maidacundo/text2sql_phi_molora
0236
1{2 "architectures": [3 "PhiMoLForCausalLM"4 ],5 "attention_dropout": 0.0,6 "bos_token_id": 1,7 "embd_pdrop": 0.0,8 "eos_token_id": 2,9 "expert_rank": 2,10 "hidden_act": "gelu_new",11 "hidden_size": 32,12 "initializer_range": 0.02,13 "intermediate_size": 64,14 "layer_norm_eps": 1e-05,15 "max_position_embeddings": 2048,16 "model_type": "phi",17 "num_attention_heads": 32,18 "num_experts_per_tok": 2,19 "num_hidden_layers": 24,20 "num_key_value_heads": 32,21 "num_local_experts": 8,22 "output_router_logits": false,23 "partial_rotary_factor": 0.5,24 "qk_layernorm": false,25 "resid_pdrop": 0.0,26 "rope_scaling": null,27 "rope_theta": 10000.0,28 "router_aux_loss_coef": 0.001,29 "tie_word_embeddings": false,30 "torch_dtype": "float32",31 "transformers_version": "4.37.0",32 "use_cache": false,33 "vocab_size": 5120034}35 