58 lines
3.4 KiB
Markdown
58 lines
3.4 KiB
Markdown
---
|
|
license: llama3.2
|
|
language:
|
|
- en
|
|
base_model:
|
|
- meta-llama/Llama-3.2-3B-Instruct
|
|
pipeline_tag: text-generation
|
|
library_name: transformers
|
|
tags:
|
|
- text-generation-inference
|
|
---
|
|
# **Llama-3.2-3B-Math-Oct**
|
|
|
|
Llama-3.2-3B-Math-Oct is a math role-play model designed to solve mathematical problems and enhance the reasoning capabilities of 3B-parameter models. These models have proven highly effective in context understanding, reasoning, and mathematical problem-solving, based on the Llama 3.2 is an auto-regressive language model that uses an optimized transformer architecture. The tuned versions use supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF) to align with human preferences for helpfulness and safety.
|
|
|
|
|
|
# **Use with transformers**
|
|
|
|
Starting with `transformers >= 4.43.0` onward, you can run conversational inference using the Transformers `pipeline` abstraction or by leveraging the Auto classes with the `generate()` function.
|
|
|
|
Make sure to update your transformers installation via `pip install --upgrade transformers`.
|
|
|
|
```python
|
|
import torch
|
|
from transformers import pipeline
|
|
|
|
model_id = "prithivMLmods/Llama-3.2-3B-Math-Oct"
|
|
pipe = pipeline(
|
|
"text-generation",
|
|
model=model_id,
|
|
torch_dtype=torch.bfloat16,
|
|
device_map="auto",
|
|
)
|
|
messages = [
|
|
{"role": "system", "content": "You are a pirate chatbot who always responds in pirate speak!"},
|
|
{"role": "user", "content": "Who are you?"},
|
|
]
|
|
outputs = pipe(
|
|
messages,
|
|
max_new_tokens=256,
|
|
)
|
|
print(outputs[0]["generated_text"][-1])
|
|
```
|
|
# **Intended Use**
|
|
1. **Mathematical Problem Solving**: Llama-3.2-3B-Math-Oct is designed for solving a wide range of mathematical problems, including arithmetic, algebra, calculus, and probability.
|
|
2. **Reasoning Enhancement**: It enriches logical reasoning capabilities, helping users understand and solve complex mathematical concepts.
|
|
3. **Context Understanding**: The model is highly effective in interpreting problem statements, mathematical scenarios, and context-heavy equations.
|
|
4. **Educational Support**: It serves as a learning tool for students, educators, and enthusiasts, providing step-by-step explanations for mathematical solutions.
|
|
5. **Scenario Simulation**: The model can role-play specific mathematical scenarios, such as tutoring, creating math problems, or acting as a math assistant.
|
|
|
|
# **Limitations**
|
|
1. **Accuracy Constraints**: While effective in many cases, the model may occasionally provide incorrect solutions, particularly for highly complex or unconventional problems.
|
|
2. **Parameter Limitation**: Being a 3B-parameter model, it might lack the precision and capacity of larger models for intricate problem-solving.
|
|
3. **Lack of Domain-Specific Expertise**: The model may struggle with problems requiring niche mathematical knowledge or specialized fields like advanced topology or quantum mechanics.
|
|
4. **Dependency on Input Clarity**: Ambiguous or poorly worded problem statements might lead to incorrect interpretations and solutions.
|
|
5. **Inability to Learn Dynamically**: The model cannot improve its understanding or reasoning dynamically without retraining.
|
|
6. **Non-Mathematical Queries**: While optimized for mathematics, the model may underperform in general-purpose tasks compared to models designed for broader use cases.
|
|
7. **Computational Resources**: Deploying the model may require significant computational resources for real-time usage. |