Instructions to use CausalLM/35b-beta-long with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use CausalLM/35b-beta-long with Transformers:

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

pipe = pipeline("text-generation", model="CausalLM/35b-beta-long")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("CausalLM/35b-beta-long")
model = AutoModelForCausalLM.from_pretrained("CausalLM/35b-beta-long")
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 CausalLM/35b-beta-long with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "CausalLM/35b-beta-long"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "CausalLM/35b-beta-long",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/CausalLM/35b-beta-long

SGLang

How to use CausalLM/35b-beta-long 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 "CausalLM/35b-beta-long" \
    --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": "CausalLM/35b-beta-long",
		"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 "CausalLM/35b-beta-long" \
        --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": "CausalLM/35b-beta-long",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use CausalLM/35b-beta-long with Docker Model Runner:
```
docker model run hf.co/CausalLM/35b-beta-long
```

Anyone have gguf quants?

by lemon07r - opened Apr 29, 2024

Discussion

lemon07r

Apr 29, 2024

A big thanks for this farewell gift. Might be one of the best models in this size we have for a while since finetuning for 32b/35b is slow (this is the only good one from what I can tell). I'm wondering if anyone has gguf quants for this model.

JosephusCheung

CausalLM.org org Apr 29, 2024

I think the implementation of bpe tokenizer from llama.cpp is still incorrect, and it won't work as expected unless someone fix that. And it's the same case with cohere-command-r. The only different thing is that I replaced the special tokens to those from chatml.

I would recommend aphrodite-engine for accelerated inference with f16.

0-hero

May 6, 2024

Just added them. Issue is fixed in llama.cpp
QuantFactory/CausalLM-35b-beta-long-GGUF

tastypear

May 11, 2024

bartowski/35b-beta-long-GGUF is OK with the latest llama.cpp or koboldcpp.

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