{ | |
"_attn_implementation_autoset": true, | |
"a_low_rank_dim": 96, | |
"attn": null, | |
"attn_mode": "chunk", | |
"bos_token_id": 1, | |
"decay_low_rank_dim": 96, | |
"eos_token_id": 2, | |
"fuse_cross_entropy": true, | |
"fuse_norm": true, | |
"gate_low_rank_dim": 256, | |
"head_dim": 64, | |
"hidden_act": "sqrelu", | |
"hidden_ratio": 4.0, | |
"hidden_size": 2048, | |
"initializer_range": 0.02, | |
"intermediate_size": 8192, | |
"max_position_embeddings": 2048, | |
"model_type": "rwkv7", | |
"norm_bias": true, | |
"norm_eps": 1e-05, | |
"norm_first": true, | |
"num_heads": null, | |
"num_hidden_layers": 24, | |
"tie_word_embeddings": false, | |
"torch_dtype": "float32", | |
"transformers_version": "4.48.1", | |
"use_cache": true, | |
"v_low_rank_dim": 64, | |
"vocab_size": 65536 | |
} | |