Instructions to use microsoft/DialoGPT-large with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use microsoft/DialoGPT-large with Transformers:

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

pipe = pipeline("text-generation", model="microsoft/DialoGPT-large")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("microsoft/DialoGPT-large")
model = AutoModelForCausalLM.from_pretrained("microsoft/DialoGPT-large")
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 microsoft/DialoGPT-large with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/microsoft/DialoGPT-large

SGLang

How to use microsoft/DialoGPT-large 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 "microsoft/DialoGPT-large" \
    --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": "microsoft/DialoGPT-large",
		"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 "microsoft/DialoGPT-large" \
        --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": "microsoft/DialoGPT-large",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use microsoft/DialoGPT-large with Docker Model Runner:
```
docker model run hf.co/microsoft/DialoGPT-large
```

Wauplin HF Staff commited on Feb 28, 2024

Commit

58feb13

verified ·

1 Parent(s): 3391fef

Add `eos_token` to the tokenizer config.

Browse files

If we merge the ChatWidget right now, it would not work since the widget would not be able to format the chat_template correctly from the information available via API (https://huggingface.co/api/models/microsoft/DialoGPT-large?config=True). This is because to get `eos_token` one need to get `eos_token_id` from [config.json](https://huggingface.co/microsoft/DialoGPT-large/blob/main/config.json) and then reading [`vocab.json`](https://huggingface.co/microsoft/DialoGPT-large/blob/main/vocab.json) to check which token is associated with this id.

This PR fixes this by adding `eos_token` directly to `tokenizer_config.json` cc @julien-c @osanseviero @sbrandeis

Files changed (1) hide show

tokenizer_config.json +2 -1

tokenizer_config.json CHANGED Viewed

@@ -1,4 +1,5 @@
 {
   "model_max_length": 1024,
-  "chat_template": "{% for message in messages %}{{ message.content }}{{ eos_token }}{% endfor %}"
 }

 {
   "model_max_length": 1024,
+  "chat_template": "{% for message in messages %}{{ message.content }}{{ eos_token }}{% endfor %}",
+  "eos_token": "<|endoftext|>"
 }