初始化项目，由ModelHub XC社区提供模型

Model: Flexan/DQN-Labs-dqnMath-v0.2-3.8B-HF-GGUF Source: Original Platform
2026-04-11 17:12:55 +08:00
commit 42a0807df9
16 changed files with 171 additions and 0 deletions
--- a/README.md
+++ b/README.md
@@ -0,0 +1,80 @@
+---
+language:
+- en
+library_name: mlx
+license: mit
+license_link: https://huggingface.co/microsoft/Phi-4-mini-instruct-reasoning/resolve/main/LICENSE
+pipeline_tag: text-generation
+tags:
+- nlp
+- math
+- code
+- mlx
+widget:
+- messages:
+  - role: user
+    content: How to solve 3*x^2+4*x+5=1?
+base_model: DQN-Labs/dqnMath-v0.2-3.8B-HF
+---
+
+# GGUF Files for dqnMath-v0.2-3.8B-HF
+
+These are the GGUF files for [DQN-Labs/dqnMath-v0.2-3.8B-HF](https://huggingface.co/DQN-Labs/dqnMath-v0.2-3.8B-HF).
+
+## Downloads
+
+| GGUF Link | Quantization | Description |
+| ---- | ----- | ----------- |
+| [Download](https://huggingface.co/Flexan/DQN-Labs-dqnMath-v0.2-3.8B-HF-GGUF/resolve/main/dqnMath-v0.2-3.8B-HF.Q2_K.gguf) | Q2_K | Lowest quality |
+| [Download](https://huggingface.co/Flexan/DQN-Labs-dqnMath-v0.2-3.8B-HF-GGUF/resolve/main/dqnMath-v0.2-3.8B-HF.Q3_K_S.gguf) | Q3_K_S |  |
+| [Download](https://huggingface.co/Flexan/DQN-Labs-dqnMath-v0.2-3.8B-HF-GGUF/resolve/main/dqnMath-v0.2-3.8B-HF.IQ3_S.gguf) | IQ3_S | Integer quant, preferable over Q3_K_S |
+| [Download](https://huggingface.co/Flexan/DQN-Labs-dqnMath-v0.2-3.8B-HF-GGUF/resolve/main/dqnMath-v0.2-3.8B-HF.IQ3_M.gguf) | IQ3_M | Integer quant |
+| [Download](https://huggingface.co/Flexan/DQN-Labs-dqnMath-v0.2-3.8B-HF-GGUF/resolve/main/dqnMath-v0.2-3.8B-HF.Q3_K_M.gguf) | Q3_K_M |  |
+| [Download](https://huggingface.co/Flexan/DQN-Labs-dqnMath-v0.2-3.8B-HF-GGUF/resolve/main/dqnMath-v0.2-3.8B-HF.Q3_K_L.gguf) | Q3_K_L |  |
+| [Download](https://huggingface.co/Flexan/DQN-Labs-dqnMath-v0.2-3.8B-HF-GGUF/resolve/main/dqnMath-v0.2-3.8B-HF.IQ4_XS.gguf) | IQ4_XS | Integer quant |
+| [Download](https://huggingface.co/Flexan/DQN-Labs-dqnMath-v0.2-3.8B-HF-GGUF/resolve/main/dqnMath-v0.2-3.8B-HF.Q4_K_S.gguf) | Q4_K_S | Fast with good performance |
+| [Download](https://huggingface.co/Flexan/DQN-Labs-dqnMath-v0.2-3.8B-HF-GGUF/resolve/main/dqnMath-v0.2-3.8B-HF.Q4_K_M.gguf) | Q4_K_M | **Recommended:** Perfect mix of speed and performance |
+| [Download](https://huggingface.co/Flexan/DQN-Labs-dqnMath-v0.2-3.8B-HF-GGUF/resolve/main/dqnMath-v0.2-3.8B-HF.Q5_K_S.gguf) | Q5_K_S |  |
+| [Download](https://huggingface.co/Flexan/DQN-Labs-dqnMath-v0.2-3.8B-HF-GGUF/resolve/main/dqnMath-v0.2-3.8B-HF.Q5_K_M.gguf) | Q5_K_M |  |
+| [Download](https://huggingface.co/Flexan/DQN-Labs-dqnMath-v0.2-3.8B-HF-GGUF/resolve/main/dqnMath-v0.2-3.8B-HF.Q6_K.gguf) | Q6_K | Very good quality |
+| [Download](https://huggingface.co/Flexan/DQN-Labs-dqnMath-v0.2-3.8B-HF-GGUF/resolve/main/dqnMath-v0.2-3.8B-HF.Q8_0.gguf) | Q8_0 | Best quality |
+| [Download](https://huggingface.co/Flexan/DQN-Labs-dqnMath-v0.2-3.8B-HF-GGUF/resolve/main/dqnMath-v0.2-3.8B-HF.f16.gguf) | f16 | Full precision, don't bother; use a quant |
+
+## Note from Flexan
+
+I provide GGUFs and quantizations of publicly available models that do not have a GGUF equivalent available yet,
+usually for models **I deem interesting and wish to try out**.
+
+If there are some quants missing that you'd like me to add, you may request one in the community tab.
+If you want to request a public model to be converted, you can also request that in the community tab.
+If you have questions regarding this model, please refer to [the original model repo](https://huggingface.co/DQN-Labs/dqnMath-v0.2-3.8B-HF).
+
+You can find more info about me and what I do [here](https://huggingface.co/Flexan/Flexan).
+
+# DQN-Labs/dqnMath-v0.2-3.8B-HF
+
+This model [DQN-Labs/dqnMath-v0.2-3.8B-HF](https://huggingface.co/DQN-Labs/dqnMath-v0.2-3.8B-HF) was
+converted to MLX format from [microsoft/Phi-4-mini-reasoning](https://huggingface.co/microsoft/Phi-4-mini-reasoning)
+using mlx-lm version **0.30.7**.
+
+## Use with mlx
+
+```bash
+pip install mlx-lm
+```
+
+```python
+from mlx_lm import load, generate
+
+model, tokenizer = load("DQN-Labs/dqnMath-v0.2-3.8B-HF")
+
+prompt = "hello"
+
+if tokenizer.chat_template is not None:
+    messages = [{"role": "user", "content": prompt}]
+    prompt = tokenizer.apply_chat_template(
+        messages, add_generation_prompt=True, return_dict=False,
+    )
+
+response = generate(model, tokenizer, prompt=prompt, verbose=True)
+```