Update README.md
Browse files
README.md
CHANGED
|
@@ -12,4 +12,93 @@ base_model:
|
|
| 12 |
- Qwen/Qwen2.5-Coder-3B-Instruct
|
| 13 |
pipeline_tag: text-generation
|
| 14 |
library_name: transformers
|
| 15 |
-
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 12 |
- Qwen/Qwen2.5-Coder-3B-Instruct
|
| 13 |
pipeline_tag: text-generation
|
| 14 |
library_name: transformers
|
| 15 |
+
---
|
| 16 |
+
|
| 17 |
+
# **Omega-Qwen2.5-Coder-3B**
|
| 18 |
+
|
| 19 |
+
> **Omega-Qwen2.5-Coder-3B** is a compact and high-efficiency code-focused model fine-tuned on **Qwen2.5-Coder-3B-Instruct**, using the symbolic-rich **Open-Omega-Forge-1M** dataset. Designed specifically for **hard-coded tasks** and deterministic computation, this model runs in a **"thinking-disabled"** mode—delivering precise, structured outputs with minimal hallucination, making it ideal for rigorous coding workflows and embedded logic applications.
|
| 20 |
+
|
| 21 |
+
> Thinking: Disabled
|
| 22 |
+
|
| 23 |
+
> \[!note]
|
| 24 |
+
> GGUF: [https://huggingface.co/prithivMLmods/Omega-Qwen2.5-Coder-3B-GGUF](https://huggingface.co/prithivMLmods/Omega-Qwen2.5-Coder-3B-GGUF)
|
| 25 |
+
|
| 26 |
+
## **Key Features**
|
| 27 |
+
|
| 28 |
+
1. **Purpose-Built for Hard Coding**
|
| 29 |
+
Specially tuned to perform precise, low-level code generation with **minimal reasoning overhead**. Ideal for edge-case algorithms, embedded scripting, and deterministic logic patterns.
|
| 30 |
+
|
| 31 |
+
2. **Optimized Qwen2.5 Foundation**
|
| 32 |
+
Built on **Qwen2.5-Coder-3B-Instruct**, benefiting from its robust token handling, instruction following, and multilingual code representation.
|
| 33 |
+
|
| 34 |
+
3. **Backed by Open-Omega-Forge-1M Dataset**
|
| 35 |
+
Trained on a curated mix of code, math, and logic problems focused on **symbolic clarity** and **STEM coherence**, drawn from sources like OpenCodeReasoning, MathX-5M, OpenMathReasoning, and more.
|
| 36 |
+
|
| 37 |
+
4. **Thinking Disabled Mode**
|
| 38 |
+
The model avoids overgeneralizing or injecting speculative reasoning. It executes tasks **as-is**—perfect for structured prompts, tight constraints, and automation pipelines.
|
| 39 |
+
|
| 40 |
+
5. **Structured Output Control**
|
| 41 |
+
Outputs in **JSON**, **YAML**, **Python**, **Markdown**, and **LaTeX**, tailored for script generation, data serialization, and scientific formatting.
|
| 42 |
+
|
| 43 |
+
6. **Efficient 3B Deployment**
|
| 44 |
+
Lightweight and scalable for **mid-tier GPUs**, **offline dev environments**, or **local inference systems**, while maintaining solid performance on symbolic tasks.
|
| 45 |
+
|
| 46 |
+
---
|
| 47 |
+
|
| 48 |
+
## **Quickstart with Transformers**
|
| 49 |
+
|
| 50 |
+
```python
|
| 51 |
+
from transformers import AutoModelForCausalLM, AutoTokenizer
|
| 52 |
+
|
| 53 |
+
model_name = "prithivMLmods/Omega-Qwen2.5-Coder-3B"
|
| 54 |
+
|
| 55 |
+
model = AutoModelForCausalLM.from_pretrained(
|
| 56 |
+
model_name,
|
| 57 |
+
torch_dtype="auto",
|
| 58 |
+
device_map="auto"
|
| 59 |
+
)
|
| 60 |
+
tokenizer = AutoTokenizer.from_pretrained(model_name)
|
| 61 |
+
|
| 62 |
+
prompt = "Write a Python function to compute factorial iteratively."
|
| 63 |
+
|
| 64 |
+
messages = [
|
| 65 |
+
{"role": "system", "content": "You are a deterministic code generator. No assumptions. No extra explanations."},
|
| 66 |
+
{"role": "user", "content": prompt}
|
| 67 |
+
]
|
| 68 |
+
|
| 69 |
+
text = tokenizer.apply_chat_template(
|
| 70 |
+
messages,
|
| 71 |
+
tokenize=False,
|
| 72 |
+
add_generation_prompt=True
|
| 73 |
+
)
|
| 74 |
+
|
| 75 |
+
model_inputs = tokenizer([text], return_tensors="pt").to(model.device)
|
| 76 |
+
|
| 77 |
+
generated_ids = model.generate(
|
| 78 |
+
**model_inputs,
|
| 79 |
+
max_new_tokens=256
|
| 80 |
+
)
|
| 81 |
+
generated_ids = [
|
| 82 |
+
output_ids[len(input_ids):] for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids)
|
| 83 |
+
]
|
| 84 |
+
|
| 85 |
+
response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
|
| 86 |
+
print(response)
|
| 87 |
+
```
|
| 88 |
+
|
| 89 |
+
---
|
| 90 |
+
|
| 91 |
+
## **Intended Use**
|
| 92 |
+
|
| 93 |
+
* Embedded logic and deterministic function generation
|
| 94 |
+
* Script automation and toolchain integration
|
| 95 |
+
* Codegen under fixed constraints or symbolic inputs
|
| 96 |
+
* Lightweight STEM applications on edge devices or offline clusters
|
| 97 |
+
* Tools where **"no thinking" = better stability**
|
| 98 |
+
|
| 99 |
+
## **Limitations**
|
| 100 |
+
|
| 101 |
+
* Not suitable for high-level reasoning or open-ended thought processes
|
| 102 |
+
* General chat performance is minimal by design
|
| 103 |
+
* Lacks emotional intelligence or creative composition capability
|
| 104 |
+
* Assumes user provides **clear, explicit instructions** for best results
|