bumblebee-testing/tiny-random-Gemma3TextForSequenceClassification
054
1{2 "_sliding_window_pattern": 2,3 "architectures": [4 "Gemma3TextForSequenceClassification"5 ],6 "attention_bias": false,7 "attention_dropout": 0.0,8 "attention_probs_dropout_prob": 0.1,9 "attn_logit_softcapping": null,10 "bos_token_id": 2,11 "dtype": "float32",12 "eos_token_id": 1,13 "final_logit_softcapping": null,14 "head_dim": 256,15 "hidden_act": "gelu_pytorch_tanh",16 "hidden_activation": "gelu_pytorch_tanh",17 "hidden_dropout_prob": 0.1,18 "hidden_size": 32,19 "initializer_range": 0.02,20 "intermediate_size": 37,21 "layer_types": [22 "sliding_attention",23 "full_attention"24 ],25 "max_position_embeddings": 512,26 "model_type": "gemma3_text",27 "num_attention_heads": 4,28 "num_hidden_layers": 2,29 "num_key_value_heads": 2,30 "pad_token_id": 0,31 "query_pre_attn_scalar": 8,32 "rms_norm_eps": 1e-06,33 "rope_local_base_freq": 10000.0,34 "rope_scaling": null,35 "rope_theta": 1000000.0,36 "sliding_window": 128,37 "sliding_window_pattern": 2,38 "transformers_version": "4.57.0",39 "type_vocab_size": 16,40 "use_bidirectional_attention": false,41 "use_cache": true,42 "vocab_size": 102443}44 