Downloads · 30 days
16
8% of all-time downloads
granite-4.0-h-small-quantized.w8a8
fedora-copr/granite-4.0-h-small-quantized.w8a8
granite-4.0-h-small-quantized.w8a8 is a text generation model from fedora-copr. Use it when you need the model to write or continue text. The card lists the license as apache-2.0.
This is a W8A8 INT8 quantized version of ibm-granite/granite-4.0-h-small.
All-time downloads
205
Public
Parameters
32.2B
32.7 GB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors32.7 GB · 100%
Parameter types
How the weights are stored.
I831.8B · 99%
Try a prompt
Base models
Model card
From the Hugging Face model README
fedora-copr/granite-4.0-h-small-quantized.w8a8
This is a W8A8 INT8 quantized version of ibm-granite/granite-4.0-h-small.
Model Details
-
Quantized by: Jiri Podivin [email protected]
-
Architecture: Granite-4.0 Hybrid MoE (Mamba + Transformer)
-
Quantization: INT8 Weight & Activation (W8A8)
-
Engine Support: vLLM (0.6.0+)
Implementation
The quantization was performed using the llm-compressor library.
vLLM Serving
vllm serve fedora-copr/granite-4.0-h-small-quantized.w8a8 --quantization compressed-tensors
- Task
- Text Generation
- Type
- granitemoehybrid
- License
- apache-2.0
- Languages
- en
- Created
- Mar 4, 2026
- Updated
- Mar 5, 2026