{ "vocab_size": 2048, "hidden_size": 256, "intermediate_size": 1488, "num_hidden_layers": 3, "num_attention_heads": 8, "num_key_value_heads": 1, "head_dim": 32, "ngram_buckets": 156250, "loop_schedule": [ 0, 1, 1, 2 ], "refresh_kernel_size": 9, "max_position_embeddings": 4096, "rope_theta": 100000.0, "rms_norm_eps": 1e-05, "initializer_range": 0.02, "use_cache": false, "transformers_version": "5.16.1", "architectures": [ "CherryAlphaForCausalLM" ], "output_hidden_states": false, "return_dict": true, "dtype": null, "chunk_size_feed_forward": 0, "is_encoder_decoder": false, "id2label": { "0": "LABEL_0", "1": "LABEL_1" }, "label2id": { "LABEL_0": 0, "LABEL_1": 1 }, "problem_type": null, "_name_or_path": "Banaxi-Tech/cherry-alpha", "tie_word_embeddings": true, "bos_token_id": 1, "eos_token_id": 2, "pad_token_id": 0, "unk_token_id": 3, "model_type": "cherry_alpha", "output_attentions": false, "auto_map": { "AutoConfig": "configuration_cherry_alpha.CherryAlphaConfig", "AutoModelForCausalLM": "modeling_cherry_alpha.CherryAlphaForCausalLM" }, "torch_dtype": "float32" }