So before people take credit, I found the oai os a min after they uploaded and saved the config and other stuff before it was removed. It’s an OS model and coming soon so kinda feels like ruining a surprise
Eh It’s going to come out anyway now Config: {"num_hidden_layers": 36, "num_experts": 128, "experts_per_token": 4, "vocab_size": 201088, "hidden_size": 2880, "intermediate_size": 2880, "swiglu_limit": 7.0, "head_dim": 64, "num_attention_heads": 64, "num_key_value_heads": 8, "sliding_window": 128, "initial_context_length": 4096, "rope_theta": 150000, "rope_scaling_factor": 32.0, "rope_ntk_alpha": 1, "rope_ntk_beta": 32}
@teortaxesTex
13,17K