Qwen3-0.6B-FP8-KV

EliovpAI/Qwen3-0.6B-FP8-KV

Qwen3-0.6B-FP8-KV is a text generation model from EliovpAI. Use it when you need the model to write or continue text. It is set up for transformers.

Lightweight OCP FP8e4m3 quant of Qwen3-0.6B with end-to-end KV-cache FP8 support, built with AMD Quark for ROCm.

Downloads · 30 days

20% of all-time downloads

All-time downloads

139

Public

Parameters

596M

752 MB on disk

Likes

Public

Repo makeup

Click a slice to open those files.

.safetensors752 MB · 99%

How the weights are stored.

F8_E4M3440M · 74%

From the Hugging Face model README

Lightweight OCP FP8_e4m3 quant of Qwen3-0.6B with end-to-end KV-cache FP8 support, built with AMD Quark for ROCm.

Qwen3-0.6B-FP8-KV is an OCP-standard FP8_e4m3 quantization of Qwen/Qwen3-0.6B, produced with AMD Quark.

Quantizer: AMD Quark v0.9+
Numeric Format: OCP FP8_e4m3, symmetric per-tensor
Scope: All Linear layers (excl. lm_head), activations and the KV cache
Block Size: 128 (OCP-aligned)
Calibration: 128 Pile samples
Metadata: scales & block info in JSON; weights in SafeTensors

We measured perplexity on WikiText2:

This model inherits the Qwen3-0.6B license.