Instructions to use obiwit/qwen2.5-3b-dpo-vanilla-no-tools with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use obiwit/qwen2.5-3b-dpo-vanilla-no-tools with Transformers:

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

pipe = pipeline("text-generation", model="obiwit/qwen2.5-3b-dpo-vanilla-no-tools", device_map="auto")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("obiwit/qwen2.5-3b-dpo-vanilla-no-tools")
model = AutoModelForCausalLM.from_pretrained("obiwit/qwen2.5-3b-dpo-vanilla-no-tools", 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=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use obiwit/qwen2.5-3b-dpo-vanilla-no-tools with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "obiwit/qwen2.5-3b-dpo-vanilla-no-tools"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "obiwit/qwen2.5-3b-dpo-vanilla-no-tools",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/obiwit/qwen2.5-3b-dpo-vanilla-no-tools

SGLang

How to use obiwit/qwen2.5-3b-dpo-vanilla-no-tools 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 "obiwit/qwen2.5-3b-dpo-vanilla-no-tools" \
    --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": "obiwit/qwen2.5-3b-dpo-vanilla-no-tools",
		"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 "obiwit/qwen2.5-3b-dpo-vanilla-no-tools" \
        --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": "obiwit/qwen2.5-3b-dpo-vanilla-no-tools",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use obiwit/qwen2.5-3b-dpo-vanilla-no-tools with Docker Model Runner:
```
docker model run hf.co/obiwit/qwen2.5-3b-dpo-vanilla-no-tools
```

obiwit commited on Jun 16, 2025

Commit

734a5b0

verified ·

1 Parent(s): a898a72

Model save

Browse files

Files changed (6) hide show

README.md +1 -1
all_results.json +3 -3
config.json +1 -1
train_results.json +3 -3
trainer_state.json +3 -3
training_args.bin +1 -1

README.md CHANGED Viewed

@@ -27,7 +27,7 @@ print(output["generated_text"])
 ## Training procedure
-[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/bborges/CriteriaPreferences/runs/4sx9g2ym)
 This model was trained with DPO, a method introduced in [Direct Preference Optimization: Your Language Model is Secretly a Reward Model](https://huggingface.co/papers/2305.18290).

 ## Training procedure
+[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/bborges/CriteriaPreferences/runs/meb4alvj)
 This model was trained with DPO, a method introduced in [Direct Preference Optimization: Your Language Model is Secretly a Reward Model](https://huggingface.co/papers/2305.18290).

all_results.json CHANGED Viewed

@@ -15,8 +15,8 @@
     "eval_steps_per_second": 0.492,
     "total_flos": 0.0,
     "train_loss": 0.0,
-    "train_runtime": 18.8042,
     "train_samples": 3438778,
-    "train_samples_per_second": 182872.522,
-    "train_steps_per_second": 1428.667
 }

     "eval_steps_per_second": 0.492,
     "total_flos": 0.0,
     "train_loss": 0.0,
+    "train_runtime": 20.8147,
     "train_samples": 3438778,
+    "train_samples_per_second": 165208.979,
+    "train_steps_per_second": 1290.673
 }

config.json CHANGED Viewed

@@ -24,7 +24,7 @@
   "tie_word_embeddings": true,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.46.3",
-  "use_cache": true,
   "use_mrope": false,
   "use_sliding_window": false,
   "vocab_size": 151665

   "tie_word_embeddings": true,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.46.3",
+  "use_cache": false,
   "use_mrope": false,
   "use_sliding_window": false,
   "vocab_size": 151665

train_results.json CHANGED Viewed

@@ -2,8 +2,8 @@
     "epoch": 0.9999813887699838,
     "total_flos": 0.0,
     "train_loss": 0.0,
-    "train_runtime": 18.8042,
     "train_samples": 3438778,
-    "train_samples_per_second": 182872.522,
-    "train_steps_per_second": 1428.667
 }

     "epoch": 0.9999813887699838,
     "total_flos": 0.0,
     "train_loss": 0.0,
+    "train_runtime": 20.8147,
     "train_samples": 3438778,
+    "train_samples_per_second": 165208.979,
+    "train_steps_per_second": 1290.673
 }

trainer_state.json CHANGED Viewed

@@ -40398,9 +40398,9 @@
       "step": 26865,
       "total_flos": 0.0,
       "train_loss": 0.0,
-      "train_runtime": 18.8042,
-      "train_samples_per_second": 182872.522,
-      "train_steps_per_second": 1428.667
     }
   ],
   "logging_steps": 10,

       "step": 26865,
       "total_flos": 0.0,
       "train_loss": 0.0,
+      "train_runtime": 20.8147,
+      "train_samples_per_second": 165208.979,
+      "train_steps_per_second": 1290.673
     }
   ],
   "logging_steps": 10,

training_args.bin CHANGED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6eb07e19092483aa01195a9866c272305219978e8d7683d95f384d14ef17c049
 size 7864

 version https://git-lfs.github.com/spec/v1
+oid sha256:dfa920440d870bd978c9bfc20c34f8c14db806f66ae160525b7d6434ab31f248
 size 7864