Language Modeling on WikiText-103 (Hardware Efficiency)
6.88PPLT3C
Evaluation results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| T3CBackbone=TinyLlama-1.1B, Budget=Tiny2026.01 | 6.88 | 7.6 | 9.8 | 92 | 118 | 151 | 198 | 210 | 0.26 | |
| T3CBackbone=TinyLlama-1.1B, Budget=Med2026.01 | 7.02 | 7.9 | 10.1 | 96 | 122 | 156 | 204 | 225 | 0.24 | |
| KD-PTQBackbone=TinyLlama-1.1B, Bit-width=8b2026.01 | 7.04 | 8.8 | 11 | 121 | 154 | 189 | 246 | 418 | 0.33 | |
| T3CBackbone=TinyLlama-1.1B, Budget=Max2026.01 | 7.1 | 8.2 | 10.5 | 99 | 127 | 160 | 210 | 244 | 0.27 | |
| MP-QATBackbone=TinyLlama-1.1B2026.01 | 7.12 | 9.3 | 11.9 | 133 | 168 | 209 | 268 | 436 | 0.36 | |
| LR+FTBackbone=TinyLlama-1.1B2026.01 | 7.22 | 9.1 | 11.6 | 128 | 162 | 201 | 258 | 312 | 0.38 | |
| SparseGPTBackbone=TinyLlama-1.1B2026.01 | 7.36 | 8.5 | 10.8 | 112 | 144 | 176 | 228 | 238 | 0.41 | |
| PTQBackbone=TinyLlama-1.1B, Bit-width=4b2026.01 | 7.58 | 8.7 | 11.2 | 125 | 159 | 196 | 252 | 280 | 0.49 |