| { |
| "architectures": [ |
| "SparseLlamaForCausalLM" |
| ], |
| "attention_bias": false, |
| "attention_dropout": 0.0, |
| "bos_token_id": 0, |
| "dtype": "bfloat16", |
| "eos_token_id": 0, |
| "head_dim": 64, |
| "hidden_act": "silu", |
| "hidden_size": 2048, |
| "initializer_range": 0.02, |
| "intermediate_size": 5632, |
| "is_llama_config": true, |
| "max_position_embeddings": 2048, |
| "mlp_bias": false, |
| "model_type": "llama_sparse_relu", |
| "num_attention_heads": 32, |
| "num_hidden_layers": 38, |
| "num_key_value_heads": 32, |
| "pretraining_tp": 1, |
| "rms_norm_eps": 1e-05, |
| "rope_interleaved": false, |
| "rope_scaling": null, |
| "rope_theta": 100000, |
| "sparsity_baseline_run": false, |
| "sparsity_do_track": true, |
| "sparsity_end_warmup_step": null, |
| "sparsity_gated_mlp": true, |
| "sparsity_l1_coeff": 2e-05, |
| "sparsity_max_seq_len": 2048, |
| "sparsity_max_seq_length": 2048, |
| "sparsity_mlp_execution_logic": "training", |
| "sparsity_new_intermediate_size": 5632, |
| "sparsity_per_device_train_batch_size": 8, |
| "sparsity_reinitialize": true, |
| "sparsity_start_warmup_step": null, |
| "sparsity_tokens_until_dead": null, |
| "sparsity_train_batch_size": 512, |
| "tie_word_embeddings": true, |
| "transformers_version": "4.56.2", |
| "use_cache": true, |
| "vocab_size": 49152 |
| } |
|
|
