{
  "_name_or_path": "/home/dinalt/ai_assets/models/walsh_instruct",
  "activation_args": {},
  "activation_cls": "torch.nn.GELU",
  "architectures": [
    "HFCausalModel"
  ],
  "attention_args": {
    "beta": 0.25,
    "dropout": 0.1
  },
  "attention_cls": ".CausalSelfAttention",
  "auto_map": {
    "AutoConfig": "modelling_walsh.Config",
    "AutoModelForCausalLM": "modelling_walsh.HFCausalModel"
  },
  "d_embed": 2048,
  "dim_feedforward": 8192,
  "dropout": 0.1,
  "embdding_cls": "torch.nn.Embedding",
  "embedding_args": {},
  "feedforward_args": {
    "beta": 0.25,
    "bias": true
  },
  "feedforward_cls": ".FeedforwardLayer",
  "head_args": {},
  "head_cls": ".Transformer",
  "init_gain": 1.0,
  "layer_args": {
    "alpha": 2.828427124746
  },
  "layer_cls": ".DeepnetLayer",
  "layer_stack_args": {},
  "layer_stack_cls": ".TransformerLayerStack",
  "loss_function": ".causal_loss",
  "max_sequence_length": 16384,
  "model_type": "walsh-causal-v1",
  "norm_args": {
    "normalized_shape": 2084
  },
  "norm_cls": "torch.nn.LayerNorm",
  "num_attention_heads": 32,
  "num_hidden_layers": 32,
  "output_proj_args": {},
  "output_proj_cls": "torch.nn.Linear",
  "pad_index": null,
  "positional_encoder_args": {
    "d_embed": 2048,
    "gain": 0.3333,
    "max_seq": 16384
  },
  "positional_encoder_cls": ".RSWalshPositionalEncoder",
  "torch_dtype": "bfloat16",
  "transformer_args": {},
  "transformer_cls": ".Transformer",
  "transformers_version": "4.37.2",
  "vocab_size": 32000
}