Instructions to use prithivMLmods/Neumind-Math-7B-Instruct with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use prithivMLmods/Neumind-Math-7B-Instruct with Transformers:

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

pipe = pipeline("text-generation", model="prithivMLmods/Neumind-Math-7B-Instruct")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("prithivMLmods/Neumind-Math-7B-Instruct")
model = AutoModelForCausalLM.from_pretrained("prithivMLmods/Neumind-Math-7B-Instruct")
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 prithivMLmods/Neumind-Math-7B-Instruct with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "prithivMLmods/Neumind-Math-7B-Instruct"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "prithivMLmods/Neumind-Math-7B-Instruct",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/prithivMLmods/Neumind-Math-7B-Instruct

SGLang

How to use prithivMLmods/Neumind-Math-7B-Instruct 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 "prithivMLmods/Neumind-Math-7B-Instruct" \
    --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": "prithivMLmods/Neumind-Math-7B-Instruct",
		"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 "prithivMLmods/Neumind-Math-7B-Instruct" \
        --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": "prithivMLmods/Neumind-Math-7B-Instruct",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use prithivMLmods/Neumind-Math-7B-Instruct with Docker Model Runner:
```
docker model run hf.co/prithivMLmods/Neumind-Math-7B-Instruct
```

prithivMLmods commited on Dec 5, 2024

Commit

ac9c818

verified ·

1 Parent(s): 52da960

Update README.md

Browse files

Files changed (1) hide show

README.md +56 -0

README.md CHANGED Viewed

@@ -21,6 +21,8 @@ tags:
 ### Neumind-Math-7B-Instruct Model Files
 | File Name                          | Size       | Description                              | Upload Status  |
 |------------------------------------|------------|------------------------------------------|----------------|
 | `.gitattributes`                   | 1.57 kB    | Git attributes configuration file        | Uploaded       |
@@ -40,3 +42,57 @@ tags:
 | `vocab.json`                       | 2.78 MB    | Vocabulary for tokenization              | Uploaded       |
 ---

 ### Neumind-Math-7B-Instruct Model Files
+The **Neumind-Math-7B-Instruct** is a fine-tuned model based on **Qwen2.5-7B-Instruct**, optimized for mathematical reasoning, step-by-step problem-solving, and instruction-based tasks in the mathematics domain. The model is designed for applications requiring structured reasoning, numerical computations, and mathematical proof generation.
 | File Name                          | Size       | Description                              | Upload Status  |
 |------------------------------------|------------|------------------------------------------|----------------|
 | `.gitattributes`                   | 1.57 kB    | Git attributes configuration file        | Uploaded       |
 | `vocab.json`                       | 2.78 MB    | Vocabulary for tokenization              | Uploaded       |
 ---
+### **Key Features:**
+1. **Mathematical Reasoning:**
+   Specifically fine-tuned for solving mathematical problems, including arithmetic, algebra, calculus, and geometry.
+2. **Step-by-Step Problem Solving:**
+   Provides detailed, logical solutions for complex mathematical tasks and demonstrates problem-solving methodologies.
+3. **Instructional Applications:**
+   Tailored for use in educational settings, such as tutoring systems, math content creation, and interactive learning tools.
+---
+### **Training Details:**
+- **Base Model:** [Qwen2.5-7B-Instruct](#)
+- **Dataset:** Trained on **AI-MO/NuminaMath-CoT**, a large dataset of mathematical problems and chain-of-thought (CoT) reasoning. The dataset contains **860k problems** across various difficulty levels, enabling the model to tackle a wide spectrum of mathematical tasks.
+---
+### **Capabilities:**
+- **Complex Problem Solving:**
+   Solves a wide range of mathematical problems, from basic arithmetic to advanced calculus and algebraic equations.
+- **Chain-of-Thought Reasoning:**
+   Excels in step-by-step logical reasoning, making it suitable for tasks requiring detailed explanations.
+- **Instruction-Based Generation:**
+   Ideal for generating educational content, such as worked examples, quizzes, and tutorials.
+---
+### **Usage Instructions:**
+1. **Model Setup:**
+   Download all model shards and the associated configuration files. Ensure the files are correctly placed for seamless loading.
+2. **Inference:**
+   Load the model using frameworks like PyTorch and Hugging Face Transformers. Ensure the `pytorch_model.bin.index.json` file is in the same directory for shard-based loading.
+3. **Customization:**
+   Adjust generation parameters using `generation_config.json` to optimize outputs for your specific application.
+---
+### **Applications:**
+- **Education:**
+   Interactive math tutoring, content creation, and step-by-step problem-solving tools.
+- **Research:**
+   Automated theorem proving and symbolic mathematics.
+- **General Use:**
+   Solving everyday mathematical queries and generating numerical datasets.
+---