| [ | |
| { | |
| "alloc_gb": 0.027, | |
| "amp": false, | |
| "attn_backend": "sublinear", | |
| "batch_size": 1, | |
| "block": 64, | |
| "elapsed_s": 0.471, | |
| "error": null, | |
| "grad_checkpoint": true, | |
| "loss": 18.1082, | |
| "ok": true, | |
| "peak_alloc_gb": 0.031, | |
| "peak_reserved_gb": 0.057, | |
| "reserved_gb": 0.057, | |
| "sublinear_chunk": 16, | |
| "sublinear_max_anchors": 16, | |
| "sublinear_stride": 8, | |
| "sublinear_window": 16, | |
| "tokens_per_s_synthetic": 135.9 | |
| }, | |
| { | |
| "alloc_gb": 0.027, | |
| "amp": false, | |
| "attn_backend": "sublinear", | |
| "batch_size": 1, | |
| "block": 96, | |
| "elapsed_s": 0.149, | |
| "error": null, | |
| "grad_checkpoint": true, | |
| "loss": 17.9191, | |
| "ok": true, | |
| "peak_alloc_gb": 0.033, | |
| "peak_reserved_gb": 0.057, | |
| "reserved_gb": 0.057, | |
| "sublinear_chunk": 16, | |
| "sublinear_max_anchors": 16, | |
| "sublinear_stride": 8, | |
| "sublinear_window": 16, | |
| "tokens_per_s_synthetic": 644.1 | |
| } | |
| ] |