Instructions to use KaraKaraWarehouse/UnFimbulvetr-20B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use KaraKaraWarehouse/UnFimbulvetr-20B with Transformers:

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

pipe = pipeline("text-generation", model="KaraKaraWarehouse/UnFimbulvetr-20B")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("KaraKaraWarehouse/UnFimbulvetr-20B")
model = AutoModelForCausalLM.from_pretrained("KaraKaraWarehouse/UnFimbulvetr-20B", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use KaraKaraWarehouse/UnFimbulvetr-20B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "KaraKaraWarehouse/UnFimbulvetr-20B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "KaraKaraWarehouse/UnFimbulvetr-20B",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/KaraKaraWarehouse/UnFimbulvetr-20B

SGLang

How to use KaraKaraWarehouse/UnFimbulvetr-20B 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 "KaraKaraWarehouse/UnFimbulvetr-20B" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "KaraKaraWarehouse/UnFimbulvetr-20B",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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 "KaraKaraWarehouse/UnFimbulvetr-20B" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "KaraKaraWarehouse/UnFimbulvetr-20B",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use KaraKaraWarehouse/UnFimbulvetr-20B with Docker Model Runner:
```
docker model run hf.co/KaraKaraWarehouse/UnFimbulvetr-20B
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

UnFimbulvetr-20B

Waifu to catch your attention

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

NOTE: Only tested this just for a bit. YMMV.

Next Day Tests...

Downloaded the GGUF model that someone quantized... And... nope. No.

Do not use model.

Merge Details

Merge Method

This model was merged using the passthrough merge method.

Models Merged

The following models were included in the merge:

Sao10K/Fimbulvetr-11B-v2

Configuration

The following YAML configuration was used to produce this model:

slices:
  - sources:
    - model: FimbMagic
      layer_range: [0, 13]
  - sources:
    - model: FimbMagic
      layer_range: [8, 13]
  - sources:
    - model: FimbMagic
      layer_range: [12, 36]
  - sources:
    - model: FimbMagic
      layer_range: [12, 36]
  - sources:
    - model: FimbMagic
      layer_range: [36, 48]
  - sources:
    - model: FimbMagic
      layer_range: [36, 48]
merge_method: passthrough
dtype: bfloat16

Additional Notes

Fimbulvetr 11B is still a very good model. This model is for extreme trailblazers who wants to test stuff!

Eval results? Don't bother.

Last one before I sleep: I'm so sorry Sao10K...

Downloads last month: 14

Safetensors

Model size

20B params

Tensor type

BF16

Model tree for KaraKaraWarehouse/UnFimbulvetr-20B

Base model

Sao10K/Fimbulvetr-11B-v2

Finetuned

(10)

this model

Quantizations

1 model