.2-3B-Instruct-abliterated

This llama model was trained 2x faster with Unsloth and Huggingface's TRL library.

hf (pretrained=bralynn/omnim,trust_remote_code=True), gen_kwargs: (None), limit: None, num_fewshot: 0, batch_size: auto (16)

Tasks	Version	Filter	n-shot	Metric		Value		Stderr
mmlu	2	none		acc	↑	0.6000	±	0.0040
- humanities	2	none		acc	↑	0.5768	±	0.0070
- formal_logic	1	none	0	acc	↑	0.4206	±	0.0442
- high_school_european_history	1	none	0	acc	↑	0.7273	±	0.0348
- high_school_us_history	1	none	0	acc	↑	0.7304	±	0.0311
- high_school_world_history	1	none	0	acc	↑	0.7890	±	0.0266
- international_law	1	none	0	acc	↑	0.7107	±	0.0414
- jurisprudence	1	none	0	acc	↑	0.6296	±	0.0467
- logical_fallacies	1	none	0	acc	↑	0.7239	±	0.0351
- moral_disputes	1	none	0	acc	↑	0.6532	±	0.0256
- moral_scenarios	1	none	0	acc	↑	0.5196	±	0.0167
- philosophy	1	none	0	acc	↑	0.6592	±	0.0269
- prehistory	1	none	0	acc	↑	0.6636	±	0.0263
- professional_law	1	none	0	acc	↑	0.4518	±	0.0127
- world_religions	1	none	0	acc	↑	0.7544	±	0.0330
- other	2	none		acc	↑	0.6662	±	0.0082
- business_ethics	1	none	0	acc	↑	0.5500	±	0.0500
- clinical_knowledge	1	none	0	acc	↑	0.6415	±	0.0295
- college_medicine	1	none	0	acc	↑	0.5954	±	0.0374
- global_facts	1	none	0	acc	↑	0.4000	±	0.0492
- human_aging	1	none	0	acc	↑	0.5964	±	0.0329
- management	1	none	0	acc	↑	0.7670	±	0.0419
- marketing	1	none	0	acc	↑	0.8462	±	0.0236
- medical_genetics	1	none	0	acc	↑	0.7000	±	0.0461
- miscellaneous	1	none	0	acc	↑	0.7586	±	0.0153
- nutrition	1	none	0	acc	↑	0.6732	±	0.0269
- professional_accounting	1	none	0	acc	↑	0.4752	±	0.0298
- professional_medicine	1	none	0	acc	↑	0.7757	±	0.0253
- virology	1	none	0	acc	↑	0.4639	±	0.0388
- social sciences	2	none		acc	↑	0.6640	±	0.0083
- econometrics	1	none	0	acc	↑	0.3947	±	0.0460
- high_school_geography	1	none	0	acc	↑	0.7121	±	0.0323
- high_school_government_and_politics	1	none	0	acc	↑	0.7668	±	0.0305
- high_school_macroeconomics	1	none	0	acc	↑	0.5744	±	0.0251
- high_school_microeconomics	1	none	0	acc	↑	0.6218	±	0.0315
- high_school_psychology	1	none	0	acc	↑	0.7835	±	0.0177
- human_sexuality	1	none	0	acc	↑	0.6412	±	0.0421
- professional_psychology	1	none	0	acc	↑	0.5866	±	0.0199
- public_relations	1	none	0	acc	↑	0.6455	±	0.0458
- security_studies	1	none	0	acc	↑	0.6367	±	0.0308
- sociology	1	none	0	acc	↑	0.7861	±	0.0290
- us_foreign_policy	1	none	0	acc	↑	0.8200	±	0.0386
- stem	2	none		acc	↑	0.5068	±	0.0086
- abstract_algebra	1	none	0	acc	↑	0.2700	±	0.0446
- anatomy	1	none	0	acc	↑	0.6370	±	0.0415
- astronomy	1	none	0	acc	↑	0.6579	±	0.0386
- college_biology	1	none	0	acc	↑	0.7222	±	0.0375
- college_chemistry	1	none	0	acc	↑	0.4100	±	0.0494
- college_computer_science	1	none	0	acc	↑	0.4300	±	0.0498
- college_mathematics	1	none	0	acc	↑	0.3000	±	0.0461
- college_physics	1	none	0	acc	↑	0.3627	±	0.0478
- computer_security	1	none	0	acc	↑	0.6600	±	0.0476
- conceptual_physics	1	none	0	acc	↑	0.5064	±	0.0327
- electrical_engineering	1	none	0	acc	↑	0.5448	±	0.0415
- elementary_mathematics	1	none	0	acc	↑	0.4233	±	0.0254
- high_school_biology	1	none	0	acc	↑	0.7194	±	0.0256
- high_school_chemistry	1	none	0	acc	↑	0.5567	±	0.0350
- high_school_computer_science	1	none	0	acc	↑	0.5800	±	0.0496
- high_school_mathematics	1	none	0	acc	↑	0.3630	±	0.0293
- high_school_physics	1	none	0	acc	↑	0.4238	±	0.0403
- high_school_statistics	1	none	0	acc	↑	0.4444	±	0.0339
- machine_learning	1	none	0	acc	↑	0.4821	±	0.0474

Groups	Version	Filter	Metric		Value		Stderr
mmlu	2	none	acc	↑	0.6000	±	0.0040
- humanities	2	none	acc	↑	0.5768	±	0.0070
- other	2	none	acc	↑	0.6662	±	0.0082
- social sciences	2	none	acc	↑	0.6640	±	0.0083
- stem	2	none	acc	↑	0.5068	±	0.0086

Downloads last month: 19

Safetensors

Model size

4B params

Tensor type

BF16

Model tree for bralynn/omnim

Base model

meta-llama/Llama-3.2-3B-Instruct

Finetuned

huihui-ai/Llama-3.2-3B-Instruct-abliterated

Finetuned

(20)

this model

Finetunes

3 models