File size: 1,025 Bytes
884aa0c
 
 
1121776
 
 
884aa0c
 
e6a1a0e
 
884aa0c
 
 
e6a1a0e
 
884aa0c
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
---
base_model:
- Qwen/QwQ-32B
license: apache-2.0
datasets:
- Rombo-Org/Optimized_Reasoning
---

# QwQ-32B-nvfp4

**Format:** NVFP4 — weights & activations quantized to FP4 with dual scaling.  
**Base model:** `Qwen/QwQ-32B`  
**How it was made:** One-shot calibration with LLM Compressor (NVFP4 recipe), long-seq calibration with Rombo-Org/Optimized_Reasoning.

> Notes: Keep `lm_head` in high precision; calibrate on long, domain-relevant sequences.

Check the original model card for information about this model. 

# Running the model with VLLM in Docker
```sh
 sudo docker run --runtime nvidia --gpus all    -p  8000:8000   --ipc=host   vllm/vllm-openai:latest --model Firworks/QwQ-32B-nvfp4   --dtype auto   --max-model-len 32768
```
This was tested on a B200 cloud instance.

If there are other models you're interested in seeing quantized to NVFP4 for use on the DGX Spark, or other modern Blackwell (or newer) cards let me know. I'm trying to make more NVFP4 models available to allow more people to try them out.