Text Classification
Transformers
Safetensors
English
Chinese
qwen2
feature-extraction
reward model
custom_code
text-embeddings-inference
Instructions to use Qwen/Qwen2.5-Math-RM-72B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use Qwen/Qwen2.5-Math-RM-72B with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-classification", model="Qwen/Qwen2.5-Math-RM-72B", trust_remote_code=True)# Load model directly from transformers import AutoTokenizer, AutoModel tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen2.5-Math-RM-72B", trust_remote_code=True) model = AutoModel.from_pretrained("Qwen/Qwen2.5-Math-RM-72B", trust_remote_code=True, device_map="auto") - Notebooks
- Google Colab
- Kaggle
example to service the RM
#2
by weqweasdas - opened
Great work!
I assume that this RM is a process RM? I am curious whether you could provide an example to service the RM to evaluate the intermediate step of the CoT. Many thanks!
As mentioned in the technical report (https://arxiv.org/pdf/2409.12122), we label the responses with the correct answers as positive, while those with incorrect answers as negative, so it is an outcome RM. If you want to try scoring the intermediate step, you might consider directly inputting a step response.
Zhenru changed discussion status to closed