Instructions to use rinna/qwen2.5-bakeneko-32b-instruct-v2-awq with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use rinna/qwen2.5-bakeneko-32b-instruct-v2-awq with Transformers:

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

pipe = pipeline("text-generation", model="rinna/qwen2.5-bakeneko-32b-instruct-v2-awq")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("rinna/qwen2.5-bakeneko-32b-instruct-v2-awq")
model = AutoModelForCausalLM.from_pretrained("rinna/qwen2.5-bakeneko-32b-instruct-v2-awq", 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 rinna/qwen2.5-bakeneko-32b-instruct-v2-awq with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "rinna/qwen2.5-bakeneko-32b-instruct-v2-awq"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "rinna/qwen2.5-bakeneko-32b-instruct-v2-awq",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/rinna/qwen2.5-bakeneko-32b-instruct-v2-awq

SGLang

How to use rinna/qwen2.5-bakeneko-32b-instruct-v2-awq 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 "rinna/qwen2.5-bakeneko-32b-instruct-v2-awq" \
    --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": "rinna/qwen2.5-bakeneko-32b-instruct-v2-awq",
		"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 "rinna/qwen2.5-bakeneko-32b-instruct-v2-awq" \
        --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": "rinna/qwen2.5-bakeneko-32b-instruct-v2-awq",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use rinna/qwen2.5-bakeneko-32b-instruct-v2-awq with Docker Model Runner:
```
docker model run hf.co/rinna/qwen2.5-bakeneko-32b-instruct-v2-awq
```

keisawada commited on Mar 23, 2025

Commit

49eb39c

verified ·

1 Parent(s): 1917901

Update README.md

Browse files

Files changed (1) hide show

README.md +13 -1

README.md CHANGED Viewed

@@ -21,6 +21,14 @@ library_name: transformers
 This model is a 4-bit quantized model for [rinna/qwen2.5-bakeneko-32b-instruct-v2](https://huggingface.co/rinna/qwen2.5-bakeneko-32b-instruct-v2) using [AutoAWQ](https://github.com/casper-hansen/AutoAWQ). The quantized version is 4x smaller than the original model and thus requires less memory and provides faster inference.
 See [rinna/qwen2.5-bakeneko-32b-instruct-v2](https://huggingface.co/rinna/qwen2.5-bakeneko-32b-instruct-v2) for details about model architecture and data.
 * **Contributors**
@@ -28,6 +36,10 @@ See [rinna/qwen2.5-bakeneko-32b-instruct-v2](https://huggingface.co/rinna/qwen2.
     - [Xinqi Chen](https://huggingface.co/Keely0419)
     - [Kei Sawada](https://huggingface.co/keisawada)
 ---
 # Benchmarking
@@ -44,7 +56,7 @@ See [rinna/qwen2.5-bakeneko-32b-instruct-v2](https://huggingface.co/rinna/qwen2.
 | [Qwen/QwQ-32B](https://huggingface.co/Qwen/QwQ-32B) | 76.12 | 8.58 | 8.25
 | [rinna/qwq-bakeneko-32b](https://huggingface.co/rinna/qwq-bakeneko-32b) | 78.31 | 8.81 | 8.52
-For detailed benchmarking results, please refer to [rinna's LM benchmark page](https://rinnakk.github.io/research/benchmarks/lm/index.html).
 ---

 This model is a 4-bit quantized model for [rinna/qwen2.5-bakeneko-32b-instruct-v2](https://huggingface.co/rinna/qwen2.5-bakeneko-32b-instruct-v2) using [AutoAWQ](https://github.com/casper-hansen/AutoAWQ). The quantized version is 4x smaller than the original model and thus requires less memory and provides faster inference.
+| Model Type | Model Name
+| :-   | :-
+| Japanese Continual Pre-Training Model | Qwen2.5 Bakeneko 32B [[HF]](https://huggingface.co/rinna/qwen2.5-bakeneko-32b)
+| Instruction-Tuning Model | Qwen2.5 Bakeneko 32B Instruct [[HF]](https://huggingface.co/rinna/qwen2.5-bakeneko-32b-instruct)[[AWQ]](https://huggingface.co/rinna/qwen2.5-bakeneko-32b-instruct-awq)[[GGUF]](https://huggingface.co/rinna/qwen2.5-bakeneko-32b-instruct-gguf)[[GPTQ int8]](https://huggingface.co/rinna/qwen2.5-bakeneko-32b-instruct-gptq-int8)[[GPTQ int4]](https://huggingface.co/rinna/qwen2.5-bakeneko-32b-instruct-gptq-int4)
+| DeepSeek R1 Distill Qwen2.5 Merged Reasoning Model | DeepSeek R1 Distill Qwen2.5 Bakeneko 32B [[HF]](https://huggingface.co/rinna/deepseek-r1-distill-qwen2.5-bakeneko-32b)[[AWQ]](https://huggingface.co/rinna/deepseek-r1-distill-qwen2.5-bakeneko-32b-awq)[[GGUF]](https://huggingface.co/rinna/deepseek-r1-distill-qwen2.5-bakeneko-32b-gguf)[[GPTQ int8]](https://huggingface.co/rinna/deepseek-r1-distill-qwen2.5-bakeneko-32b-gptq-int8)[[GPTQ int4]](https://huggingface.co/rinna/deepseek-r1-distill-qwen2.5-bakeneko-32b-gptq-int4)
+| QwQ Merged Reasoning Model | QwQ Bakeneko 32B [[HF]](https://huggingface.co/rinna/qwq-bakeneko-32b)[[AWQ]](https://huggingface.co/rinna/qwq-bakeneko-32b-awq)[[GGUF]](https://huggingface.co/rinna/qwq-bakeneko-32b-gguf)[[GPTQ int8]](https://huggingface.co/rinna/qwq-bakeneko-32b-gptq-int8)[[GPTQ int4]](https://huggingface.co/rinna/qwq-bakeneko-32b-gptq-int4)
+| QwQ Bakeneko Merged Instruction-Tuning Model | Qwen2.5 Bakeneko 32B Instruct V2 [[HF]](https://huggingface.co/rinna/qwen2.5-bakeneko-32b-instruct-v2)[[AWQ]](https://huggingface.co/rinna/qwen2.5-bakeneko-32b-instruct-v2-awq)[[GGUF]](https://huggingface.co/rinna/qwen2.5-bakeneko-32b-instruct-v2-gguf)[[GPTQ int8]](https://huggingface.co/rinna/qwen2.5-bakeneko-32b-instruct-v2-gptq-int8)[[GPTQ int4]](https://huggingface.co/rinna/qwen2.5-bakeneko-32b-instruct-v2-gptq-int4)
 See [rinna/qwen2.5-bakeneko-32b-instruct-v2](https://huggingface.co/rinna/qwen2.5-bakeneko-32b-instruct-v2) for details about model architecture and data.
 * **Contributors**
     - [Xinqi Chen](https://huggingface.co/Keely0419)
     - [Kei Sawada](https://huggingface.co/keisawada)
+* **Release date**
+    February 19, 2025
 ---
 # Benchmarking
 | [Qwen/QwQ-32B](https://huggingface.co/Qwen/QwQ-32B) | 76.12 | 8.58 | 8.25
 | [rinna/qwq-bakeneko-32b](https://huggingface.co/rinna/qwq-bakeneko-32b) | 78.31 | 8.81 | 8.52
+For detailed benchmarking results, please refer to [rinna's LM benchmark page (Sheet 20250319)](https://rinnakk.github.io/research/benchmarks/lm/index.html).
 ---