Instructions to use shpotes/codegen-350M-mono with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use shpotes/codegen-350M-mono with Transformers:

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

pipe = pipeline("text-generation", model="shpotes/codegen-350M-mono")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("shpotes/codegen-350M-mono")
model = AutoModelForCausalLM.from_pretrained("shpotes/codegen-350M-mono")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use shpotes/codegen-350M-mono with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "shpotes/codegen-350M-mono"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "shpotes/codegen-350M-mono",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/shpotes/codegen-350M-mono

SGLang

How to use shpotes/codegen-350M-mono 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 "shpotes/codegen-350M-mono" \
    --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": "shpotes/codegen-350M-mono",
		"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 "shpotes/codegen-350M-mono" \
        --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": "shpotes/codegen-350M-mono",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use shpotes/codegen-350M-mono with Docker Model Runner:
```
docker model run hf.co/shpotes/codegen-350M-mono
```

Santiago Hincapie-Potes commited on Jun 22, 2022

Commit

249d3f7

1 Parent(s): d8bd706

feat: improve onnx support

Browse files

Files changed (1) hide show

modelling_codegen.py +4 -1

modelling_codegen.py CHANGED Viewed

@@ -40,7 +40,10 @@ def fixed_pos_embedding(x, seq_dim=1, seq_len=None):
     if seq_len is None:
         seq_len = x.shape[seq_dim]
     inv_freq = 1.0 / (10000 ** (torch.arange(0, dim, 2) / dim))
-    sinusoid_inp = torch.einsum("i , j -> i j", torch.arange(seq_len), inv_freq).to(x.device).float()
     return torch.sin(sinusoid_inp), torch.cos(sinusoid_inp)

     if seq_len is None:
         seq_len = x.shape[seq_dim]
     inv_freq = 1.0 / (10000 ** (torch.arange(0, dim, 2) / dim))
+    # original
+    # sinusoid_inp = torch.einsum("i , j -> i j", torch.arange(seq_len), inv_freq).to(x.device).float()
+    # QHD fix onnx error by https://github.com/microsoft/onnxruntime/discussions/10121#discussioncomment-1987845
+    sinusoid_inp = torch.einsum("i , j -> i j", torch.arange(seq_len).float(), inv_freq).to(x.device).float()
     return torch.sin(sinusoid_inp), torch.cos(sinusoid_inp)