| --- |
| license: apache-2.0 |
| base_model: Qwen2.5-1.5B-Instruct |
| tags: |
| - code |
| - morphy |
| --- |
| **Morphy-1.5B** is the first model in the Morphy lineup (or family). The model was trained entirely on our proprietary datasets and is optimized for fast performance on local machines. |
|
|
| --- |
|
|
| ## ✨ Strengths (What the model excels at) |
|
|
| * 💻 **Code Generation** — the model is highly capable of writing quality code across a wide variety of programming languages. |
| * 📝 **Documentation Creation** — it excels at composing and structuring `.md` (Markdown) files of any complexity. |
|
|
| --- |
|
|
| ## ⚠️ Important Features and Limitations |
|
|
| Since this is a compact model with 1.5B parameters, it is important for users to keep the following nuances in mind: |
|
|
| * 🧠 **Frequent Hallucinations:** Due to its small size, the model may frequently hallucinate when answering questions. |
| * 🔬 **Complex and Scientific Topics:** The model is prone to making mistakes in complex topics, as well as queries related to physics or biology. |
| * 🛑 **Coding Errors:** The small parameter count can lead to hallucinations even in code, albeit with a lower probability. |
|
|
| > **❗ Recommendation:** We strongly advise always **checking and testing the code** generated by Morphy before deployment. |
|
|
| --- |
|
|
| ## 🛠️ Technical Specifications |
|
|
| | Parameter | Value | |
| | :--- | :--- | |
| | **Name** | Morphy-1.5B | |
| | **Generation** | First model in the family | |
| | **Data Type** | Custom Datasets | |
| | **Purpose** | Code / Markdown Generation | |
