Instructions to use replit/replit-code-v1-3b with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use replit/replit-code-v1-3b with Transformers:

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

pipe = pipeline("text-generation", model="replit/replit-code-v1-3b", trust_remote_code=True)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("replit/replit-code-v1-3b", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("replit/replit-code-v1-3b", trust_remote_code=True)

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use replit/replit-code-v1-3b with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "replit/replit-code-v1-3b"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "replit/replit-code-v1-3b",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/replit/replit-code-v1-3b

SGLang

How to use replit/replit-code-v1-3b 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 "replit/replit-code-v1-3b" \
    --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": "replit/replit-code-v1-3b",
		"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 "replit/replit-code-v1-3b" \
        --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": "replit/replit-code-v1-3b",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use replit/replit-code-v1-3b with Docker Model Runner:
```
docker model run hf.co/replit/replit-code-v1-3b
```

Expected minimum hardware requirements for inference?

by zeroing - opened May 3, 2023

Discussion

zeroing

May 3, 2023

Title is self-explanatory. 😁

pirroh

Replit org May 3, 2023

Not easy to give you a reliable answer, given that "minimum" requirement could be also CPU inference... if you are willing to wait minutes for a few tokens 😄
That said, we are hosting our demo on an NVIDIA A10G, and it looks pretty fast!
Further quantization with LLM.int8() would help you load the model even on smaller GPUs.

merlinarer

May 4, 2023

Not easy to give you a reliable answer, given that "minimum" requirement could be also CPU inference... if you are willing to wait minutes for a few tokens 😄
That said, we are hosting our demo on an NVIDIA A10G, and it looks pretty fast!
Further quantization with LLM.int8() would help you load the model even on smaller GPUs.

hello, is the demo hosted naively with transfomers as described in model card ? I thought it is hosted with optimized method such as fastertransfomer.

madhavatreplit

May 5, 2023

It's hosted natively on the GPU as described in the model card with bfloat16 precision, but without any flash attention — i.e. the attn_impl kwarg defaults to 'torch'.

luoji12345

May 17, 2023

It's hosted natively on the GPU as described in the model card with bfloat16 precision, but without any flash attention — i.e. the attn_impl kwarg defaults to 'torch'.

When I set attn_impl to flash , the flash_attn can not be matched with alibi , so wpe layer must be newly initialized.
My question is :

Whether you will release the flash_attn version model ? That model have the wpe layer that match the flash_attn config.
How faster is the flash_attn than torch ? Will it be faster in the inference process ?
Weather you will release faster_transformer version of code ? It will benefit many people.

vgrichina

May 17, 2023

ggml just merged CPU inference implementation:
https://github.com/ggerganov/ggml/tree/master/examples/replit

it's pretty fast on my M1 16GB MB Air

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