---
license: mit
pipeline_tag: text-generation
tags:
- code
- deepseek_v3
- qwen
- int4
- conversational
---

# DeepSeek-R1-Distill-Qwen-32B-AWQ wint4

Distillation of DeepSeek-R1 to Qwen 32B, quantized using AWQ to wint4. It fits on any 24GB VRAM GPU or 32GB URAM device!

## MMLU-PRO

The MMLU-PRO dataset evaluates subjects across 14 distinct fields using a 5-shot accuracy measurement. Each task assesses models following the methodology of the original MMLU implementation, with each having ten possible choices.

### Measure

- **Accuracy**: Evaluated as "exact_match"

### Shots

- **Shots**: 5-shot

### Results Table

| Tasks                     | Version | Filter        | n-shot | Metric     | Direction | Value | Stderr |
|---------------------------|---------|---------------|--------|------------|-----------|-------|--------|
| mmlu_pro                  | 2       | custom-extract|        | exact_match| ↑         | 0.5875| 0.0044 |
| biology                   | 1       | custom-extract| 5      | exact_match| ↑         | 0.7978| 0.0150 |
| business                  | 1       | custom-extract| 5      | exact_match| ↑         | 0.5982| 0.0175 |
| chemistry                 | 1       | custom-extract| 5      | exact_match| ↑         | 0.4691| 0.0148 |
| computer_science          | 1       | custom-extract| 5      | exact_match| ↑         | 0.6122| 0.0241 |
| economics                 | 1       | custom-extract| 5      | exact_match| ↑         | 0.7346| 0.0152 |
| engineering               | 1       | custom-extract| 5      | exact_match| ↑         | 0.3891| 0.0157 |
| health                    | 1       | custom-extract| 5      | exact_match| ↑         | 0.6345| 0.0168 |
| history                   | 1       | custom-extract| 5      | exact_match| ↑         | 0.6168| 0.0249 |
| law                       | 1       | custom-extract| 5      | exact_match| ↑         | 0.4596| 0.0150 |
| math                      | 1       | custom-extract| 5      | exact_match| ↑         | 0.6425| 0.0130 |
| other                     | 1       | custom-extract| 5      | exact_match| ↑         | 0.6223| 0.0160 |
| philosophy                | 1       | custom-extract| 5      | exact_match| ↑         | 0.5731| 0.0222 |
| physics                   | 1       | custom-extract| 5      | exact_match| ↑         | 0.5073| 0.0139 |
| psychology                | 1       | custom-extract| 5      | exact_match| ↑         | 0.7494| 0.0154 |

## Groups

| Groups    | Version | Filter        | n-shot | Metric     | Direction | Value | Stderr |
|-----------|---------|---------------|--------|------------|-----------|-------|--------|
| mmlu_pro  | 2       | custom-extract|        | exact_match| ↑         | 0.5875| 0.0044 |