Wan2.2-T2V-A14B-GGUF is a quantized version in GGUF format of the 14-billion parameter Wan2.2 text-to-video model. It enables local inference of text-to-video generation on compatible hardware through reduced memory requirements offered by multiple quantization levels.
The repository provides GGUF files at quantization levels ranging from Q2_K at 5.3 GB to Q8_0 at 15.4 GB. Specific variants include Q3_K_S at 6.51 GB, Q3_K_M at 7.17 GB, Q4_K_S at 8.75 GB, Q4_0 at 8.56 GB, Q4_1 at 9.26 GB, Q4_K_M at 9.65 GB, Q5_K_S at 10.1 GB, Q5_0 at 10.3 GB, Q5_1 at 11 GB, Q5_K_M at 10.8 GB, and Q6_K at 12 GB. This model is a direct conversion of the original Wan-AI/Wan2.2-T2V-A14B, so all original licensing terms and usage restrictions apply.
It integrates with the ComfyUI custom node ComfyUI-GGUF developed by city96. Model files are placed in the ComfyUI/models/unet directory, with further setup details available in the associated GitHub readme. The underlying architecture is listed as wan. The repository records 94,545 downloads in the last month and carries an apache-2.0 license.
Wan2.2 T2V A14B is a Text to video project. It focuses on running high-quality text-to-video generation models locally with significantly reduced memory requirements. Wan2.2 T2V A14B is an open-source project aimed at developers. Wan2.2 T2V A14B is open source under the Apache-2.0 license. It ships for the web and API.
QuantStack builds and maintains Wan2.2 T2V A14B, and it first shipped in 2024. The project is developed in the open on GitHub with 3.8k stars. Key capabilities include text-to-Video, GGUF Quantization, and ComfyUI Compatible.
Summary written by a language model from the project’s public pages.
What PulseGate has recorded for this listing
Closest matches by what these projects do