8 Commits (5f6ffc987a1e1c5380cdcfd1160aeeb8ea7ad831)

Author SHA1 Message Date
BlinkDL dc26998708 torch jit 3 years ago
BlinkDL 75929cbbba torch jit (xx% faster inference) 3 years ago
BlinkDL 2567c8c904 rescale to avoid FP16 overflow 3 years ago
BlinkDL aef9f6f7ef fp16 3 years ago
BlinkDL 8a4a41a3aa better 3 years ago
BlinkDL daed379db2 bf16 inference - 15G VRAM for 7b model 3 years ago
BlinkDL 77bcaa8247 +AE model 3 years ago
BlinkDL 3a7e6a6aa3 faster inference 3 years ago