Stitched HIGGS Llama3 8B mixed-precision model variants.
-
inference-optimization/llama3_8b_5.0_bits_mode_heuristic_stiched
5B • Updated • 10 -
inference-optimization/llama3_8b_5.0_bits_mode_hybrid_stiched
5B • Updated • 11 -
inference-optimization/llama3_8b_5.0_bits_mode_noise_stiched
5B • Updated • 5 -
inference-optimization/llama3_8b_5.5_bits_mode_heuristic_stiched
6B • Updated • 4