inference/model_config.json

{
  "_name_or_path": "DeepXR/Helion-OSC",
  "architectures": [
    "HelionOSCForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "bos_token_id": 1,
  "eos_token_id": 2,
  "pad_token_id": 0,
  "hidden_act": "swiglu",
  "hidden_size": 8192,
  "initializer_range": 0.02,
  "intermediate_size": 28672,
  "max_position_embeddings": 32768,
  "model_type": "helion-osc",
  "num_attention_heads": 64,
  "num_hidden_layers": 80,
  "num_key_value_heads": 8,
  "pretraining_tp": 1,
  "rms_norm_eps": 1e-5,
  "rope_scaling": {
    "factor": 8.0,
    "high_freq_factor": 4.0,
    "low_freq_factor": 1.0,
    "original_max_position_embeddings": 8192,
    "rope_type": "llama3"
  },
  "rope_theta": 500000.0,
  "tie_word_embeddings": false,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.40.0",
  "use_cache": true,
  "vocab_size": 128256,
  "sliding_window": null,
  "mlp_bias": false,
  "head_dim": 128,
  "quantization_config": null,
  "gradient_checkpointing": false,
  "use_flash_attention_2": true,
  "attn_implementation": "flash_attention_2",
  "rope_interleaved": false,
  "output_attentions": false,
  "output_hidden_states": false,
  "use_return_dict": true,
  "torchscript": false,
  "use_bfloat16": true,
  "tf_legacy_loss": false,
  "pruned_heads": {},
  "chunk_size_feed_forward": 0,
  "is_encoder_decoder": false,
  "is_decoder": false,
  "cross_attention_hidden_size": null,
  "add_cross_attention": false,
  "tie_encoder_decoder": false,
  "max_length": 32768,
  "min_length": 0,
  "do_sample": false,
  "early_stopping": false,
  "num_beams": 1,
  "num_beam_groups": 1,
  "diversity_penalty": 0.0,
  "temperature": 0.7,
  "top_k": 50,
  "top_p": 0.95,
  "typical_p": 1.0,
  "repetition_penalty": 1.05,
  "length_penalty": 1.0,
  "no_repeat_ngram_size": 3,
  "encoder_no_repeat_ngram_size": 0,
  "bad_words_ids": null,
  "num_return_sequences": 1,
  "output_scores": false,
  "return_dict_in_generate": false,
  "forced_bos_token_id": null,
  "forced_eos_token_id": null,
  "remove_invalid_values": false,
  "exponential_decay_length_penalty": null,
  "suppress_tokens": null,
  "begin_suppress_tokens": null,
  "architectures_info": {
    "model_description": "Helion-OSC is a large-scale transformer model optimized for code generation and mathematical reasoning",
    "total_parameters": "405B",
    "active_parameters": "37B",
    "num_experts": 160,
    "experts_per_token": 6,
    "shared_expert_intermediate_size": 14336,
    "moe_intermediate_size": 4096,
    "routed_scaling_factor": 1.0,
    "scoring_func": "softmax",
    "aux_loss_alpha": 0.001,
    "seq_aux": true,
    "topk_method": "gready",
    "n_group": 1,
    "topk_group": 3,
    "num_shared_experts": 2,
    "norm_topk_prob": false,
    "route_scale": 1.0,
    "moe_layer_freq": 1,
    "first_k_dense_replace": 1,
    "architecture_type": "MoE-Transformer"
  },
  "safetensors_configuration": {
    "total_shards": 116,
    "shard_size_gb": 3.01,
    "total_size_gb": 349.16,
    "weight_map_file": "model.safetensors.index.json",
    "shard_pattern": "model-{:05d}-of-{:05d}.safetensors",
    "max_shard_size": "3.1GB",
    "metadata_included": true
  }
}