File size: 713 Bytes
2c96c3a | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 | {
"model": "OpenPerov Flash",
"artifact": "single combined LoRA adapter",
"method": "rank concatenation; unit coefficient for each stage; no rank truncation",
"rank": 48,
"alpha": 96,
"dtype": "float32",
"tensors": 992,
"bytes": 1400885904,
"sha256": "0cc0706fd2ed78240c492b03ce3112e97dfcc08f935d5976b8375451c0c47067",
"all_input_tensor_blocks_preserved": true,
"probe_count": 12,
"maximum_float64_relative_delta_error": 2.804487023000676e-16,
"checkpoint_numerics": "The additive LoRA updates are preserved. Sequential low-precision full-model merges can have different rounding; this exported adapter has not been re-benchmarked.",
"base_model": "Qwen/Qwen3.6-27B"
}
|