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

Libraries

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

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

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

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("prithivMLmods/Math-IIO-7B-Instruct")
model = AutoModelForCausalLM.from_pretrained("prithivMLmods/Math-IIO-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

vLLM

How to use prithivMLmods/Math-IIO-7B-Instruct with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "prithivMLmods/Math-IIO-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/Math-IIO-7B-Instruct",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

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

SGLang

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

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

lbourdois commited on Apr 28, 2025

Commit

a9e5ad4

verified ·

1 Parent(s): 7e40d00

Improve language tag

Browse files

Hi! As the model is multilingual, this is a PR to add other languages than English to the language tag to improve the referencing. Note that 29 languages are announced in the README, but only 13 are explicitly listed. I was therefore only able to add these 13 languages.

Files changed (1) hide show

README.md +92 -80

README.md CHANGED Viewed

@@ -1,80 +1,92 @@
----
-license: creativeml-openrail-m
-datasets:
-- prithivMLmods/Math-IIO-68K-Mini
-language:
-- en
-base_model:
-- Qwen/Qwen2.5-7B-Instruct
-pipeline_tag: text-generation
-library_name: transformers
-tags:
-- safetensors
-- qwen2.5
-- 7B
-- Instruct
-- Math
-- CoT
-- one-shot
----
-![aaa.png](https://cdn-uploads.huggingface.co/production/uploads/65bb837dbfb878f46c77de4c/faLfR-doaWP_BLUkOQrbq.png)
-### **Math IIO 7B Instruct**
-The **Math IIO 7B Instruct** is a fine-tuned language model based on the robust **Qwen2.5-7B-Instruct** architecture. This model has been specifically trained to excel in single-shot mathematical reasoning and instruction-based tasks, making it a reliable choice for educational, analytical, and problem-solving applications.
-### **Key Features:**
-1. **Math-Optimized Capabilities:**
-   The model is designed to handle complex mathematical problems, step-by-step calculations, and reasoning tasks.
-2. **Instruction-Tuned:**
-   Fine-tuned for better adherence to structured queries and task-oriented prompts, enabling clear and concise outputs.
-3. **Large Vocabulary:**
-   Equipped with an extensive tokenizer configuration and custom tokens to ensure precise mathematical notation support.
-### Single Shot Answers
-![solution.png](https://cdn-uploads.huggingface.co/production/uploads/65bb837dbfb878f46c77de4c/4Zq6crBrbFLDqfKlDwBMU.png)
-### Math-IIO File Structure
-| File Name [ Uploaded file ]                         | Size       | Description                                   | Upload Status  |
-|------------------------------------|------------|-----------------------------------------------|----------------|
-| `.gitattributes`                   | 1.57 kB    | Git attributes configuration file             | Uploaded       |
-| `README.md`                        | 263 Bytes  | README file with minimal details              | Updated        |
-| `added_tokens.json`                | 657 Bytes  | Custom added tokens for tokenizer             | Uploaded       |
-| `config.json`                      | 861 Bytes  | Model configuration file                      | Uploaded       |
-| `generation_config.json`           | 281 Bytes  | Configuration for text generation settings    | Uploaded       |
-| `merges.txt`                       | 1.82 MB    | Merge rules for byte pair encoding tokenizer  | Uploaded       |
-| `pytorch_model-00001-of-00004.bin` | 4.88 GB    | First part of model weights (PyTorch)         | Uploaded (LFS) |
-| `pytorch_model-00002-of-00004.bin` | 4.93 GB    | Second part of model weights (PyTorch)        | Uploaded (LFS) |
-| `pytorch_model-00003-of-00004.bin` | 4.33 GB    | Third part of model weights (PyTorch)         | Uploaded (LFS) |
-| `pytorch_model-00004-of-00004.bin` | 1.09 GB    | Fourth part of model weights (PyTorch)        | Uploaded (LFS) |
-| `pytorch_model.bin.index.json`     | 28.1 kB    | Index JSON file for model weights             | Uploaded       |
-| `special_tokens_map.json`          | 644 Bytes  | Map of special tokens used by the tokenizer   | Uploaded       |
-| `tokenizer.json`                   | 11.4 MB    | Tokenizer settings and vocab                  | Uploaded (LFS) |
-| `tokenizer_config.json`            | 7.73 kB    | Configuration for tokenizer                   | Uploaded       |
-| `vocab.json`                       | 2.78 MB    | Vocabulary for tokenizer                      | Uploaded       |
-| Model Type | Size | Context Length | Link |
-|------------|------|----------------|------|
-| GGUF | 7B | - | [🤗 Math-IIO-7B-Instruct-GGUF](https://huggingface.co/prithivMLmods/Math-IIO-7B-Instruct-GGUF) |
-### **Training Details:**
-- **Base Model:** [Qwen/Qwen2.5-7B-Instruct](#)
-- **Dataset:** Trained on **Math-IIO-68K-Mini**, a curated dataset with 68.8k high-quality examples focusing on mathematical instructions, equations, and logic-based queries.
-### **Capabilities:**
-- **Problem-Solving:** Solves mathematical problems ranging from basic arithmetic to advanced calculus and linear algebra.
-- **Educational Use:** Explains solutions step-by-step, making it a valuable teaching assistant.
-- **Analysis & Reasoning:** Handles logical reasoning tasks and computational queries effectively.
-### **How to Use:**
-1. Download all model files, ensuring the PyTorch weights and tokenizer configurations are included.
-2. Load the model in your Python environment using frameworks like PyTorch or Hugging Face Transformers.
-3. Use the provided configurations (`config.json` and `generation_config.json`) for optimal inference.
----

+---
+license: creativeml-openrail-m
+datasets:
+- prithivMLmods/Math-IIO-68K-Mini
+language:
+- zho
+- eng
+- fra
+- spa
+- por
+- deu
+- ita
+- rus
+- jpn
+- kor
+- vie
+- tha
+- ara
+base_model:
+- Qwen/Qwen2.5-7B-Instruct
+pipeline_tag: text-generation
+library_name: transformers
+tags:
+- safetensors
+- qwen2.5
+- 7B
+- Instruct
+- Math
+- CoT
+- one-shot
+---
+![aaa.png](https://cdn-uploads.huggingface.co/production/uploads/65bb837dbfb878f46c77de4c/faLfR-doaWP_BLUkOQrbq.png)
+### **Math IIO 7B Instruct**
+The **Math IIO 7B Instruct** is a fine-tuned language model based on the robust **Qwen2.5-7B-Instruct** architecture. This model has been specifically trained to excel in single-shot mathematical reasoning and instruction-based tasks, making it a reliable choice for educational, analytical, and problem-solving applications.
+### **Key Features:**
+1. **Math-Optimized Capabilities:**
+   The model is designed to handle complex mathematical problems, step-by-step calculations, and reasoning tasks.
+2. **Instruction-Tuned:**
+   Fine-tuned for better adherence to structured queries and task-oriented prompts, enabling clear and concise outputs.
+3. **Large Vocabulary:**
+   Equipped with an extensive tokenizer configuration and custom tokens to ensure precise mathematical notation support.
+### Single Shot Answers
+![solution.png](https://cdn-uploads.huggingface.co/production/uploads/65bb837dbfb878f46c77de4c/4Zq6crBrbFLDqfKlDwBMU.png)
+### Math-IIO File Structure
+| File Name [ Uploaded file ]                         | Size       | Description                                   | Upload Status  |
+|------------------------------------|------------|-----------------------------------------------|----------------|
+| `.gitattributes`                   | 1.57 kB    | Git attributes configuration file             | Uploaded       |
+| `README.md`                        | 263 Bytes  | README file with minimal details              | Updated        |
+| `added_tokens.json`                | 657 Bytes  | Custom added tokens for tokenizer             | Uploaded       |
+| `config.json`                      | 861 Bytes  | Model configuration file                      | Uploaded       |
+| `generation_config.json`           | 281 Bytes  | Configuration for text generation settings    | Uploaded       |
+| `merges.txt`                       | 1.82 MB    | Merge rules for byte pair encoding tokenizer  | Uploaded       |
+| `pytorch_model-00001-of-00004.bin` | 4.88 GB    | First part of model weights (PyTorch)         | Uploaded (LFS) |
+| `pytorch_model-00002-of-00004.bin` | 4.93 GB    | Second part of model weights (PyTorch)        | Uploaded (LFS) |
+| `pytorch_model-00003-of-00004.bin` | 4.33 GB    | Third part of model weights (PyTorch)         | Uploaded (LFS) |
+| `pytorch_model-00004-of-00004.bin` | 1.09 GB    | Fourth part of model weights (PyTorch)        | Uploaded (LFS) |
+| `pytorch_model.bin.index.json`     | 28.1 kB    | Index JSON file for model weights             | Uploaded       |
+| `special_tokens_map.json`          | 644 Bytes  | Map of special tokens used by the tokenizer   | Uploaded       |
+| `tokenizer.json`                   | 11.4 MB    | Tokenizer settings and vocab                  | Uploaded (LFS) |
+| `tokenizer_config.json`            | 7.73 kB    | Configuration for tokenizer                   | Uploaded       |
+| `vocab.json`                       | 2.78 MB    | Vocabulary for tokenizer                      | Uploaded       |
+| Model Type | Size | Context Length | Link |
+|------------|------|----------------|------|
+| GGUF | 7B | - | [🤗 Math-IIO-7B-Instruct-GGUF](https://huggingface.co/prithivMLmods/Math-IIO-7B-Instruct-GGUF) |
+### **Training Details:**
+- **Base Model:** [Qwen/Qwen2.5-7B-Instruct](#)
+- **Dataset:** Trained on **Math-IIO-68K-Mini**, a curated dataset with 68.8k high-quality examples focusing on mathematical instructions, equations, and logic-based queries.
+### **Capabilities:**
+- **Problem-Solving:** Solves mathematical problems ranging from basic arithmetic to advanced calculus and linear algebra.
+- **Educational Use:** Explains solutions step-by-step, making it a valuable teaching assistant.
+- **Analysis & Reasoning:** Handles logical reasoning tasks and computational queries effectively.
+### **How to Use:**
+1. Download all model files, ensuring the PyTorch weights and tokenizer configurations are included.
+2. Load the model in your Python environment using frameworks like PyTorch or Hugging Face Transformers.
+3. Use the provided configurations (`config.json` and `generation_config.json`) for optimal inference.
+---