| { | |
| "schema": "alpha-native-model-config-v1", | |
| "architecture": "gpt", | |
| "model": { | |
| "vocabSize": 12288, | |
| "blockSize": 512, | |
| "nLayer": 16, | |
| "nEmbd": 512, | |
| "nHead": 8, | |
| "dropout": 0, | |
| "ffnActivation": "gelu", | |
| "normType": "layernorm", | |
| "posEnc": "learned", | |
| "tieEmbeddings": false, | |
| "attentionSoftCap": 30 | |
| }, | |
| "tokenizer": "bpe-byte-12k", | |
| "parameterCount": 63210496, | |
| "checkpointFormat": "alpha-native-full-training-state" | |
| } | |