bumblebee-testing/tiny-random-Qwen3ForSequenceClassification

1{2  "architectures": [3    "Qwen3ForSequenceClassification"4  ],5  "attention_bias": false,6  "attention_dropout": 0.0,7  "attention_probs_dropout_prob": 0.1,8  "dtype": "float32",9  "head_dim": 128,10  "hidden_act": "gelu",11  "hidden_dropout_prob": 0.1,12  "hidden_size": 32,13  "initializer_range": 0.02,14  "intermediate_size": 37,15  "layer_types": [16    "full_attention",17    "full_attention"18  ],19  "max_position_embeddings": 512,20  "max_window_layers": 28,21  "model_type": "qwen3",22  "num_attention_heads": 4,23  "num_hidden_layers": 2,24  "num_key_value_heads": 2,25  "pad_token_id": 0,26  "rms_norm_eps": 1e-06,27  "rope_scaling": null,28  "rope_theta": 10000.0,29  "sliding_window": null,30  "tie_word_embeddings": false,31  "transformers_version": "4.57.0",32  "type_vocab_size": 16,33  "use_cache": true,34  "use_sliding_window": false,35  "vocab_size": 102436}37