Skip to content

[CUDA] Fixing quantization uint8 packing bug for NF4 and FP4 #1681

[CUDA] Fixing quantization uint8 packing bug for NF4 and FP4

[CUDA] Fixing quantization uint8 packing bug for NF4 and FP4 #1681

Job Run time
12s
19s
9m 52s
9m 12s
18s
1m 36s
8m 8s
5m 30s
5m 26s
5m 30s
5m 54s
3m 58s
7m 28s
7m 15s
5m 6s
5m 54s
5m 22s
6m 44s
5m 25s
5m 6s
5m 53s
4m 30s
6m 48s
4m 22s
5m 14s
5m 18s
6m 56s
5m 48s
5m 40s
3m 55s
5m 30s
6m 54s
6m 31s
6m 28s
6m 37s
4m 58s
6m 38s
1m 11s
44s
2m 29s
39s
16s
16s
0s
0s
3h 27m 50s