Instructions to use zeroxjason200/TheDrummer_Skyfall-31B-SLERP-v4.1_v4.2-60_40 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use zeroxjason200/TheDrummer_Skyfall-31B-SLERP-v4.1_v4.2-60_40 with Transformers:

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

pipe = pipeline("text-generation", model="zeroxjason200/TheDrummer_Skyfall-31B-SLERP-v4.1_v4.2-60_40")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForMultimodalLM

tokenizer = AutoTokenizer.from_pretrained("zeroxjason200/TheDrummer_Skyfall-31B-SLERP-v4.1_v4.2-60_40")
model = AutoModelForMultimodalLM.from_pretrained("zeroxjason200/TheDrummer_Skyfall-31B-SLERP-v4.1_v4.2-60_40")
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 Settings

vLLM

How to use zeroxjason200/TheDrummer_Skyfall-31B-SLERP-v4.1_v4.2-60_40 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "zeroxjason200/TheDrummer_Skyfall-31B-SLERP-v4.1_v4.2-60_40"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "zeroxjason200/TheDrummer_Skyfall-31B-SLERP-v4.1_v4.2-60_40",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/zeroxjason200/TheDrummer_Skyfall-31B-SLERP-v4.1_v4.2-60_40

SGLang

How to use zeroxjason200/TheDrummer_Skyfall-31B-SLERP-v4.1_v4.2-60_40 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 "zeroxjason200/TheDrummer_Skyfall-31B-SLERP-v4.1_v4.2-60_40" \
    --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": "zeroxjason200/TheDrummer_Skyfall-31B-SLERP-v4.1_v4.2-60_40",
		"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 "zeroxjason200/TheDrummer_Skyfall-31B-SLERP-v4.1_v4.2-60_40" \
        --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": "zeroxjason200/TheDrummer_Skyfall-31B-SLERP-v4.1_v4.2-60_40",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use zeroxjason200/TheDrummer_Skyfall-31B-SLERP-v4.1_v4.2-60_40 with Docker Model Runner:
```
docker model run hf.co/zeroxjason200/TheDrummer_Skyfall-31B-SLERP-v4.1_v4.2-60_40
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

TheDrummer_Skyfall-31B-SLERP-v4.1_v4.2-60_40

This is a merge of pre-trained language models created using mergekitty.

Merge Details

I like Skyfall 4.2, it had some nice new prose, felt more emotionally smart, but it had some 'rough' edges for me that the previous version didn't. This merge is exploring a mix of both to try and bring the new prose and emotional smarts to the older model with it's sharper scene and spatial memory.

I tried various percentages of each model and also different techniques, such as merging in increasing intensity across the layers, and this one turned out my favourite.

Merge Method

This model was merged using the SLERP merge method.

Models Merged

The following models were included in the merge:

Configuration

The following YAML configuration was used to produce this model:

slices:
  - sources:
      - model: TheDrummer/Skyfall-31B-v4.1
        layer_range: [0, 54]
      - model: TheDrummer/Skyfall-31B-v4.2
        layer_range: [0, 54]
merge_method: slerp
base_model: TheDrummer/Skyfall-31B-v4.1
parameters:
  t: 0.4
  nuslerp_flatten: false
dtype: bfloat16

Downloads last month: 3

Safetensors

Model size

31B params

Tensor type

BF16

Model tree for zeroxjason200/TheDrummer_Skyfall-31B-SLERP-v4.1_v4.2-60_40

TheDrummer/Skyfall-31B-v4.1

TheDrummer/Skyfall-31B-v4.2

Merge model

this model

Quantizations

1 model