{ "step": 700, "stage": "sft", "val_loss": 0.5377802371978759, "version": 4, "params": 100047616, "dtype": "bf16", "config": { "vocab_size": 8192, "block_size": 1024, "n_layer": 34, "n_head": 8, "n_embd": 512, "dropout": 0.0, "n_kv_head": 2, "qk_norm": true, "arch_version": 4 }, "files": [ "model.pt", "model-00001-of-00005.pt", "model-00002-of-00005.pt", "model-00003-of-00005.pt", "model-00004-of-00005.pt", "model-00005-of-00005.pt" ] }