Instructions to use mrcuddle/Dark-Hermes3-Llama3.2-3B-Func with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use mrcuddle/Dark-Hermes3-Llama3.2-3B-Func with Transformers:

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

pipe = pipeline("text-generation", model="mrcuddle/Dark-Hermes3-Llama3.2-3B-Func")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("mrcuddle/Dark-Hermes3-Llama3.2-3B-Func")
model = AutoModelForCausalLM.from_pretrained("mrcuddle/Dark-Hermes3-Llama3.2-3B-Func")
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 mrcuddle/Dark-Hermes3-Llama3.2-3B-Func with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "mrcuddle/Dark-Hermes3-Llama3.2-3B-Func"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "mrcuddle/Dark-Hermes3-Llama3.2-3B-Func",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/mrcuddle/Dark-Hermes3-Llama3.2-3B-Func

SGLang

How to use mrcuddle/Dark-Hermes3-Llama3.2-3B-Func 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 "mrcuddle/Dark-Hermes3-Llama3.2-3B-Func" \
    --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": "mrcuddle/Dark-Hermes3-Llama3.2-3B-Func",
		"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 "mrcuddle/Dark-Hermes3-Llama3.2-3B-Func" \
        --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": "mrcuddle/Dark-Hermes3-Llama3.2-3B-Func",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use mrcuddle/Dark-Hermes3-Llama3.2-3B-Func with Docker Model Runner:
```
docker model run hf.co/mrcuddle/Dark-Hermes3-Llama3.2-3B-Func
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

See axolotl config

axolotl version: 0.8.0.dev0

base_model: mrcuddle/Dark-Hermes3-Llama3.2-3B
hub_model_id: mrcuddle/Dark-Hermes3-Llama3.2-3B-Func
dataloader_num_workers: 8
datasets:
- chat_template: alpaca
  field_messages: conversations
  message_property_mappings:
    content: value
    role: from
  path: NousResearch/hermes-function-calling-v1
  split: train
  type: chat_template
eval_steps: 500
evaluation_strategy: steps
fp16: true
gradient_accumulation_steps: 4
gradient_checkpointing: true
learning_rate: 2e-5
logging_dir: /content/outputs/logs
logging_steps: 50
lr_scheduler: linear
lr_scheduler_type: linear
micro_batch_size: 2
num_train_epochs: 3
optimizer: adamw_torch  # Or another optimizer of your choice
output_dir: /content/outputs
overwrite_output_dir: true
per_device_train_batch_size: 8
save_steps: 500
save_total_limit: 2
use_peft: false
val_set_size: 0.05
warmup_steps: 100
unsloth: true  # Enable Unsloth if supported by your training framework

Dark-Hermes3-Llama3.2-3B-Func

This model is a fine-tuned version of mrcuddle/Dark-Hermes3-Llama3.2-3B on the NousResearch/hermes-function-calling-v1 dataset.

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 2e-05
train_batch_size: 2
eval_batch_size: 2
seed: 42
gradient_accumulation_steps: 4
total_train_batch_size: 8
optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 100
num_epochs: 1.0

Training results

Training Loss	Epoch	Step	Validation Loss
No log	0.0889	1	0.3864

Framework versions

Transformers 4.49.0
Pytorch 2.5.1+cu124
Datasets 3.2.0
Tokenizers 0.21.0

Downloads last month: 3

Safetensors

Model size

3B params

Tensor type

F32

BF16

Model tree for mrcuddle/Dark-Hermes3-Llama3.2-3B-Func

Base model

NousResearch/Hermes-3-Llama-3.2-3B

Finetuned

mrcuddle/Dark-Hermes3-Llama3.2-3B

Finetuned

(5)

this model

Quantizations

2 models

mrcuddle
/

Dark-Hermes3-Llama3.2-3B-Func

Dark-Hermes3-Llama3.2-3B-Func

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

Model tree for mrcuddle/Dark-Hermes3-Llama3.2-3B-Func

Dataset used to train mrcuddle/Dark-Hermes3-Llama3.2-3B-Func