{ | |
"_name_or_path": "rhymes-ai/Aria", | |
"architectures": [ | |
"AriaForConditionalGeneration" | |
], | |
"auto_map": { | |
"AutoConfig": "modeling_aria.AriaConfig", | |
"AutoModelForCausalLM": "modeling_aria.AriaForConditionalGeneration" | |
}, | |
"ignore_index": -100, | |
"image_token_index": 9, | |
"initializer_range": 0.02, | |
"model_type": "aria", | |
"pad_token_id": 2, | |
"projector_patch_to_query_dict": { | |
"1225": 128, | |
"4900": 256 | |
}, | |
"text_config": { | |
"hidden_size": 2560, | |
"intermediate_size": 1664, | |
"max_position_embeddings": 65536, | |
"model_type": "aria_text", | |
"moe_num_experts": 64, | |
"moe_topk": 6, | |
"num_attention_heads": 20, | |
"num_experts_per_tok": 6, | |
"num_hidden_layers": 28, | |
"num_key_value_heads": 20, | |
"rope_theta": 5000000, | |
"vocab_size": 100352 | |
}, | |
"torch_dtype": "bfloat16", | |
"transformers_version": "4.47.0.dev0", | |
"vision_config": { | |
"_flash_attn_2_enabled": true, | |
"architectures": [ | |
"AriaVisionModel" | |
], | |
"attention_heads": 16, | |
"image_size": 980, | |
"intermediate_size": 4304, | |
"model_type": "idefics3_vision", | |
"num_hidden_layers": 27, | |
"patch_size": 14, | |
"torch_dtype": "bfloat16" | |
}, | |
"vision_feature_layer": -1 | |
} | |