reducto

qwen36-extract-bs8-ckpt888-20260703-fp8

Deploy Dedicated

Dedicated Endpoints

Run this model inference on single tenant GPU with unmatched speed and reliability at scale.

Learn more

Get help setting up a custom Dedicated Endpoints.

Talk with our engineer to get a quote for reserved GPU instances with discounts.

README

qwen36 extract checkpoint 888 FP8

This is an FP8 compressed-tensors export of /workspace/home/apurv/rdt/extract-grpo/hf-only/qwen36-27b-ckpt888.

Quantization scheme: FP8_DYNAMIC
Ignored modules: lm_head, re:.*visual.*
Intended runtime: vLLM / Transformers with compressed-tensors support

The source checkpoint was trained with FP8 mixed precision, but the public checkpoint stores BF16 weights. This repo contains an inference-oriented FP8 conversion of those BF16 weights.

Hub repo: reducto/qwen36-extract-bs8-ckpt888-20260703-fp8

Model provider

reducto

Model tree

Base

reducto/qwen36-extract-bs8-ckpt888-20260703

Quantized

this model

Modalities

Input

Video, Text, Image

Output

Text

Pricing

Dedicated Endpoints

View details

Supported Functionality

Model APIs

Dedicated Endpoints

Container

More information

Model card

Explore FriendliAI today

Get started Talk to an engineer