{ "architectures": [ "Qwen2ForCausalLM" ], "attention_dropout": 0.0, "attention_partition_spec": [ [ "dp", "fsdp" ], "sp", "tp", null ], "attn_mechanism": "sharded_vanilla", "axis_dims": [ 1, -1, 1, 1 ], "axis_names": [ "dp", "fsdp", "tp", "sp" ], "backend": null, "bias_partition_spec": [ [ "dp", "fsdp" ], null, null, null ], "bits": null, "block_b": 1, "block_k": 128, "block_k_dkv": 128, "block_k_dq": 128, "block_k_major": 128, "block_k_major_dkv": 128, "block_k_major_dq": 128, "block_q": 128, "block_q_dkv": 128, "block_q_dq": 128, "block_q_major_dkv": 128, "bos_token_id": 151643, "easy_method": "train", "embd_pdrop": 0.0, "eos_token_id": 151645, "fcm_max_ratio": 0.0, "fcm_min_ratio": 0.0, "generation_attention_partition_spec": [ [ "dp", "fsdp" ], null, "tp", null ], "generation_bias_partition_spec": [ [ "dp", "fsdp" ], null, null, null ], "generation_query_partition_spec": [ [ "dp", "fsdp" ], "sp", null, null ], "hidden_act": "silu", "hidden_size": 4096, "initializer_range": 0.02, "intermediate_size": 11008, "key_partition_spec": [ [ "dp", "fsdp" ], "sp", "tp", null ], "max_position_embeddings": 32768, "max_window_layers": 28, "model_type": "qwen2", "num_attention_heads": 32, "num_hidden_layers": 32, "num_key_value_heads": 32, "number_rep_kv": 1, "query_partition_spec": [ [ "dp", "fsdp" ], "sp", "tp", null ], "resid_pdrop": 0.0, "rms_norm_eps": 1e-06, "rope_scaling": null, "rope_theta": 10000.0, "scan_attention_layers": false, "scan_layers": true, "scan_mlp_chunk_size": 1024, "scan_ring_attention": false, "shard_attention_computation": true, "sliding_window": 32768, "tie_word_embeddings": false, "torch_dtype": "float32", "transformers_version": "4.41.0.dev0", "use_cache": true, "use_scan_mlp": false, "use_sharded_kv_caching": true, "use_sharding_constraint": false, "use_sliding_window": false, "value_partition_spec": [ [ "dp", "fsdp" ], "sp", "tp", null ], "vocab_size": 151936 }