Instructions to use experimentalmachines/QwenGrad-DPO with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use experimentalmachines/QwenGrad-DPO with Transformers:

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

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

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("experimentalmachines/QwenGrad-DPO")
model = AutoModelForCausalLM.from_pretrained("experimentalmachines/QwenGrad-DPO", device_map="auto")
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=256)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use experimentalmachines/QwenGrad-DPO with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/experimentalmachines/QwenGrad-DPO

SGLang

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

Docker Model Runner
How to use experimentalmachines/QwenGrad-DPO with Docker Model Runner:
```
docker model run hf.co/experimentalmachines/QwenGrad-DPO
```

Arjhine Ty commited on Sep 12

Commit

d462ebe

verified ·

1 Parent(s): 69b1760

Replace confirmatory results table with matplotlib chart

Browse files

Files changed (2) hide show

README.md +4 -4
qwengrad-dpo-results.png +0 -0

README.md CHANGED Viewed

@@ -36,10 +36,10 @@ This repository is published by **Experimental Machines**. It contains the promo
 ## Results (pre-registered internal confirmatory partition, 1,277 examples)
-| run | `call_f1` | precision | recall | over_call | clarification | unsupported |
-|---|---:|---:|---:|---:|---:|---:|
-| M0-final-v2 @1800 | 0.7470 | 0.7350 | 0.7594 | 0.1505 | 0.7682 | 0.5430 |
-| **M1-v2 @30** | **0.7548** | 0.7358 | **0.7748** | 0.1529 | 0.7655 | 0.5386 |
 M1 preserves the M0 calibrated frontier and makes a small improvement in call F1 and recall. It does
 not materially reduce over-calling; this is calibration retention and slight improvement, not a large

 ## Results (pre-registered internal confirmatory partition, 1,277 examples)
+![QwenGrad-DPO confirmatory metrics: M0-final-v2 @1800 versus M1-v2 @30](qwengrad-dpo-results.png)
+The chart reports the exact values for `call_f1`, precision, recall, `over_call`, clarification,
+and unsupported. Higher is better for every metric except `over_call`.
 M1 preserves the M0 calibrated frontier and makes a small improvement in call F1 and recall. It does
 not materially reduce over-calling; this is calibration retention and slight improvement, not a large

qwengrad-dpo-results.png ADDED Viewed