| { | |
| "architectures": [ | |
| "Model" | |
| ], | |
| "attention_dropout": 0.0, | |
| "bos_token_id": 1, | |
| "clip_qkv": 8, | |
| "eos_token_id": 2, | |
| "fused_attention_norm": true, | |
| "gated_mlp": true, | |
| "glm_tokens": null, | |
| "gradient_checkpointing": false, | |
| "hidden_act": "silu", | |
| "hidden_size": 1024, | |
| "initializer_range": 0.02, | |
| "intermediate_size": 3072, | |
| "max_num_sequences": 512, | |
| "max_position_embeddings": 65536, | |
| "memory_optimized_mlp": true, | |
| "mlm_loss_coef": 1.0, | |
| "model_type": "progen3", | |
| "moe_expert_selection": "switch", | |
| "moe_grouped_gemm": true, | |
| "moe_implementation": "megablocks", | |
| "moe_memory_optimized": true, | |
| "moe_world_size": 1, | |
| "msa_style_attention": true, | |
| "no_ffn_gradient_checkpointing": true, | |
| "num_attention_heads": 16, | |
| "num_experts": 8, | |
| "num_experts_per_tok": 2, | |
| "num_hidden_layers": 17, | |
| "num_key_value_heads": 16, | |
| "output_router_weights": false, | |
| "pad_token_id": 0, | |
| "quantize_inputs_num_bits": 8, | |
| "quantize_rematerialize_num_bits": 8, | |
| "quantize_scatter_num_bits": 8, | |
| "rms_norm_eps": 1e-05, | |
| "rope_theta": 100000.0, | |
| "router_aux_loss_coef": 0.05, | |
| "tie_word_embeddings": false, | |
| "torch_dtype": "bfloat16", | |
| "transformers_version": "4.42.4", | |
| "use_cache": true, | |
| "vocab_size": 134 | |
| } | |