| --- |
| license: apache-2.0 |
| base_model: Qwen/Qwen3-32B |
| base_model_relation: quantized |
| library_name: epure-runtime |
| pipeline_tag: text-generation |
| language: |
| - en |
| tags: |
| - exeaon |
| - epure |
| - compressed |
| - quantized |
| --- |
| |
| # Exeaon1-Kese-32B |
|
|
| Compressed with E-PURE. Runs with the free |
| [`epure-runtime`](https://github.com/ExeaonLM/epure-runtime), and **stays |
| compressed in memory** — the dense weight is never assembled. |
|
|
| > **Requires `epure-runtime>=0.2.4`.** Earlier versions load every model as |
| > float16. Qwen3-32B is trained in bfloat16 and its activations exceed the |
| > float16 range, so on 0.2.3 this model overflows to infinity partway through |
| > the second layer and produces garbage on prompts of ordinary length. This is |
| > not a soft warning: the failure is silent on short prompts and total on real |
| > ones. |
| > |
| > ``` |
| > pip install "epure-runtime>=0.2.4" |
| > ``` |
|
|
| | | | |
| |---|---| |
| | Base model | [Qwen/Qwen3-32B](https://huggingface.co/Qwen/Qwen3-32B) | |
| | Size on disk | **16.42 GB** | |
| | Compression | **3.99×** | |
| | Format | `.ebin` | |
|
|
| ## Quality |
|
|
| Measured against the base model on the same GPU, the same harness version and |
| the same `limit`. Not copied from anyone's README. |
|
|
| | benchmark | Qwen3-32B | Exeaon1-Kese-32B | delta | |
| |---|---|---|---| |
| | ARC-Challenge | 57.33 | 57.00 | -0.33 | |
| | ARC-Easy | 82.67 | 82.67 | +0.00 | |
| | HellaSwag | 69.33 | 69.67 | +0.34 | |
| | PIQA | 82.33 | 83.00 | +0.67 | |
| | **mean** | **72.91** | **73.09** | **+0.17** | |
|
|
| **Retention: 100.2%** of base mean accuracy. |
|
|
| The compressed model scores fractionally above the original. That is not a |
| gain — it is measurement noise at `limit=300`, where one item is worth 0.33 |
| points. What it does mean is that the loss from compression is smaller than |
| this benchmark can resolve. |
|
|
| Scored at `limit=300`, so roughly 2–3 points of standard error per task. Read |
| the mean, not any single row. |
|
|
| ## Usage |
|
|
| ```python |
| pip install "epure-runtime>=0.2.4" |
| ``` |
|
|
| ```python |
| import epure |
| |
| model, tok = epure.load("model.ebin", device="cuda") |
| print(tok.decode(model.generate(**tok("Hello", return_tensors="pt").to("cuda"), |
| max_new_tokens=50)[0])) |
| ``` |
|
|
| ## Notes on speed |
|
|
| On a large GPU this runs slower than dense fp16. A vendor tensor-core GEMM is |
| heavily tuned and a large card has bandwidth to spare, so trading compute for |
| memory loses there. The win is fitting in less memory — running where the dense |
| model does not fit at all, on a smaller card, or alongside more of them. |
|
|
| Qwen3-32B needs roughly 61 GB in bfloat16. This container needs 16.42 GB on |
| disk and stays packed in memory. |
|
|
| ## Licence |
|
|
| The runtime is Apache-2.0 and free. The compression method that produced this |
| container is proprietary to Zenux Plimver Technologies LTD. |
|
|
