This quant was generated from cerebras/Qwen3-Coder-REAP-25B-A3B using a customized imatrix calibration based on the eaddario/imatrix-calibration datasets.
The code_small dataset is varied and rich in python queries, but lacks significant amounts of C and C++. Therefore, unique C and C++ queries from code_huge were added, in addition to the math_micro dataset for queries containing math instructions.
The calibration set text and .dat files are included.
- Downloads last month
- 298
Hardware compatibility
Log In
to view the estimation
4-bit
Inference Providers
NEW
This model isn't deployed by any Inference Provider.
🙋
Ask for provider support