Instructions to use KaraKaraWarehouse/Llama-EveningMirai-3.3-70B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use KaraKaraWarehouse/Llama-EveningMirai-3.3-70B with Transformers:

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

pipe = pipeline("text-generation", model="KaraKaraWarehouse/Llama-EveningMirai-3.3-70B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("KaraKaraWarehouse/Llama-EveningMirai-3.3-70B")
model = AutoModelForCausalLM.from_pretrained("KaraKaraWarehouse/Llama-EveningMirai-3.3-70B", 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=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Inference
Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use KaraKaraWarehouse/Llama-EveningMirai-3.3-70B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "KaraKaraWarehouse/Llama-EveningMirai-3.3-70B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "KaraKaraWarehouse/Llama-EveningMirai-3.3-70B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/KaraKaraWarehouse/Llama-EveningMirai-3.3-70B

SGLang

How to use KaraKaraWarehouse/Llama-EveningMirai-3.3-70B 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/Llama-EveningMirai-3.3-70B" \
    --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": "KaraKaraWarehouse/Llama-EveningMirai-3.3-70B",
		"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 "KaraKaraWarehouse/Llama-EveningMirai-3.3-70B" \
        --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": "KaraKaraWarehouse/Llama-EveningMirai-3.3-70B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use KaraKaraWarehouse/Llama-EveningMirai-3.3-70B with Docker Model Runner:
```
docker model run hf.co/KaraKaraWarehouse/Llama-EveningMirai-3.3-70B
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

L3.3-MiraiFanfare-Evening

Good evening. Welcome to Mirai Fanfare (Evening Edition / Evening Timeslot)

Talking to a bunch of folks on featherless.ai discord, a lot still remember Mirai Fanfare because of how unhinged it was.
This merge is partly a continuation of Mirai 2.5 with some additional modern models such as Wayfarer and Asobi (Asobi isn't new I know).

What's new this time is the base switch to Forgotten Safeword for more of your deranged needs.

Not much else to say other than to test it...

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

Model Vibes

Man is this model super HORNY when you give it the chance!
Acronyms: Bends and twists words. Which is a first from me.
Tends to write more novel/story like sentences.
Retains formatting for character focused dialogues. Seems to be quite good if you want to write game dialogues.
JP to English translations seems okay, but nothing novel or super interesting.

Merge Details

Merge Method

This model was merged using the TIES merge method using ReadyArt/Forgotten-Safeword-70B-v5.0 as a base.

Models Merged

The following models were included in the merge:

Configuration

The following YAML configuration was used to produce this model:

models:
  - model: Blackroot/Mirai-70B-1.0
    parameters:
      density: 0.4
      weight: 0.35
  - model: EVA-UNIT-01/EVA-LLaMA-3.33-70B-v0.0
    parameters:
      density: 0.7
      weight: 1
  - model: LatitudeGames/Wayfarer-Large-70B-Llama-3.3
    parameters:
      density: 0.4
      weight: 0.6
  - model: nbeerbower/llama3.1-kartoffeldes-70B
    parameters:
      density: 0.66
      weight: 0.3
  - model: nbeerbower/Llama3-Asobi-70B
    parameters:
      density: 0.5
      weight: 0.7

merge_method: ties
base_model: ReadyArt/Forgotten-Safeword-70B-v5.0
parameters:
  normalize: true
dtype: bfloat16