Instructions to use clibrain/mamba-2.8b-instruct-openhermes with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use clibrain/mamba-2.8b-instruct-openhermes with Transformers:

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

pipe = pipeline("text-generation", model="clibrain/mamba-2.8b-instruct-openhermes")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("clibrain/mamba-2.8b-instruct-openhermes", dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use clibrain/mamba-2.8b-instruct-openhermes with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "clibrain/mamba-2.8b-instruct-openhermes"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "clibrain/mamba-2.8b-instruct-openhermes",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/clibrain/mamba-2.8b-instruct-openhermes

SGLang

How to use clibrain/mamba-2.8b-instruct-openhermes 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 "clibrain/mamba-2.8b-instruct-openhermes" \
    --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": "clibrain/mamba-2.8b-instruct-openhermes",
		"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 "clibrain/mamba-2.8b-instruct-openhermes" \
        --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": "clibrain/mamba-2.8b-instruct-openhermes",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use clibrain/mamba-2.8b-instruct-openhermes with Docker Model Runner:
```
docker model run hf.co/clibrain/mamba-2.8b-instruct-openhermes
```

Training With EOS

by assafbk - opened Jun 8, 2024

Discussion

assafbk

Jun 8, 2024

Thanks for all of your great work!
During training, have you encountered a situation where the model predicts the EOS token at the beginning of the response?
And if so, were you able to mitigate this behaviour?
Thanks in advance!

mrm8488

CliBrAIn org Jun 9, 2024

Hi, @assafbk
AFAIK we didn't have that problem.
Thanks!

mrm8488 changed discussion status to closed Jun 9, 2024

Ajibola

Jun 9, 2024

The model hasn't converged. Continue training

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