Instructions to use TeeZee/Kyllene-34B-v1.1 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use TeeZee/Kyllene-34B-v1.1 with Transformers:

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

pipe = pipeline("text-generation", model="TeeZee/Kyllene-34B-v1.1")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("TeeZee/Kyllene-34B-v1.1")
model = AutoModelForCausalLM.from_pretrained("TeeZee/Kyllene-34B-v1.1")
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 TeeZee/Kyllene-34B-v1.1 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "TeeZee/Kyllene-34B-v1.1"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "TeeZee/Kyllene-34B-v1.1",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/TeeZee/Kyllene-34B-v1.1

SGLang

How to use TeeZee/Kyllene-34B-v1.1 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 "TeeZee/Kyllene-34B-v1.1" \
    --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": "TeeZee/Kyllene-34B-v1.1",
		"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 "TeeZee/Kyllene-34B-v1.1" \
        --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": "TeeZee/Kyllene-34B-v1.1",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use TeeZee/Kyllene-34B-v1.1 with Docker Model Runner:
```
docker model run hf.co/TeeZee/Kyllene-34B-v1.1
```

What is this?

by brucethemoose - opened Jan 21, 2024

Discussion

brucethemoose

Jan 21, 2024

•

edited Jan 21, 2024

A bagel finetune?

TeeZee

Owner Jan 21, 2024

A merge of 4 models, details in here: https://huggingface.co/TeeZee/Kyllene-34B-v1.1/blob/main/merge-config.yml, proper readme on the way.

brucethemoose

Jan 22, 2024

Oh I see, this is mergemonster? Never used it before, but still, interesting.

TeeZee

Owner Jan 24, 2024

Yes, I wanted to try something different, it kind of works ;). README is ready, if you want, take a look at my another model for RP/ERP https://huggingface.co/TeeZee/DarkForest-20B-v1.0

Nexesenex

Jan 28, 2024

•

edited Jan 29, 2024

For information, guys, I benched your models :

Bruce recent merges :

yi-34b-200k-dare-merge-v5.Q4_K_M.gguf,-,Hellaswag,85.75,,400,2024-01-26 01:40:00,,34b,Yi,200000,,,GGUF,Brucethemoose,TheBloke,
yi-34b-200k-dare-merge-v5.Q4_K_M.gguf,-,Hellaswag_Bin,78.5,,400,2024-01-26 01:40:00,,34b,Yi,200000,,,GGUF,Brucethemoose,TheBloke,
yi-34b-200k-dare-merge-v5.Q4_K_M.gguf,-,Arc-Challenge,58.52842809,,299,2024-01-26 05:40:00,,34b,Yi,200000,,,GGUF,Brucethemoose,TheBloke,
yi-34b-200k-dare-merge-v5.Q4_K_M.gguf,-,Arc-Easy,80.87719298,,570,2024-01-26 05:40:00,,34b,Yi,200000,,,GGUF,Brucethemoose,TheBloke,
yi-34b-200k-dare-merge-v5.Q4_K_M.gguf,-,MMLU,39.93610224,,313,2024-01-26 05:40:00,,34b,Yi,200000,,,GGUF,Brucethemoose,TheBloke,
yi-34b-200k-dare-merge-v5.Q4_K_M.gguf,-,Thruthful-QA,33.65973072,,817,2024-01-26 05:40:00,,34b,Yi,200000,,,GGUF,Brucethemoose,TheBloke,
yi-34b-200k-dare-merge-v5.Q4_K_M.gguf,-,Winogrande,77.0324,,1267,2024-01-26 05:40:00,,34b,Yi,200000,,,GGUF,Brucethemoose,TheBloke,
yi-34b-200k-dare-merge-v5.Q4_K_M.gguf,-,wikitext,5.5128,512,512,2024-01-26 01:40:00,,34b,Yi,200000,,,GGUF,Brucethemoose,TheBloke,
yi-34b-200k-dare-merge-v5.Q4_K_M.gguf,-,wikitext,4.6294,4096,4096,2024-01-26 01:40:00,,34b,Yi,200000,,,GGUF,Brucethemoose,TheBloke,

Yi-34B-200K-DARE-merge-v7-AR-b1952-iMat-c32_ch2500-Q4_K_M.gguf,-,Hellaswag,85.25,,400,2024-01-26 01:40:00,,34b,Yi,200000,,,GGUF,Brucethemoose,Nexesenex,
Yi-34B-200K-DARE-merge-v7-AR-b1952-iMat-c32_ch2500-Q4_K_M.gguf,-,Hellaswag_Bin,80,,400,2024-01-26 01:40:00,,34b,Yi,200000,,,GGUF,Brucethemoose,Nexesenex,
Yi-34B-200K-DARE-merge-v7-AR-b1952-iMat-c32_ch2500-Q4_K_M.gguf,-,Arc-Challenge,57.19063545,,299,2024-01-26 05:40:00,,34b,Yi,200000,,,GGUF,Brucethemoose,Nexesenex,
Yi-34B-200K-DARE-merge-v7-AR-b1952-iMat-c32_ch2500-Q4_K_M.gguf,-,Arc-Easy,79.12280702,,570,2024-01-26 05:40:00,,34b,Yi,200000,,,GGUF,Brucethemoose,Nexesenex,
Yi-34B-200K-DARE-merge-v7-AR-b1952-iMat-c32_ch2500-Q4_K_M.gguf,-,MMLU,38.91285591,,1159,2024-01-26 05:40:00,,34b,Yi,200000,,,GGUF,Brucethemoose,Nexesenex,
Yi-34B-200K-DARE-merge-v7-AR-b1952-iMat-c32_ch2500-Q4_K_M.gguf,-,Thruthful-QA,33.41493268,19.8590,817,2024-01-26 05:40:00,,34b,Yi,200000,,,GGUF,Brucethemoose,Nexesenex,
Yi-34B-200K-DARE-merge-v7-AR-b1952-iMat-c32_ch2500-Q4_K_M.gguf,-,Winogrande,78.1373,,1267,2024-01-26 05:40:00,,34b,Yi,200000,,,GGUF,Brucethemoose,Nexesenex,
Yi-34B-200K-DARE-merge-v7-AR-b1952-iMat-c32_ch2500-Q4_K_M.gguf,-,wikitext,5.1353,512,512,2024-01-26 01:40:00,,34b,Yi,200000,,,GGUF,Brucethemoose,Nexesenex,
Yi-34B-200K-DARE-merge-v7-AR-b1952-iMat-c32_ch2500-Q4_K_M.gguf,-,wikitext,4.5414,2048,2048,2024-01-26 01:40:00,,34b,Yi,200000,,,GGUF,Brucethemoose,Nexesenex,
Yi-34B-200K-DARE-merge-v7-AR-b1952-iMat-c32_ch2500-Q4_K_M.gguf,-,wikitext,4.3967,4096,4096,2024-01-26 01:40:00,,34b,Yi,200000,,,GGUF,Brucethemoose,Nexesenex,
Yi-34B-200K-DARE-merge-v7-AR-b1952-iMat-c32_ch2500-Q4_K_M.gguf,-,wikitext,4.4457,8192,8192,2024-01-26 01:40:00,,34b,Yi,200000,,,GGUF,Brucethemoose,Nexesenex,

Yi-34B-200K-DARE-megamerge-v8-b1952-iMat-c32_ch2500-Q4_K_M.gguf,-,Hellaswag,84.5,,400,2024-01-26 01:40:00,,34b,Yi,200000,,,GGUF,Brucethemoose,Nexesenex,
Yi-34B-200K-DARE-megamerge-v8-b1952-iMat-c32_ch2500-Q4_K_M.gguf,-,Hellaswag_Bin,79,,400,2024-01-26 01:40:00,,34b,Yi,200000,,,GGUF,Brucethemoose,Nexesenex,
Yi-34B-200K-DARE-megamerge-v8-b1952-iMat-c32_ch2500-Q4_K_M.gguf,-,Arc-Challenge,57.52508361,,299,2024-01-26 05:40:00,,34b,Yi,200000,,,GGUF,Brucethemoose,Nexesenex,
Yi-34B-200K-DARE-megamerge-v8-b1952-iMat-c32_ch2500-Q4_K_M.gguf,-,Arc-Easy,78.59649123,,570,2024-01-26 05:40:00,,34b,Yi,200000,,,GGUF,Brucethemoose,Nexesenex,
Yi-34B-200K-DARE-megamerge-v8-b1952-iMat-c32_ch2500-Q4_K_M.gguf,-,MMLU,40.89456869,,313,2024-01-26 05:40:00,,34b,Yi,200000,,,GGUF,Brucethemoose,Nexesenex,
Yi-34B-200K-DARE-megamerge-v8-b1952-iMat-c32_ch2500-Q4_K_M.gguf,-,Thruthful-QA,34.76132191,,817,2024-01-26 05:40:00,,34b,Yi,200000,,,GGUF,Brucethemoose,Nexesenex,
Yi-34B-200K-DARE-megamerge-v8-b1952-iMat-c32_ch2500-Q4_K_M.gguf,-,Winogrande,77.9795,,1267,2024-01-26 05:40:00,,34b,Yi,200000,,,GGUF,Brucethemoose,Nexesenex,
Yi-34B-200K-DARE-megamerge-v8-b1952-iMat-c32_ch2500-Q4_K_M.gguf,-,wikitext,5.0681,512,512,2024-01-26 01:40:00,,34b,Yi,200000,,,GGUF,Brucethemoose,Nexesenex,
Yi-34B-200K-DARE-megamerge-v8-b1952-iMat-c32_ch2500-Q4_K_M.gguf,-,wikitext,4.5052,2048,2048,2024-01-26 01:40:00,,34b,Yi,200000,,,GGUF,Brucethemoose,Nexesenex,
Yi-34B-200K-DARE-megamerge-v8-b1952-iMat-c32_ch2500-Q4_K_M.gguf,-,wikitext,4.3656,4096,4096,2024-01-26 01:40:00,,34b,Yi,200000,,,GGUF,Brucethemoose,Nexesenex,
Yi-34B-200K-DARE-megamerge-v8-b1952-iMat-c32_ch2500-Q4_K_M.gguf,-,wikitext,4.3190,8192,8192,2024-01-26 01:40:00,,34b,Yi,200000,,,GGUF,Brucethemoose,Nexesenex,

Teezee :

Kyllene-34B-v1.1-b1989-iMat-c32_ch3250-Q4_K_S.gguf,-,Hellaswag,85,,400,2024-01-28 01:40:00,,34b,Yi,200000,,,GGUF,TeeZee,Nexesenex,
Kyllene-34B-v1.1-b1989-iMat-c32_ch3250-Q4_K_S.gguf,-,Hellaswag,85.2,,1000,2024-01-28 01:40:00,,34b,Yi,200000,,,GGUF,TeeZee,Nexesenex,
Kyllene-34B-v1.1-b1989-iMat-c32_ch3250-Q4_K_S.gguf,-,Hellaswag,84.6,,2000,2024-01-28 01:40:00,,34b,Yi,200000,,,GGUF,TeeZee,Nexesenex,
Kyllene-34B-v1.1-b1989-iMat-c32_ch3250-Q4_K_S.gguf,-,Hellaswag_Bin,81,,400,2024-01-28 01:40:00,,34b,Yi,200000,,,GGUF,TeeZee,Nexesenex,
Kyllene-34B-v1.1-b1989-iMat-c32_ch3250-Q4_K_S.gguf,-,Hellaswag_Bin,83.5,,1000,2024-01-28 01:40:00,,34b,Yi,200000,,,GGUF,TeeZee,Nexesenex,
Kyllene-34B-v1.1-b1989-iMat-c32_ch3250-Q4_K_S.gguf,-,Hellaswag_Bin,82.95,,2000,2024-01-28 01:40:00,,34b,Yi,200000,,,GGUF,TeeZee,Nexesenex,
Kyllene-34B-v1.1-b1989-iMat-c32_ch3250-Q4_K_S.gguf,-,Arc-Challenge,61.53846154,,299,2024-01-28 05:40:00,,34b,Yi,200000,,,GGUF,TeeZee,Nexesenex,
Kyllene-34B-v1.1-b1989-iMat-c32_ch3250-Q4_K_S.gguf,-,Arc-Easy,80.35087719,,570,2024-01-28 05:40:00,,34b,Yi,200000,,,GGUF,TeeZee,Nexesenex,
Kyllene-34B-v1.1-b1989-iMat-c32_ch3250-Q4_K_S.gguf,-,MMLU,43.13099042,,313,2024-01-28 05:40:00,,34b,Yi,200000,,,GGUF,TeeZee,Nexesenex,
Kyllene-34B-v1.1-b1989-iMat-c32_ch3250-Q4_K_S.gguf,-,Thruthful-QA,35.00611995,,817,2024-01-28 05:40:00,,34b,Yi,200000,,,GGUF,TeeZee,Nexesenex,
Kyllene-34B-v1.1-b1989-iMat-c32_ch3250-Q4_K_S.gguf,-,Winogrande,79.3212,,1267,2024-01-28 05:40:00,,34b,Yi,200000,,,GGUF,TeeZee,Nexesenex,
Kyllene-34B-v1.1-b1989-iMat-c32_ch3250-Q4_K_S.gguf,-,wikitext,5.1703,512,512,2024-01-28 01:40:00,,34b,Yi,200000,,,GGUF,TeeZee,Nexesenex,

Kyllene-34B-v1.1-b1989-iMat-c32_ch3250-Q4_K_M.gguf,-,Hellaswag,84.75,,400,2024-01-28 01:40:00,,34b,Yi,200000,,,GGUF,TeeZee,Nexesenex,
Kyllene-34B-v1.1-b1989-iMat-c32_ch3250-Q4_K_M.gguf,-,Hellaswag,85.6,,1000,2024-01-28 01:40:00,,34b,Yi,200000,,,GGUF,TeeZee,Nexesenex,
Kyllene-34B-v1.1-b1989-iMat-c32_ch3250-Q4_K_M.gguf,-,Hellaswag,84.9,,2000,2024-01-28 01:40:00,,34b,Yi,200000,,,GGUF,TeeZee,Nexesenex,
Kyllene-34B-v1.1-b1989-iMat-c32_ch3250-Q4_K_M.gguf,-,Hellaswag_Bin,81,,400,2024-01-28 01:40:00,,34b,Yi,200000,,,GGUF,TeeZee,Nexesenex,
Kyllene-34B-v1.1-b1989-iMat-c32_ch3250-Q4_K_M.gguf,-,Hellaswag_Bin,83.4,,1000,2024-01-28 01:40:00,,34b,Yi,200000,,,GGUF,TeeZee,Nexesenex,
Kyllene-34B-v1.1-b1989-iMat-c32_ch3250-Q4_K_M.gguf,-,Hellaswag_Bin,82.9,,2000,2024-01-28 01:40:00,,34b,Yi,200000,,,GGUF,TeeZee,Nexesenex,
Kyllene-34B-v1.1-b1989-iMat-c32_ch3250-Q4_K_M.gguf,-,Arc-Challenge,60.53511706,,299,2024-01-28 05:40:00,,34b,Yi,200000,,,GGUF,TeeZee,Nexesenex,
Kyllene-34B-v1.1-b1989-iMat-c32_ch3250-Q4_K_M.gguf,-,Arc-Easy,80.52631579,,570,2024-01-28 05:40:00,,34b,Yi,200000,,,GGUF,TeeZee,Nexesenex,
Kyllene-34B-v1.1-b1989-iMat-c32_ch3250-Q4_K_M.gguf,-,MMLU,42.49201278,,313,2024-01-28 05:40:00,,34b,Yi,200000,,,GGUF,TeeZee,Nexesenex,
Kyllene-34B-v1.1-b1989-iMat-c32_ch3250-Q4_K_M.gguf,-,Thruthful-QA,34.39412485,,817,2024-01-28 05:40:00,,34b,Yi,200000,,,GGUF,TeeZee,Nexesenex,
Kyllene-34B-v1.1-b1989-iMat-c32_ch3250-Q4_K_M.gguf,-,Winogrande,79.4791,,1267,2024-01-28 05:40:00,,34b,Yi,200000,,,GGUF,TeeZee,Nexesenex,
Kyllene-34B-v1.1-b1989-iMat-c32_ch3250-Q4_K_M.gguf,-,wikitext,5.1679,512,512,2024-01-28 01:40:00,,34b,Yi,200000,,,GGUF,TeeZee,Nexesenex,
Kyllene-34B-v1.1-b1989-iMat-c32_ch3250-Q4_K_M.gguf,-,wikitext,4.3623,4096,4096,2024-01-28 01:40:00,,34b,Yi,200000,,,GGUF,TeeZee,Nexesenex,
Kyllene-34B-v1.1-b1989-iMat-c32_ch3250-Q4_K_M.gguf,-,wikitext,4.4061,8192,8192,2024-01-28 01:40:00,,34b,Yi,200000,,,GGUF,TeeZee,Nexesenex,

Kyllene-34B-v1.1-b1989-iMat-c32_ch3250-Q5_K_S.gguf,-,Hellaswag,85.25,,400,2024-01-28 01:40:00,,34b,Yi,200000,,,GGUF,TeeZee,Nexesenex,
Kyllene-34B-v1.1-b1989-iMat-c32_ch3250-Q5_K_S.gguf,-,Hellaswag,85.6,,1000,2024-01-28 01:40:00,,34b,Yi,200000,,,GGUF,TeeZee,Nexesenex,
Kyllene-34B-v1.1-b1989-iMat-c32_ch3250-Q5_K_S.gguf,-,Hellaswag,84.95,,2000,2024-01-28 01:40:00,,34b,Yi,200000,,,GGUF,TeeZee,Nexesenex,
Kyllene-34B-v1.1-b1989-iMat-c32_ch3250-Q5_K_S.gguf,-,Hellaswag_Bin,81.25,,400,2024-01-28 01:40:00,,34b,Yi,200000,,,GGUF,TeeZee,Nexesenex,
Kyllene-34B-v1.1-b1989-iMat-c32_ch3250-Q5_K_S.gguf,-,Hellaswag_Bin,83.3,,1000,2024-01-28 01:40:00,,34b,Yi,200000,,,GGUF,TeeZee,Nexesenex,
Kyllene-34B-v1.1-b1989-iMat-c32_ch3250-Q5_K_S.gguf,-,Hellaswag_Bin,83,,2000,2024-01-28 01:40:00,,34b,Yi,200000,,,GGUF,TeeZee,Nexesenex,
Kyllene-34B-v1.1-b1989-iMat-c32_ch3250-Q5_K_S.gguf,-,Arc-Challenge,60.20066890,,299,2024-01-28 05:40:00,,34b,Yi,200000,,,GGUF,TeeZee,Nexesenex,
Kyllene-34B-v1.1-b1989-iMat-c32_ch3250-Q5_K_S.gguf,-,Arc-Easy,81.05263158,,570,2024-01-28 05:40:00,,34b,Yi,200000,,,GGUF,TeeZee,Nexesenex,
Kyllene-34B-v1.1-b1989-iMat-c32_ch3250-Q5_K_S.gguf,-,MMLU,42.17252396,,313,2024-01-28 05:40:00,,34b,Yi,200000,,,GGUF,TeeZee,Nexesenex,
Kyllene-34B-v1.1-b1989-iMat-c32_ch3250-Q5_K_S.gguf,-,Thruthful-QA,36.96450428,,817,2024-01-28 05:40:00,,34b,Yi,200000,,,GGUF,TeeZee,Nexesenex,
Kyllene-34B-v1.1-b1989-iMat-c32_ch3250-Q5_K_S.gguf,-,Winogrande,79.5580,,1267,2024-01-28 05:40:00,,34b,Yi,200000,,,GGUF,TeeZee,Nexesenex,
Kyllene-34B-v1.1-b1989-iMat-c32_ch3250-Q5_K_S.gguf,-,wikitext,5.1806,512,512,2024-01-28 01:40:00,,34b,Yi,200000,,,GGUF,TeeZee,Nexesenex,

And Teezee's merge takes the lead. Then I tested it on ST, and the benchs are confirmed : it's a delight to chat with, this model really elaborates its RP answers in a compelling way. Thanks to you both, guys! (I'm using your merges for a while, Bruce, and until today they were the only one really working for Yi imho).

As for your 57b model, Teezee, it it a match or even better than this one, especially at long context? Capy 34b is serious, and Bagel 34b is very promising but prone to degenerate at long context.

TeeZee

Owner Jan 29, 2024

Thank you Nexesenex for benchmarking, I'm glad you enjoy the model, more to come soon :)

Nexesenex

Jan 29, 2024

•

edited Jan 29, 2024

You're welcome, Teezee. I'm very enthusiastic about Kyllene 34b, its output has really a natural feel, and Gryphe's mergemonster seems to be an awesome tool to help to achieve that.

By the way, I suggest you to get a look at Smaug, which is a solid bench performer and might not be a cheat like the UNA-Beagles34 models seem to be :
https://huggingface.co/abacusai/Smaug-34B-v0.1/discussions/2

I didn't chat with Smaug much (and already forgot my impressions because I focused on Kyllene 34b while my GPU is busy atm, but it could be a good component for mergemonster.

In any case, I can't wait to test.. and bench your next models!

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment