jamba-1.9b-9 / config.json
neeleshg23's picture
Training in progress, step 2000
48bab2b verified
{
"_name_or_path": "neeleshg23/jamba-1.9b-8",
"architectures": [
"JambaForCausalLM"
],
"attention_dropout": 0.0,
"attn_layer_offset": 4,
"attn_layer_period": 8,
"bos_token_id": 1,
"d_model": 1024,
"eos_token_id": 2,
"expert_layer_offset": 1,
"expert_layer_period": 2,
"hidden_act": "silu",
"hidden_size": 4096,
"initializer_range": 0.02,
"intermediate_size": 14336,
"mamba_conv_bias": true,
"mamba_d_conv": 4,
"mamba_d_state": 16,
"mamba_dt_rank": 256,
"mamba_expand": 2,
"mamba_proj_bias": false,
"max_position_embeddings": 2048,
"model_type": "jamba",
"num_attention_heads": 16,
"num_experts": 4,
"num_experts_per_tok": 2,
"num_hidden_layers": 3,
"num_key_value_heads": 8,
"num_logits_to_keep": 1,
"output_router_logits": false,
"pad_token_id": 0,
"rms_norm_eps": 1e-06,
"router_aux_loss_coef": 0.001,
"sliding_window": null,
"tie_word_embeddings": true,
"torch_dtype": "bfloat16",
"transformers_version": "4.46.2",
"use_cache": true,
"use_mamba_kernels": true,
"vocab_size": 128256
}