File size: 1,540 Bytes
06ba1d4
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
{
  "date": "2026-10-19",
  "donor_commit": "529ee4b691732e1040d6a45f05fd12f58def96ec",
  "platform": "WSL Linux, CPU only",
  "python": "3.12.3",
  "torch": "2.14.0+cpu",
  "command": "TORCH_COMPILE_DISABLE=1 OMP_NUM_THREADS=1 python -m unittest discover -s tests -v",
  "full_suite": {
    "passed": 65,
    "skipped": 0,
    "elapsed_seconds": 76.884
  },
  "align_new_parity": {
    "cases": 7,
    "max_loss_abs_error": 0.0,
    "max_gradient_abs_error": 5.960464477539063e-08,
    "dtype": "float32",
    "atol": 1e-6,
    "rtol": 1e-5,
    "coverage": "Run/xtoken chunks, final softcap, rho=0/0.1/1, active cross-chunk guard",
    "oracle": "Donor method bodies plus separately loaded original guard/head-pair copies; frozen source-verified v1 dependencies shared"
  },
  "additional_checks": [
    "Independent autograd full-response head norms and dot, including bias and softcap",
    "Vocabulary tiling invariance",
    "CPU BF16 ambient autocast preserves original credit-logit/geometry composition",
    "Teacher stop-gradient, microbatch normalization, invalid and degenerate geometry",
    "Existing Strict, author-code SimCT and MP numerical regressions/source hashes",
    "mp-align-new dry-run retains batch512, steps100, xtoken, exact_head and rho0.1",
    "git diff --check"
  ],
  "not_qualified": [
    "Target Torch 2.11 CUDA, B200/FSDP integration and full-size memory/throughput",
    "Full Transformer or Adam descent guarantees",
    "Scientific efficacy, paper-score reproduction and full training resume"
  ]
}