metadata

base_model: deepseek-ai/deepseek-math-7b-base
tags:
  - alignment-handbook
  - generated_from_trainer
widget:
  - example_title: Math problem
    messages:
      - role: user
        content: >-
          For how many values of the constant $k$ will the polynomial
          $x^{2}+kx+36$ have two distinct integer roots?
    output:
      text: >-
        ### Solution: 1. For the polynomial \\( x^2 + kx + 36 \\) to have two
        distinct integer roots, let's denote these roots by \\( r_1 \\) and \\(
        r_2 \\).\n\n2. According to Vieta's formulas, the sum of the roots \\(
        r_1 + r_2 \\) is equal to \\(-k\\), and the product of the roots \\( r_1
        \\cdot r_2 \\) is equal to 36.\n\n3. To find the distinct integer pairs
        \\((r_1, r_2)\\) whose product is 36, we need to determine all pairs of
        integers \\((r_1, r_2)\\) such that \\( r_1 \\cdot r_2 = 36 \\) and \\(
        r_1 \\neq r_2 \\).\n\n4. From the equation \\( r_1 + r_2 = -k \\), for
        each valid pair \\((r_1, r_2)\\), we can compute the corresponding value
        of \\( k \\).\n\n5. Finally, since we need the polynomial to have two
        distinct integer roots, we need to ensure that \\( r_1 \\) and \\( r_2
        \\) are distinct.\n\nLet's start by finding all pairs \\((r_1, r_2)\\)
        such that \\( r_1 \\cdot r_2 = 36 \\). We'll then determine the values
        of \\( k \\) and ensure the roots are distinct.\n```python\nimport
        itertools\n\n# Find all pairs (r1, r2) such that r1 * r2 =
        36\nproduct_36 = 36\nfactor_pairs = []\n\nfor i in range(1, product_36 +
        1):\n if product_36 % i == 0:\n pair = (i, product_36 // i)\n if pair[0]
        != pair[1]: # Ensure distinct pairs\n factor_pairs.append(pair)\n\n#
        Calculate k for each pair and ensure distinct integer
        roots\nvalid_k_values = set()\nfor r1, r2 in factor_pairs:\n if r1 !=
        r2:\n k = -(r1 + r2)\n
        valid_k_values.add(k)\n\nprint((len(valid_k_values),
        sorted(valid_k_values)))\n```\n```output\n(4, [-37, -20, -15,
        -13])\n```\nThe distinct integer values of \\( k \\) that make the
        polynomial \\( x^2 + kx + 36 \\) have two distinct integer roots are
        \\(-37, -20, -15, \\text{and} -13\\).\n\nTherefore, the number of such
        values of \\( k \\) is:\n\n\\[\n\\boxed{4}\n\\]\n
pipeline_tag: text-generation
model-index:
  - name: Numina-Math-7B
    results: []

$Numina Logo$

Model Card for NuminaMath 7B

NuminaMath is a series of language models that are trained to solve math problems using tool-integrated reasoning. NuminaMath 7B won the first progress prize of the AI Math Olympiad (AIMO), with a score of 29/50 on the public and private tests sets.

This model is a fine-tuned version of deepseek-ai/deepseek-math-7b-base with two stages of supervised fine-tuning:

Stage 1: fine-tune the base model on a large, diverse dataset of natural language math problems and solutions, where each solution is templated with Chain of Thought (CoT) to facilitate learning.
Stage 2: fine-tune the model from Stage 1 on a synthetic dataset of tool-integrated reasoning, where each math problem is decomposed into a sequence of rationales, Python programs, and their outputs. Here we followed Microsoft’s ToRA paper and prompted GPT-4 to produce solutions in the ToRA format with code execution feedback. Fine-tuning on this data produces a reasoning agent that can solve mathematical problems via a mix of natural language reasoning and use of the Python REPL to compute intermediate results.

Model description

Model type: A 7B parameter math LLM fine-tuned in two stages of supervised fine-tuning, first on a dataset with math problem-solution pairs and then on a synthetic dataset with examples of multi-step generations using tool-integrated reasoning.
Language(s) (NLP): Primarily English
License: Apache 2.0
Finetuned from model: deepseek-ai/deepseek-math-7b-base

Model Sources

Repository: Coming soon to https://github.com/huggingface/alignment-handbook

Intended uses & limitations

Here's how you can run the model using the pipeline() function from 🤗 Transformers:

import re
import torch
from transformers import pipeline

pipe = pipeline("text-generation", model="AI-MO/Numina-Math-7B", torch_dtype=torch.bfloat16, device_map="auto")

messages = [
    {"role": "user", "content": "For how many values of the constant $k$ will the polynomial $x^{2}+kx+36$ have two distinct integer roots?"},
]
prompt = pipe.tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)

gen_config = {
    "max_new_tokens": 1024,
    "do_sample": False,
    "stop_strings": ["```output"], # Generate until Python code block is complete
    "tokenizer": pipe.tokenizer,
}

outputs = pipe(prompt, **gen_config)
text = outputs[0]["generated_text"]
print(text)

# WARNING: This code will execute the python code in the string. We show this for eductional purposes only.
# Please refer to our full pipeline for a safer way to execute code.
python_code = re.findall(r"```python(.*?)```", text, re.DOTALL)[0]
exec(python_code)

The above executes a single step of Python code - for more complex problems, you will want to run the logic for several steps to obtain the final solution.

Bias, Risks, and Limitations

NuminaMath 7B was created to solve problems in the narrow domain of competition-level mathematics. As a result, the model should not be used for general chat applications. With greedy decoding, we find the model is capable of solving problems at the level of AMC 12, but often struggles generate a valid solution on harder problems at the AIME and Math Olympiad level.

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 2e-05
train_batch_size: 4
eval_batch_size: 8
seed: 42
distributed_type: multi-GPU
num_devices: 8
total_train_batch_size: 32
total_eval_batch_size: 64
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: cosine
lr_scheduler_warmup_ratio: 0.1
num_epochs: 4.0

Training results

Training Loss	Epoch	Step	Validation Loss
0.4295	1.0	1733	0.4313
0.3638	2.0	3466	0.4332
0.2951	3.0	5199	0.4704
0.2225	4.0	6932	0.5302

Framework versions

Transformers 4.40.1
Pytorch 2.3.1
Datasets 2.18.0
Tokenizers 0.19.1