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

Libraries

How to use TomGrc/FusionNet_passthrough with Transformers:

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

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

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("TomGrc/FusionNet_passthrough")
model = AutoModelForCausalLM.from_pretrained("TomGrc/FusionNet_passthrough")
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]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use TomGrc/FusionNet_passthrough with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/TomGrc/FusionNet_passthrough

SGLang

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

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

FusionNet_passthrough

Fine-tuned model on English language using passthrough Fusion method.

Model description

This is an experiment with the passthrough Fusion method of FusionNet. This model has 21.2B parameters, and this model is fine-tuned. Enjoy!

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	65.94
AI2 Reasoning Challenge (25-Shot)	69.45
HellaSwag (10-Shot)	87.72
MMLU (5-Shot)	65.28
TruthfulQA (0-shot)	67.65
Winogrande (5-shot)	81.29
GSM8k (5-shot)	24.26

Downloads last month: 119

Safetensors

Model size

21B params

Tensor type

F16

Model tree for TomGrc/FusionNet_passthrough

Quantizations

3 models

Evaluation results

normalized accuracy on AI2 Reasoning Challenge (25-Shot)
test set Open LLM Leaderboard

69.450
normalized accuracy on HellaSwag (10-Shot)
validation set Open LLM Leaderboard

87.720
accuracy on MMLU (5-Shot)
test set Open LLM Leaderboard

65.280
mc2 on TruthfulQA (0-shot)
validation set Open LLM Leaderboard

67.650
accuracy on Winogrande (5-shot)
validation set Open LLM Leaderboard

81.290
accuracy on GSM8k (5-shot)
test set Open LLM Leaderboard

24.260