Damork-Coder-27B Multimodal FP8

FP8 quantized version of PWLabs/Damork-Coder-27B-multimodal.

This model is based on Qwen/Qwen3.6-27B, merged with the Damork LoRA adapter, then quantized to FP8.

Usage

from transformers import AutoProcessor, AutoModelForImageTextToText

model_id = "PWLabs/Damork-Coder-27B-multimodal-FP8"

processor = AutoProcessor.from_pretrained(model_id, trust_remote_code=True)
model = AutoModelForImageTextToText.from_pretrained(
    model_id,
    trust_remote_code=True,
    device_map="auto",
)

Notes

The original Damork adapter was trained mainly on language-side LoRA modules. The multimodal capability is preserved from the Qwen3.6 base model.

Downloads last month
15
Safetensors
Model size
27B params
Tensor type
BF16
·
F32
·
F8_E4M3
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for fares-boutriga/Damork-Coder-27B-multimodal-FP8

Base model

Qwen/Qwen3.6-27B
Quantized
(761)
this model