VOXPrint T-pro-it-2.1 GGUF (Q4_K_M mirror)
Made for VOXPrint AI Media Suite.
Unmodified mirror of one file from t-tech/T-pro-it-2.1-GGUF, kept as a fallback download source for VOXPrint Plotweaver (24 GB tier).
| Original repository | https://huggingface.co/t-tech/T-pro-it-2.1-GGUF |
| Original commit | cf18787d8901c26ca9334812e7ff8070f2ddab4a |
| Original BF16 model | https://huggingface.co/t-tech/T-pro-it-2.1 |
| File | T-pro-it-2.1-Q4_K_M.gguf, 19,761,766,048 bytes |
| SHA-256 | b19e88cc1154b4b0c99f09bc7b57b8e5f3a6af4135581a3bfd31a2518ed58714 |
| Licence | Apache-2.0, as declared by T-Tech on the original repository (full text in LICENSE) |
The GGUF file is byte-identical to the original at that commit. Only README.md, LICENSE, the .sha256 sidecar and voxprint-model.json were added here. Other quants (Q5_0, Q5_K_M, Q5_K_S, Q6_K, Q8_0) are only on the original repository.
Credit
T-pro-it-2.1 was built by T-Tech (t-tech) on the Qwen3 model family. All rights to the model belong to its authors. This mirror is not affiliated with or endorsed by them.
From the original card: users are responsible for any additional training and oversight needed to make the model's responses meet acceptable ethical and safety standards, and for its use in industrial or commercial solutions. The model supports non-thinking mode only.
Use
Runs with llama.cpp (llama-server -m T-pro-it-2.1-Q4_K_M.gguf). About 20 GB of weights, so it needs a 24 GB GPU with a modest context, or partial CPU offload.
voxprint-model.json is the VOXPrint MDL descriptor (schema voxprint-model/1); it lists this repo, the Kaggle copy and the original repository as mirrors of the same file.