config.json

{
  "_block_types": [
    "recurrent",
    "recurrent",
    "attention"
  ],
  "_name_or_path": "./griffin-1024-c3t-8layer",
  "architectures": [
    "RecurrentGemmaForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "attention_window_size": 2048,
  "block_types": [
    "recurrent",
    "recurrent",
    "attention"
  ],
  "bos_token_id": 0,
  "conv1d_width": 4,
  "embeddings_scale_by_sqrt_dim": true,
  "eos_token_id": 0,
  "final_w_init_variance_scale": 0.25,
  "head_dim": 128,
  "hidden_activation": "gelu_pytorch_tanh",
  "hidden_size": 1024,
  "intermediate_size": 6144,
  "logits_soft_cap": 30.0,
  "lru_width": 1024,
  "model_type": "recurrent_gemma",
  "num_attention_heads": 8,
  "num_hidden_layers": 8,
  "num_key_value_heads": 2,
  "pad_token_id": 0,
  "partial_rotary_factor": 0.5,
  "rms_norm_eps": 1e-06,
  "rope_theta": 10000.0,
  "torch_dtype": "float32",
  "transformers_version": "4.40.1",
  "use_cache": true,
  "vocab_size": 65024,
  "w_init_variance_scale": 0.01
}