Instructions to use instruction-pretrain/instruction-synthesizer with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use instruction-pretrain/instruction-synthesizer with Transformers:

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

pipe = pipeline("text-generation", model="instruction-pretrain/instruction-synthesizer")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("instruction-pretrain/instruction-synthesizer")
model = AutoModelForCausalLM.from_pretrained("instruction-pretrain/instruction-synthesizer")

Inference
Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use instruction-pretrain/instruction-synthesizer with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/instruction-pretrain/instruction-synthesizer

SGLang

How to use instruction-pretrain/instruction-synthesizer 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 "instruction-pretrain/instruction-synthesizer" \
    --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": "instruction-pretrain/instruction-synthesizer",
		"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 "instruction-pretrain/instruction-synthesizer" \
        --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": "instruction-pretrain/instruction-synthesizer",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use instruction-pretrain/instruction-synthesizer with Docker Model Runner:
```
docker model run hf.co/instruction-pretrain/instruction-synthesizer
```

Can you share the training code for the instruction synthesizer？

by chenff35 - opened Jun 28, 2024

Discussion

chenff35

Jun 28, 2024

Can you share the training code for the instruction synthesizer？

instruction-pretrain

Owner Jun 29, 2024

Thanks for your interest. We plan to share the training code soon. In the meantime, could you specify the problems you encountered when training the synthesizer yourself? We may be able to help you quickly before the code release.

shailj

Jul 3, 2024

Once the dataset in generated can you share the code for pretraining

instruction-pretrain

Owner Jul 4, 2024

Okay, we will try to release the code as soon as possible.

merlinarer

Aug 20, 2024

hello, any progress?

instruction-pretrain

Owner Aug 27, 2024

Sorry for the late response. We provide guidance on data formulation and loss calculation in ft-instruction-synthesizer-collection. Beyond that, our fine-tuning process is the same as regular fine-tuning of language models. Therefore, we recommend using popular open-source code like Llama-factory for fine-tuning, which might be even easier to use than our own code.

If you still encounter implementation difficulties, please don't hesitate to contact us.

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